SCROLL NEWS / DISCOVERY
Search headlines
Find the stories shaping the conversation.
Results for " basel " 88 found 🔀 AI-powered Shuffle
Novak Djokovic Set for Basel Return After 15 Years
24-time Major champion Novak Djokovic is all set to make a surprise addition to his season this year. The Serbian, who won his 102nd ATP Tour title in Beijing a couple of weeks ago, has announced that he will be taking part in the ATP 500 in Basel later this month, marking his return to …
Is the EU research budget actually increasing?
Is a Horizon Europe budget of €158 billion a 10% cut or a significant increase to the EU’s support for research and innovation? What would seem like a question of fact has divided opinion in Brussels since the Irish presidency of the EU Council presented its negotiation box on October 10. The answer depends on factors such as how you adjust for inflation, whether you account for increases in overall EU wealth and if you care about absolute or proportional figures. The €158 billion budget in current prices, proposed by the Irish presidency of the EU Council, was quickly criticised by university lobbies and rejected by the high-level expert group led by former Portuguese science minister Manuel Heitor. It was also decried by the four pro-European political groups in the European Parliament. “How can Europe close its competitiveness gap while cutting the European Competitiveness Fund by 13% and Horizon Europe by nearly 10%?” asked a joint statement by the European People’s Party (EPP), Renew Europe, the Socialists and Democrats and the Greens/European Free Alliance. Together, these parties hold 63% of seats in the Parliament.In stark contrast, Irish foreign affairs minister Helen McEntee said that the proposal “significantly increases investment in competitiveness, research and innovation,” Meanwhile, a presidency press release said that, compared to the current Multiannual Financial Framework (MFF), the proposal would “double the volume of competitiveness related programmes.” So, who is right? That depends on how you make the calculation.Compared to the Commission figure or the previous budget?The first thing to notice is the difference in baselines. When the MEPs say the Irish presidency will cut Horizon Europe by “nearly 10%” they are using the European Commission proposal for the MFF from July 2025, which allocated €175 billion in current prices to Horizon Europe. However, when the Irish foreign affairs minister said the research and competitiveness budget is increasing significantly, she is comparing it to the 2021-27 MFF, which had a Horizon Europe budget of €84.9 billion in 2018 prices.Her point holds when accounting for inflation. Applying a 2% annual inflation rate, the calculation method used by the Commission, the 2021-27 Horizon Europe would have a budget of €97.5 billion in 2025 prices. In comparison, the Irish presidency’s €158 billion figure in current prices translates to €140 billion in 2025 prices. Does the ECF count?But there is more to the dispute. In its statements, the Irish presidency lumped the research budget in with the overall budget for competitiveness. The Parliament statement, however, kept the two separate. This matters because the next MFF includes plans to add a new European Competitiveness Fund (ECF) to fund programmes meant to boost the continent's competitiveness. The Irish negotiating box allocated €245 billion in current prices to this programme.When this is added to the €158 billion for Horizon Europe, as well as programmes such as the EU education programme Erasmus+, the total for “competitiveness, prosperity and security” in the Irish proposal becomes €514 billion in current prices. But it is still uncertain how big a share of this competitiveness budget would go to research. The Commission proposal of July 2025 for the MFF included clear links between the ECF and Horizon Europe, indicating some ECF money will go to research. At the same time, the structure of the ECF is still being negotiated, and a substantial part of its budget is expected to support non-research activities such as defence manufacturing, decarbonisation and digital infrastructure. Related articlesCouncil floats 9.6% budget cut to next Horizon EuropeResearch community to appeal to national governments over planned EU budget cutEU own resources ‘essential’ for future research budget, says leading MEPSo, it is by looking at competitiveness rather than research, or combining the two, that the Irish presidency can claim that its proposed budget will significantly increase investment in competitiveness, research and innovation.According to the Irish presidency proposal, Horizon Europe and other competitiveness-related programmes in the current multiannual budget received €221 billion in current prices. However, since this money is being distributed from 2021 to 2027, it is not directly comparable to the €514 billion in current prices which will be distributed from 2028 to 2034.Research as a proportion of the MFFBut the relative rise in competitiveness-related programmes points to another important distinction: whether you look at the research budget in absolute terms or as a share of the overall MFF. Here, too, the size of the research budget has risen, although only slightly. In the 2021-27 MFF, Horizon Europe accounted for 7% of the total budget. In the Irish presidency’s proposal, it amounts to 8.7%. The share for the total competitiveness envelope went from 17.7% in the current MFF to 28.1% in the Irish presidency proposal.This proportional share may seem irrelevant to universities looking for research funding, but it is becoming significant as budget negotiations between the Parliament, the Commission and the Council begin. On October 2, a group of 17 southern and eastern European countries sent out a statement calling for the EU agriculture and cohesion budgets to be protected. The Parliament’s co-rapporteur on the budget, Siegfried Mureșan of the EPP, has also said that the research envelope is likely to be cut disproportionately as the budget is slimmed down during negotiations.
Nearly All Soldiers Completed Army Brain Health Test, Service Says
The DOD established the Neurocognitive Assessment Tool (NCAT) in 2024 to track soldiers' baseline cognitive health.
How worried should Catholics be about the pope’s health?
Pope Leo XIV announced on Sunday that he will undergo surgery this week to remove a “nodule” from his lung. The pope made the announcement at the end of his Sunday Angelus address, asking for prayers. “Dear all, recently, during some checkups, the doctors discovered a nodule in my lung. It was detected at a very early stage and was well-localized,” Leo said. “The doctors felt it was necessary to intervene, and so, following their advice, I decided to have surgery to remove it. The procedure is already scheduled for the next few days. I thank you for your spiritual support and I entrust myself to your prayers.” The Vatican press office subsequently added that Leo is scheduled to undergo surgery at the Gamelli hospital on Tuesday, Oct. 13. The pope is expected to spend up to a week in the Gamelli recovering, after which, for the moment, he is still expected to undertake his largest and longest to-date Apostolic journey, heading to Argentina, Uruguay, and Peru, beginning Nov. 6. Prayers have, understandably, poured in from dioceses around the world, accompanied by the best wishes of heads of state and international institutions. So how worried should Catholics be for the health of the pope? The Vatican’s messaging around Leo’s pending surgery is, in many ways, a new approach to papal health issues. But, despite the pope’s personal candour, the Holy See’s messaging perhaps raises as many questions as it answers. Subscribe — Leo’s decision to announce his own medical status and pending hospital stay during the Sunday Angelus is, it should be said, already something of a big deal. Traditionally, the health of popes simply hasn’t been discussed publicly or officially, until circumstances made it absolutely necessary. Memorably, in 2023 Pope Francis was rushed to the Gemelli in an ambulance, exhibiting signs of cardiac arrest according to many close to him, only for the Sala Stampa to put out a statement insisting he was merely going in for “pre-scheduled” tests. That Pope Leo has decided to announce his condition himself, with remarkable frankness, is a significant comms policy shift around the Apostolic Palace. This might reflect a deliberately more open posture, as opposed to the traditional, otiose evasiveness, and the first fruits of an American pope and American prefect of the Dicastery for Communications favoring honesty as a baseline. Equally, it must be allowed, it could be that the pope’s situation is serious enough that there would be no keeping it under wraps or playing it off, and the decision was made to be as open as possible to head off speculation about the papal health. Certainly, Leo appeared to make every effort to sound reassuring — stressing that, whatever the diagnosis is, exactly, it is “well-localized” and “early stage.” Of course, for everyone listening, talk of a “nodule” in a lung will immediately send thoughts — and prayers — running to concerns about words not used, like “cancer,” “tumor,” “begnin” or “malignant.” To be clear, none of those words may yet be medically appropriate. And there is every reason to believe and assume that Leo is giving almost as much information as he may have himself at this stage. But the context of the pope’s announcement and the timing of his surgery do reasonably raise reasonably serious concerns for Leo. We don’t know, and this isn’t to say anyone has the right to expect to know, when the routine medical checks made this discovery. It could have been last week, last month, or even earlier. As such, it isn’t really possible to say with what kind of urgency Leo is being brought in for surgery. However, there is clearly some serious urgency to the matter. Leo making the announcement and going into hospital this week, of all weeks, means he is having to miss the balance and conclusion of his global meeting of the heads of the world’s bishops conferences, which he convened in Rome this week. That, as the Vatican has announced, has meant some serious changes to the schedule, with the newly minted dean of the College of Cardinals, Cardinal Tagle, subbing in for the closing Mass on Wednesday, and the meeting’s organizers announcing that the final working document/summary of the meeting being delayed until Leo can read and feed into it. Those changes suggest that the timing of Leo’s surgery isn’t of his own choosing, and he is acting on the best medical advice he can get by moving urgently. Had this issue been discovered weeks ago, it seems reasonable to assume the pope would have seen it dealt with before the meeting began. And were it possible to delay it, even by a few days to allow the meeting to conclude as scheduled, the expectation would surely be he’d have opted for that. It cannot be excluded that Leo is simply having to fit in with the schedules of his medical team, but a man who occupies arguably the most important and prominent role in Italy, if not Western Europe, could generally be assumed to be the one whose calendar others work around. All this points to a very recent medical discovery leading to a “days matter” diagnosis, or at least suspected possible diagnosis. Of course, another factor weighing in the balance is Leo’s upcoming apostolic visit to three South American countries, including a return to Peru, where he spent many years as a priest and bishop. It is entirely possible that the doctors, and the pope himself, are more concerned about the potential knock-on effects of surgery, and the time needed for recovery, for a 71-year-old — even one as physically sound and actively sporty as Leo. The pope is due to depart Rome November 6 for his 10-day three-country trip, returning on Nov. 17. It may be that Leo has simply made the decision to allow maximum time for his own post-op recovery, and even allow for potentially unforeseen circumstances, so that he is at full strength for the landmark trip. While it might be an inconvenience, and clearly a last-minute wrench in the works for the Amoris meeting, it is possible that Leo simply weighs the importance of that conference more lightly in the balance than even the possibility of missing his homecoming trip to Peru, or even having to scale back his itinerary for the trip. If that is the case, it would suggest a real abundance of caution by the pope approaching his Latin American tour, and a sobering indication of where Leo ranks the non-synodal meeting currently underway in Rome. In any event, what can be said for certain is that the pope having to take this, of all weeks, out of his schedule for surgery and an extended hospital stay is extremely unfortunate timing. It is clear that this is a decision Leo has made on the advice of his doctors. Together, they clearly consider that the risks of delaying, even by a little as a few days, represents an unacceptable risk for the pope. Even expecting and allowing for Leo’s full and speedy recovery, that should be a very sobering thought for Catholics everywhere. Subscribe
House Democratic leaders condemn the search of congresswoman's home as an 'abuse of power'
“For years, Donald Trump and far-right extremists have relentlessly targeted Congresswoman Omar with vile personal attacks and baseless accusations,” House Democratic leaders said in a statement.
Before Saying Yes to the Medical Test, Know the Risks and Benefits
Tests like full-body scans marketed directly to consumers are pitched as a way to detect diseases early and establish a baseline of health markers. But there are financial, physical, and psychological risks to all medical tests — and in some cases, not getting a test could be better for your health.
Trump, Critic of Voting by Mail, Urges Republicans to Do It in Midterms
It was the latest mixed messaging from President Trump, who regularly votes by mail, but has made baseless claims that the practice is “inherently corrupt.”
Syracuse Basketball Stars Sadiq White Jr. & Olivia Schmitt Join Apex Entertainment's NIL Family
New Men's Head Coach Gerry McNamara to Meet Orange Fans on October 15 at Apex Entertainment's Destiny USA LocationSYRACUSE, NY / ACCESS Newswire / September 23, 2026 / Apex Entertainment®, the …
Scientists tested whether pets really make us happier: the answer is surprising - Futura-Sciences
Pets are often described as good for both mental and physical health, helping ease loneliness and support emotional well-being. But a new study suggests the picture is more complicated than simply owning a dog, cat, or another companion animal. Researchers at the University of Basel in Switzerland found little evidence...
Health experts urge young adults to keep primary care appointments
Mercy Health and Warren County health experts say young adults without primary care can miss baseline checks and early disease warning signs.
Engrams Embedding Entendre: Codesign for Efficient DRAM/SSD Offloading
Engram extends standard token embeddings with learned multi-token lookups. Recurring local patterns retrieve vectors directly, reducing the need to reconstruct them through attention and feed-forward layers. With Engram model architecture optimization, it allows for lower HBM capacity to be needed for models at the same quality. This does not mean there won’t be an insane demand for HBM but it just means that model architecture will continue to innovate around constraints. This model architecture design is naturally codesigned for parameter offloading: each token accesses a few embedding rows whose addresses depend on token IDs, not hidden states. The runtime can prefetch those rows from host DRAM while earlier layers compute, keeping the table outside HBM without transferring entire weight matrices. Our Memory model contains our latest estimates of quarter by quarter HBM, DRAM, & NAND supply and demand. Offloading frees HBM for model weights and KV cache, potentially supporting larger batches or more concurrent sessions. When DRAM becomes the next constraint, NVMe offers another tier. Recommendation systems already cache frequently or recently accessed embedding rows in faster memory while backing colder rows with SSDs. After NVIDIA roadmap had to change due to massively despec’ing Rubin Ultra from 1024GB to now ~200GB of HBM per chip, model architecture optimizations like emgram maybe helpful. Our DeepSeek-V4.1-Flash configuration uses roughly 189 GiB of memory for Engram. We replace it with a memory-mapped (mmap) file and measure serving performance with offload to SSD. Later on in our report, we will show our Engram offloading experiments along with the official InferenceX agentic inference serving results on Engram models like DeepSeekv4.1 Flash across all 6 NVIDIA GPU SKUs along with MI355X. Unsurprisingly, the CUDA Moat is still mogging MI355X on the ultra popular DeepSeekV4.1 Flash model. We also show how even on high capacity HBM SKUs, offloading emgrams to DRAM could result in even better performance for most of the pareto than keeping the emgram in HBM. Our benchmark has been widely reproduced, validated and/or supported by almost every major buyer of compute from Google Cloud to Microsoft Azure to Oracle, to Meta and many more. Furthermore, it has the support of the ML community including from vLLM, LMCache, SGLang, PyTorch, Huggingface and the support of major labs like OpenAI, MiniMax, ZAI, Qwen, Moonshot Kimi, etc. Star the InferenceX GitHub repository if you find the open-source benchmark and data useful!. InferenceX is the only inference benchmark in the world to have TPUv7, Jalapeño, Nvidia Rubin NVL72, AMD, and soon, SambaNova and Trainium. Due to how realistic AgentX scenario is to real world agentic inference workloads, AMD has committed to collaborating on MI455X UALoE72 too. DeepSeek did not release the original paper’s two trained Engram models. We replicated its setup on fineweb-edu using the released code and training hyperparameters, at an estimated 6E18 FLOPs per run. We observed the same U-shape scaling: Engram improved performance over pure MoE baselines. We also reproduced DeepSeek’s results, where earlier-layer representations with Engram resembled those of later layers. Like the original Engram paper, it is possible to probe Engram’s gate scores to see what n-grams DeepSeek-V4.1-Flash makes the most use of. Our gate scan finds names, code fragments, relational phrasing, and boilerplate. These examples prioritize interesting-ness over gate strength. One unexpected result was Wright : Ace Attorney. These examples suggest learned memory optimizes the training objective, not a judgment of which facts deserve storage. Licenses, bibliography fragments, API scaffolding, and website furniture can provide prediction shortcuts, so the value of additional Engram capacity may depend on what survives data preparation. This does not show that table capacity is “wasted”: the evaluation-corpus scan establishes neither training exposure nor the capacity occupied by each category. For offloading, strong gates do not identify cache-hot rows. Low gates do not automatically save reads either: computing the gate requires the retrieved key and undoes the performance gain of a fused kernel. Skipping reads would require a separate usefulness predictor before retrieval. In the original paper’s inference-time ablation, factual-knowledge benchmarks retained just 29–44% of their original performance, while reading comprehension retained 81–93%. This is due to the training–inference mismatch. The resulting degradation therefore measures this trained model’s dependence on Engram, not the performance difference between models trained with and without it. In our ablations, suppressing Engram worsens token likelihood across all evaluated domains, especially encyclopedia text and several code corpora. Surprisingly, GSM8K accuracy stays within measured run-to-run variation and removing Engram has no effect. Engram is not a detachable dictionary beside an unchanged MoE. Removing it changes downstream features and expert selection. We tested whether rerouting hurts or compensates by holding tokens fixed in a teacher-forced experiment on CRUXEval, a code-reasoning benchmark of small Python functions where the model predicts a function’s output from its code and an input, and scoring the reference answer. Removing Engram raised answer loss from 0.2848 to 0.3093 bits/token. Forcing the ablated model to use the original Engram-on expert choices made it worse still, at 0.3375 bits/token. Rerouting partially compensates for the missing memory. Memory features and expert selection work together, rather than following a clean “memory stores facts; experts reason” division. On the same CRUXeval, removing Engram during either phase reduced accuracy and increased generated tokens; removing it throughout produced the largest changes. Keeping Engram for prefill leads to more correct answers than keeping for decode likely due to semantically richer KV cache transferred to decode workers, allowing it to mitigate some of the performance loss. Engram’s table is large, but each lookup is small. DeepSeek-V4.1-Flash requests 24 rows at each of two Engram layers, about 12.4 KiB per processed token position across the model, or 3.1 KiB per GPU when split across four GPUs. Currently as of Day 7 since Model Release, MI355X is still 2-4x worse performance per dollar compared to B200 even when normalized by Mi355X’s lower TCO. Our full total cost of ownership breakdown comes from our AI Cloud TCO Model along with monthly market surveys of over 100+ gpu clouds & gpu cloud customers. On the Day 0 release of DeepSeekv4.1 Flash, NVIDIA vLLM works out of the box with zero issues across all 6 SKUs: H100, H200, B200, B300, GB200, GB300! This was thanks to the amazing work by the NVIDIA & Interact teams! In comparison, AMD vLLM did not work on day 0 for DeepSeekv4.1 Flash.Source: SemiAnalysis InferenceX AMD’s vLLM documentation points to using vllm/vllm-openai-rocm:deepseekv41-flash-0909, but from hour 0 of the model release to hour 23, AMD has not publicly released the image. AMD claims, “SPEED IS THE MOAT,” yet it has still not released it by the 23rd hour. We wish that, going forward, the AMD team has a better process for hour 0 model releases. Eventually when they did publicly release for “day 0” image support, performance-wise, it is currently up to 14.8x worse perf per dollar than H200 and up to 42x worse perf per dollar than B200/B300. The power of the CUDA MOAT is NVIDIA’s collaboration with its massive 6 million-developer community ecosystem including most of the vLLM & SGLang & Tokenspeed maintainers which means that CUDA is optimized on day 0. Overall, AMD did make significant improvements but the performance per dollar is still currently 2-4x worse than B200. AgentX Engram DRAM Offloading Improving Performance The HBM and DRAM offload use the same GPU kernel to select and dequantize rows. With HBM, it reads GPU memory; with Unified Virtual Addressing (UVA), it reads pinned host memory directly. Both support full decode graphs. Moving the table into HBM accelerates only the sparse lookup, leaving decoder computation and communication unchanged, resulting in little overall benefit while consuming memory otherwise available to KV cache. Another benefit of Engram offloaded to DRAM which means you can reduce the communication overhead by using less HBM GPUs per replica. For example, when enabling Engram offloading on B300, we are able to switch from TP4 to now TP2 which improves the pareto curve by up to 1.6x. When iso-model quality, less HBM is required as the engrams could be offloaded to host DRAM. Thus HBM bandwidth matters way more than HBM capacity. For inference workloads where memory bandwidth matters the most, 4-hi HBM provides the best $/bandwidth and therefore lowest cost per token. If China continues to make more and more revolutionary model architecture innovations, soon it could potentially 0Hi HBM stacks. [ ](https://newsletter.semianalysis.com/p/long-live-the-short-king-why-4-hi) [ Long Live the Short King: Why 4-hi HBM Wins ](https://newsletter.semianalysis.com/p/long-live-the-short-king-why-4-hi) Myron Xie, Bryan Shan, and 3 others · Sep 13 [ Read full story ](https://newsletter.semianalysis.com/p/long-live-the-short-king-why-4-hi) Moreover, on B300 and week-0 stack, moving the Engram table back to HBM did not improve results and stay within run-to-run variance. This is a result of the work optimizing DRAM offload, such as async, and overlap. SSD Offloading On B200, we created a unoptimized vLLM fork and stored the Engram tables in memory-mapped files on local SSDs. File backing lets the OS reclaim table pages when other applications need RAM. Pages already cached in memory can be served without reading the SSD again. Furthermore, note that we were unable to turn on GDS The unoptimized SSD implementation changes how rows reach the GPU. It copies row IDs to the CPU, deduplicates them, gathers the requested rows into pinned buffers, copies those rows back to the GPU and dequantizes them. This work runs between segments of the GPU execution graph. Native UVA performs row selection and dequantization directly on the GPU, avoiding the CPU round trip. A warm filesystem cache removes physical SSD reads, but leaves the coordination, row gathering and transfers. This is why a file already cached in RAM can still perform worse than a pinned DRAM table. The comparison measures the whole serving path; it does not separate the time spent on each of these operations. B200 DRAM dominates both measured SSD serving curves in total tokens per dollar and P90 interactivity. Near 125 tokens/s/user, DRAM delivers 121 million total tokens per dollar versus 52 million for SSD. For production serving, SSD offloading is likely not worth the tradeoff. On the B200 configurations we measured, SSD offloading loses on both measures: every observed SSD point has a DRAM alternative that delivers higher P90 interactivity and more total tokens per dollar. The only points where Cheaper storage does not automatically produce a cheaper inference service. Moving Engram to SSD leaves the same four expensive GPUs and the rest of the server in place. Reclaiming RAM only creates an economic benefit if it enables a cheaper server configuration or additional useful capacity. The current unoptimized path provides neither benefit, and the filesystem cache still consumes RAM when table pages are resident. Next we will look at the mechanisms and specific implementation of ngrams from DeepSeekv4.1 Flash, LongCat, Qwen3.8 Flash Next.
Health-ISAC sets nine-domain MedTech cybersecurity baseline to guide medical device procurement, deployment
Health-ISAC sets nine-domain MedTech cybersecurity baseline to guide medical device procurement, deployment and risk management
Statewide assessments demonstrate ELA and math recovery accelerate as science reaches record high
The Utah State Board of Education release 2025-2026 statewide assessment results in Utah show English Language Arts (ELA) and mathematics recovery accelerating, and science results reaching record highs. The Big Picture The 2025–2026 Utah statewide assessment results demonstrate continuous multi-year academic growth following pandemic-related disruptions. Across all core subjects, student performance increased year-over-year, reflecting sustained multi-year gains.Most notably, Science performance has exceeded pre-pandemic baseline levels, while English Language Arts (ELA) and Mathematics are closing in on full recovery. Assessment results are available interactively […]
What Connects the Last 5 US Open Men’s Champions
Five different men have lifted the US Open trophy since 2020. Their games look nothing alike - Daniil Medvedev's flat, defensive counterpunching has little in common with Carlos Alcaraz's explosive shot-making. There might be some more similarities between Novak Djokovic's clinical baseline and Jannik Sinner's mechanical precision, but they are hardly simulacrums of each other. …
Bone Health Without the Hype
I am at the Australasian Menopause Society meeting and we had a fantastic lecture on Bone Health on Day 2, called, “Everything you need to know about bones - except the boring bits!” Spoiler alert, there were no boring bits. The lecture was given by Dr. Amanda Beech, who is a senior staff specialist in Obstetric Medicine and Endocrinology at the Royal Hospital for Women, where she is director of the Bone Density Unit. I want to summarize my top 10 take-aways for you. Share Osteoporosis is a disease of low bone mass, disruption of bone microarchitecture, and fragility of the skeleton. It is a silent disease…until someone develops a fracture. The diagnosis of osteoporosis is based on either fragility/minimal trauma fractures or a T-score of less than -2.5 on a DXA (Dual-Energy X-ray Absorptiometry) or bone scan. It is important to recognize that about 50% of minimal trauma fractures occur when women have low bone mass (often referred to as osteopenia), but not osteoporosis. This tells us that bone strength is more complicated than a number on a DXA report. I am going to add a little here that wasn’t in the talk (one can only include so much in 25 minutes), as many women might read the above paragraph and think, “Shouldn’t we be treating everyone with low bone mass if this is when 50% of the fractures occur?” Many fractures happen for women with low bone mass because there are so many women with low bone mass –it is a statement about the population on a whole, not about the fracture risk of low bone mass. For you as an individual, bone density is only one component of fracture risk and many women with lower bone mass will not have a significantly elevated fracture risk. This is why individualized risk assessment matters. Subscribe We build bone in our teens and twenties, and by the time we are about thirty we have reached peak bone mass. This is why, when we discuss risk factors for osteoporosis, it is important to consider factors that might have affected us in our teens and twenties when we were actively building bone, such as nutritional restriction/energy imbalance. It is important to ask women about a history of eating disorders, and for those with such a history, getting a bone scan in perimenopause or at age 50 is likely warranted. (This image below isn’t from the talk, but is similar to one that was). Bone mass is relatively stable until about two years before the final menstrual period, and perimenopause marks a time of rapid bone loss for women. Using urinary markers of bone loss we know that the time of significantly accelerated bone loss starts about 1 year before the final period and ends about 2 years after the final period. The Study of Women’s Health Across the Nation (SWAN) shows that over the menopause transition women lose an average of 10.6% of the bone at their lumbar spine and 9.1% at the femoral neck While this wasn’t covered in the lecture, I want to add that the rapid bone loss during perimenopause is clearly related to more than estradiol, because during this time estradiol levels can be erratic and not uniformly low. Bone loss during perimenopause is likely a multi-hormone event. For example, rising FSH (follicle stimulating hormone) tracks with bone loss better than estradiol, and there is evidence that FSH may act directly on bone. While decreased estrogen matters, there is clearly more going on. FSH and bone is an active area of research. Subscribe Screening for osteoporosis was recommended fairly broadly before age 65, and was taken from the “Practitioner’s Toolkit for Managing Menopause,” from Monash University. Here is a screenshot from the toolkit that was used as a slide: I am going to add a little bit more about body weight here for clarification. Low body mass (typically 152.1 lb): 0 points (risk neutral) 60–69 kg (132.3–152.1 lb): 3 points (adds some risk)
Crypto billionaires give $97 million to Reform UK, fueling concerns about money in politics
LONDON (AP) — Reform UK, the anti-immigration party led by Brexit campaigner Nigel Farage, received 72 million pounds ($97 million) from two cryptocurrency billionaires this weekend, fueling …
Science World selling pieces of Vancouver's giant FIFA soccer ball at an expensive price
Science World's FIFA World Cup soccer ball design will soon be gone forever. Or maybe not. Pieces of it are up for sale.
Fact check: Trump, RFK Jr. and others make false claims on night 2 of GOP convention
(CNN) — President Donald Trump delivered another rapid-fire series of false claims on the second night of Republicans’ midterm convention on Thursday. Health Secretary Robert F. Kennedy Jr. did his own barrage of wrongness. And various other speakers uttered statements that were untrue, misleading or missing critical context.Here is a fact check of some of the assertions from the final day of the convention. This is not intended as a comprehensive list of Thursday’s falsehoods.Trump’s many false claimsTrump made numerous false claims, some of which he also made in his longer speech Wednesday.Trump claimed that there were “no empty seats” in the Dallas arena as he spoke Thursday, which was visibly untrue. CNN White House correspondent Alayna Treene, who was there, wrote in a live chat right after Trump made the claim that “I’m looking at rows of empty seats” and that “the idea that there are ‘no empty seats’ is just laughably untrue,” though she also said the room was mostly full. Trump has for years made false claims about supposedly never having had an empty seat.Trump further inflated one of his favorite fictional figures, claiming Thursday that “$22 trillion” in investment has come into the US; on Wednesday, he had put it at “over $20 trillion,” and even that lower figure has no apparent basis in reality. As of the time he spoke Thursday, the White House’s own website claimed there had been “$11.2 trillion” in “major investment announcements” this term, not “over $20 trillion,” and the White House figure was itself a major exaggeration of actual investment.Trump, talking about elections, lied that “there’s never been bigger cheaters” than Democrats; there is no basis for broadly describing Democrats as election cheaters. For the second straight night, Trump falsely claimed that the Biden administration allowed “25 million people” to pour into the country, though that number is nowhere close to correct. Trump again falsely claimed that former President Joe Biden and other Democrats “created the worst inflation in American history,” though even peak Biden-era inflation, 9.1% in June 2022, was far from the all-time record of 23.7%; it was a 40-year high, but the rate then fell sharply before Trump returned to office.Trump also baselessly claimed that Democrats would “defund the police” and give “voting rights and welfare to 25 million illegal aliens” if they won the midterms. Republicans will still control the White House in 2027 and 2028 and the president would be in position to veto any such changes if a Democratic-controlled Congress sought to make them, but there is no indication that Democrats want to do so. Also, even groups favoring strict immigration policies estimate that the total undocumented population is far less than “25 million” or even 20 million.As he did Wednesday, Trump claimed the US now has the world’s lowest prescription drug prices. The assertion is so broad that it is unclear what exactly Trump is referencing. But experts told CNN they highly doubt that the prices for brand-name drugs in the US are generally less than they are in other countries, and the Trump administration has not provided details to prove the claim.“It’s tough to claim that you’re getting the lowest prices anywhere in the world and then also not have any evidence to support that,” said Andrew Mulcahy, a senior health economist at RAND, while Diederik Stadig, senior healthcare economist at financial institution ING, said, “To say that the US now has the lowest drug prices in the world is at this moment not true.”And as he did Wednesday, Trump said “more Americans are working today than ever before” – which is true but nearly meaningless, since the number of people working in the US generally rises over time simply because the population rises over time.“There are more Americans working because there are more Americans. That’s the entire analysis,” Justin Wolfers, a University of Michigan professor of public policy and economics, wrote in his newsletter in August.RFK Jr.’s false claims on the pandemic, Trump’s record and moreFalse claim about Democrats and the 2024 electionKennedy unsuccessfully ran for the 2024 Democratic presidential nomination before withdrawing to run as an independent and then eventually endorsing Republican candidate Donald Trump. In Kennedy’s Thursday convention speech, he complained of how the Democrats handled the nomination race.At one point, he said, “For the first time in history, they cancelled their own convention. No voter ever cast a primary ballot for Kamala Harris, no elected delegate was ever consulted; the party of ‘No Kings’ simply anointed a queen.”Kennedy is entitled to express unhappiness about how quickly Democratic leaders coalesced around then-Vice President Harris after Biden unexpectedly dropped out of the race in July 2024, after the primaries were over and less than four months from Election Day. But it’s not true that the party canceled its convention or that delegates didn’t get any say.The convention was held; it was a four-day event in Chicago in August 2024. Democratic convention delegates officially nominated Harris in a virtual roll call vote in advance of the convention. It’s true that Harris was the only candidate on the roll call ballot, but that’s because she was the only candidate who gathered enough delegate signatures to appear on the ballot.It’s clear that party elites’ backing of Harris played a key role in the unusual nomination process. But Kennedy’s specific allegations that the convention was abandoned and delegates were cut out of the process are not true.False claim that Democrats closed ‘every church’ during the Covid-19 pandemicWhile attacking Democrats, Kennedy claimed that, during the Covid-19 pandemic, “They closed every church in our country for a year.”It’s not true that every church in the country was closed for a year, nor even that every church in a Democratic state was closed for a year.Every state issued some type of guidance or order about social distancing during the Covid-19 pandemic, but many states – including some led by Democrats and some led by Republicans – incorporated religious exemptions into those limitations on social interactions.According to a Pew Research Center analysis from April 2020, 15 states allowed religious gatherings to continue without a limit on their size. That set of states included Colorado, Michigan and Pennsylvania, which all had Democratic governors. And at least two states with Republican governors, Alaska and Idaho, were among the states that prohibited religious gatherings at least briefly.False claim that Trump ‘ended nine wars’Trump has repeatedly claimed that he has ended “eight wars.” Kennedy went one higher on Thursday, declaring, “President Trump has ended nine wars.”That’s false. Trump hasn’t even ended eight wars.Trump’s list of eight includes two situations that were never actually wars during his tenure: a diplomatic dispute between Egypt and Ethiopia and a mysterious supposed situation between Serbia and Kosovo. The list also includes a war, involving Rwanda and the Democratic Republic of Congo, that hasn’t actually ended despite a peace agreement brokered by the Trump administration.In addition, the list includes the war between Israel and Hamas, but Israel continues to conduct frequent attacks in Gaza despite a US-brokered ceasefire agreement. And it includes the brief 2025 conflict between Israel and Iran, though Israel proceeded to join the US in starting another war against Iran in 2026.Claim about reversing a decline in longevityKennedy speaking of the Trump administration, claimed Thursday: “We have now reversed a 30-year decline in longevity. “Kennedy referenced preliminary data that he said he had received from the US Centers for Disease Control and Prevention this week, but it’s unclear what metric of longevity he’s referring to.CDC data shows that US life expectancy had been steadily increasing for decades before a sharp drop during the Covid-19 pandemic. Life expectancy had already started to rebound in 2022, under the Biden administration, and reached a new record high in 2024 before Trump’s second term began.Preliminary data shared by the CDC in July suggests that life expectancy probably hit another record high in 2025.CNN has reached out to HHS for clarification on Kennedy’s comments.Claim that Democrats ‘lost 360,000 children’Speaking about Democrats, Kennedy said: “They somehow lost 360,000 children and they turned our federal government into the biggest child trafficking operation in world history.”Trump and other Republicans have made similar claims. But there’s no basis for the assertion that Democrats “lost” so many children or that the federal government was engaged in human trafficking.A 2024 report from the inspector general for the Department of Homeland Security said Immigration and Customs Enforcement had reported that more than 32,000 unaccompanied migrant children failed to appear as scheduled for immigration court hearings after being released or transferred out of custody between fiscal years 2019 and 2023 – a period that, notably, includes two years and four months under the first Trump administration. The report also said that 291,000 unaccompanied migrant children during this period were not given notices to appear in court.The report said that ICE had “no assurances” these children “are safe from trafficking, exploitation, or forced labor.” But it did not say all the children were actually “lost” and did not definitively assert that any of them were trafficked – let alone that the government was somehow involved in a trafficking operation. And, again, the report spanned parts of both a Republican administration and a Democratic administration.Aaron Reichlin-Melnick, senior fellow at the American Immigration Council, told CNN in 2024: “Long story short, no, there are not 320,000 kids missing. 32,000 kids missed court. That doesn’t mean they’re missing, it means they missed court (either because their sponsor didn’t bring them or they are teenagers who didn’t want to show up). The remaining 291,000 cases mentioned by the (Office of the Inspector General) are cases where ICE hasn’t filed the paperwork to start their immigration court cases.”Vance on inflation, crime and taxes on Social SecurityA false claim about the media and Biden-era inflationVice President JD Vance mocked the media during part of his speech on Thursday. But he made a false claim about Biden-era inflation while lampooning how the media supposedly challenged him over another false claim on that subject.Vance said: “The president and I would go during a campaign, we’d say, ‘You know, inflation is as high as it’s been in 100 years,’ and the fake news would say, ‘No, no, no, it’s as high as it’s been in 68 years.’ That ain’t good, ladies and gentlemen.”Mainstream journalists, at least, wouldn’t say that inflation under Biden was as high as it had been “in 68 years.” Rather, CNN and others noted that peak inflation under the Biden administration, 9.1% in June 2022, was the highest in about 40 years.Vance is entitled to make his general point that this correction still doesn’t make Biden-era inflation sound good, but his specific number was, again, wrong. And he didn’t acknowledge that inflation fell sharply under Biden after hitting the 40-year high, declining to 2.9% in Biden’s last full month in office, December 2024, and 3.0% in January 2025, when Trump and Vance were inaugurated.The rate was 3.4% in July 2026, the most recent month for which data is available.Trump didn’t end taxes on Social SecurityVance claimed, “We’ve ended taxes on Social Security.” Several other convention speakers, including Rep. Lisa McClain of Michigan, Rep. Ryan Mackenzie of Pennsylvania and Ohio House candidate Derek Merrin, similarly said the administration has achieved “no tax on Social Security.”It hasn’t. The big domestic policy bill Trump signed in 2025 did create an additional, temporary $6,000-per-year tax deduction for individuals age 65 and older (which phases out for individuals earning $75,000 per year or more), but as the White House has implicitly acknowledged, millions of Social Security recipients will continue to pay taxes on their benefits even before the temporary new deduction expires after 2028.The Social Security Administration said in 2025 that “the bill ensures that nearly 90% of Social Security beneficiaries will no longer pay federal income taxes on their benefits.Claim about violent crime rate under BidenWhile criticizing the Biden administration, Vice President JD Vance claimed Thursday that the US “became a country where violent crime exploded, unless, of course, you could afford private security.”But violent crime did not explode under Biden. In fact, as FactCheck.org reported last month, data collected and published by the FBI shows that the national violent crime rate was at a lower level in 2023 and 2024, the last two full calendar years of Biden’s term, than it was in 2020, the last full calendar year of Trump’s first term. The year 2020 was highly unusual in many ways – violent crime jumped amid the Covid-19 pandemic – but Biden nonetheless left office with violent crime lower than it was when he took over. In fact, the 2024 violent crime rate was even lower than the 2019 rate under Trump before the pandemic.There was an increase in violent crime during the earlier part of Biden’s term. But given the subsequent decline, it’s misleading at best to claim violent crime “exploded” with no further context.Asked earlier this month about Trump’s similar previous claims to have reversed a Biden-era upswing in violence, crime data expert Jeff Asher said in an email: “No, crime began dropping significantly in 2023 during the Biden administration. Murder fell at record rates in 2023, 2024 and 2025 and, while 2026 may not feature a record decline nationally, it is still on track to see a sizable double-digit decline this year. Violent crime similarly fell nationally in 2023 and 2024, though the drops were more muted than the decline that was recorded by the FBI in 2025.”Other GOP speakersMike Johnson’s false claim about the inflation Republicans inherited from BidenSome other speakers at the convention took aim at inflation under Biden. House Speaker Mike Johnson claimed that “Republicans inherited from the Biden administration record-high inflation.”That’s not true. The inflation rate was 2.9% in Biden’s last full month in office, December 2024, and 3.0% in January 2025, when Trump took over. That’s lower than the most recent rate, 3.4% in July 2026, and nowhere close to the all-time record of 23.7%, which was reached in 1920, or the highest point of Jimmy Carter’s presidency, 14.8%, which was reached in 1980.Johnson correctly said moments later that inflation reached “a 40-year high of 9.1% inflation” under Biden. But a 40-year high is not an all-time record. More importantly, Johnson didn’t explain that 9.1% was much higher than the rate Trump inherited or that the 9.1% rate occurred in June 2022, with roughly two-and-a-half years left in Biden’s term.The cumulative increase in prices from the beginning of the Biden administration to the end was also not the highest in US history. Federal figures show that beginning-to-end inflation under Biden was less than half of that during Carter’s term.What Johnson left out of his claim about a congressional resolutionWhile attacking Democrats as extreme, Johnson said: “Just last week, nearly every Democrat in the House of Representatives voted against condemning socialism and Marxism. They voted against it.”But the resolution didn’t use the word “Marxism.” And Johnson didn’t mention that the resolution wasn’t only about condemning socialism. Rather, it contained other language that Democratic leadership said was the focus of the party’s opposition to the resolution.The resolution’s first provision said it “condemns and denounces socialism in all its forms, including the Democratic Socialists of America, and opposes the implementation of socialist policies in the United States.” But its fifth provision said it “calls for enactment of the SAVE America Act,” Trump’s favored elections bill, which contains various provisions opposed by many Democrats.So while it’s very possible many Democrats would have opposed the resolution even without its expression of support for the SAVE America Act, the call to pass this elections bill served as a poison pill that ensured most Democrats wouldn’t support it.“House Resolution 1490 explicitly calls for enactment of the so-called SAVE America Act, which is the top legislative priority of far-right extremists,” Democratic House leader Hakeem Jeffries said in a statement last week.A New Mexico candidate left out key context in an attack on his opponentGreg Cunningham, a Republican House candidate in New Mexico, used some of his Thursday speech to criticize his Democratic opponent, Rep. Gabe Vasquez. Cunningham said of Vasquez: “He even voted against a resolution honoring the sacrifices and service of the brave men and women in law enforcement who protect, every day, our communities.”But Cunningham left out important context about the resolution. While its focus was expressing support for law enforcement, its preamble contained statements praising the Trump administration and criticizing “leftist activists,” “progressive politicians” and “sanctuary city policies” for allegedly making law enforcement less safe.In other words, a House member might vote against such a resolution, as the vast majority of Democrats did, because they disagreed with conservative components of it, not the general expressions of support for law enforcement. (CNN has reached out to Vasquez’s campaign and House office for comment on his vote.) It’s common for parties to write resolutions in a way that would make it difficult for opponents to support but would open those opponents to campaign attacks if they voted “no.”An Ohio candidate’s misleading attack about Democrats and votingMerrin, the Republican candidate running against Democratic Rep. Marcy Kaptur for a House seat in Ohio, claimed in his speech, “While Democrats talk about letting illegal immigrants vote in our election, Republicans fight for proof of citizenship, photo ID and fair elections.”It’s a big country, so it’s possible some Democrat somewhere is talking about letting undocumented people vote in some election. But Merrin’s claim is misleading at best. There is no proposal from Democrats in Congress to let undocumented people vote in the 2026 midterms – the election Merrin appeared to be referring to when he said “our election” – or vote in future federal elections.CNN has reached out to Merrin’s campaign for an explanation of his claim.House Majority Leader Steve Scalise invoked a similar theme in his speech Wednesday, claiming that Democrats had voted last week “to allow illegals to vote in our federal elections.” But there was no congressional vote on letting undocumented immigrants vote in any elections.Rather, there was a vote on the non-binding Republican resolution that included multiple provisions, including one that “reiterates the notion that American elections are for American citizens only,” one supporting Trump’s favored elections bill and one denouncing socialism.Opposing that multi-clause resolution doesn’t necessarily mean someone supports voting by any non-citizens, let alone non-citizens in the country illegally.This article has been updated with additional items.CNN’s Deja Oliver contributed to this report.The-CNN-Wire™ & © 2026 Cable News Network, Inc., a Warner Bros. Discovery Company. All rights reserved.
Domestic demand drives business sentiment turnaround in South Korea
The overall outlook surpassed the baseline of 100 for the first time since before the US-Iran conflict.
Gael Monfils to Swap Baseline for Boardroom
Gael Monfils, one of the most beloved and electrifying showmen the ATP Tour has produced in the last two decades, has confirmed that his life after tennis will look nothing like the one he has led on court for the past 20-plus years. The Frenchman, who is set to retire at the end of this …
Alcaraz wins US Open first-round match after injury layoff
Carlos Alcaraz opened his US Open title defense in solid fashion Monday, defeating Russia's Roman Safiullin in straight sets in his highly anticipated first singles match in more than four months.A powerful baseliner with a prior win over Alcaraz, Safiullin kept close in all three sets, but never looked likely to threaten the Spaniard.
Judge says Pentagon's measures against Anthropic were 'illegal and baseless'
The government labeled the company a supply chain risk earlier this year
Judge says Pentagon's measures against Anthropic were 'illegal and baseless'
The government labeled the company a supply chain risk earlier this year
Shell is attracting bids for its U.S. chemicals plants that could fetch $8 billion
ExxonMobil, LyondellBasell, Apollo Global Management, and Kuwait Petroleum have submitted non-binding offers for Shell's U.S. assets
Virginia considers calls for mandatory school nurses as healthcare gaps persist
Legislators and experts are exploring updating the state’s Standards of Quality, the school funding baseline requirements for the state and local governments, to include mandatory full-time licensed nurse positions.
Top JPMorganChase exec warns regulatory proposal could squeeze credit for millions of small businesses
JPMorganChase warns Basel III Endgame capital rules could hurt Main Street by raising borrowing costs and reducing small business lending access.
Top JPMorgan Chase exec warns regulatory proposal could squeeze credit for millions of small businesses
JPMorgan Chase warns Basel III Endgame capital rules could hurt Main Street by raising borrowing costs and reducing small business lending access.
Tiny quantum engines reveal useful energy hiding in “waste heat”
A tiny machine made from just an atom and particles of light may sound impossibly simple, but it raises a surprisingly difficult question: what counts as heat, and what energy can still do useful work? University of Basel researchers have developed a theoretical framework that brings quantum physics and thermodynamics into better agreement for these microscopic “light engines.”
Barcelona’s friendly against FC Basel is nearly impossible to watch
This is disappointing
Selena Gomez Dismisses 'Baseless' Fraud Claims Over Mental Health Startup
The singer and actor vows to fight a $1.2 million lawsuit by Wondermind investors who claim she 'falsely represented' her involvement in the wellness venture. Her lawyer calls the allegations 'completely meritless.'
Explore Our Women’s Sports Report
Download the full report to learn what makes the women’s sports audience unique and how your brand can reach them.
Inside the decade-long push to make Las Vegas a sports town
Las Vegas is a certified sports town, with the NFL’s Raiders, WNBA’s Aces, F1, NHL’s Golden Knights and the MLB and NBA likely on the way.
MrBeast Sales Slow, Angel Stock Falls, MUBI and Neon Look for Hits, and Will an NBA Franchise Set a Sales Record?
(Welcome to the Entertainment Strategy Guy, a newsletter on the entertainment industry and business strategy. I write a weekly Streaming Ratings Report and a bi-weekly strategy column, along with occasional deep dives into other topics, like today’s article. Please subscribe.) You probably remember egg prices spiking a few years ago. This led to a lot of debate between the pro-market/pro-competition/antitrust/Neo-Brandesian crowd and the center/center-right/center-left/abundance/libertarian crowd over the cause, whether it was the avian flu outbreak or whether industry consolidation led to price fixing, using avian flu as a convenient cover. Basel Musharbash wrote a whole investigation at BIG looking at it. Well, two years later, we have our answer: it was price fixing. The Justice Department found proof, via emails, of price fixing. Indeed, just the threat of an investigation brought prices down two years ago: I bring this up because at least half a dozen of my favorite political writers and pundits, who were skeptical about price fixing at the time, haven’t written any mea culpas on this issue. It’s really dispiriting and makes it much, much harder to trust their work on other issues. If you take a hard stand on an issue, then you get it wrong, you need to explain how and why you got it wrong. Or update your priors. This is how epistemological bubbles are formed! All of which is to say, it’s time for another installment of “What I Got Right, What I Got Wrong”. Last issue, I mostly focused on the “wrong” side of the equation, so today’s is a bit more balanced, but with one big mea culpa (though I still stand by my original concerns). Let’s dive right in! Subscribe CORRECTION: UFC Didn’t Release Subscriber Data (But Seems to Have Bragged About It Internally) In my analysis of the viewership of UFC Freedom 250, I wrote... “Paramount has also released datecdotes bragging about their subscriber numbers. Earlier this year, they claimed that the first UFC fight on Paramount+ brought in 1 million subscribers.” Well, that datecdote did not come from a PR rep, but from great reporting from James Faris at Business Insider. The actual quote was this: “Paramount’s flagship streaming service generated about a million new subscribers on the day of its first-ever UFC event, Paramount product chief Dane Glasgow told employees in a town hall on Tuesday morning, three staffers who attended the meeting told Business Insider.“ This wasn’t the biggest goof in the world; even Faris emphasized to me that this is more a “clarification” than a correction. Still, it sounds like, internally, the folks at Paramount were very happy about this result—and I’m guessing they were happy with this news getting out—even though the gains didn’t last through the quarter, dropping to 700K adds. RIGHT: MrBeast’s Struggles Continue... I’m very skeptical about MrBeast’s burgeoning media “empire” and, in particular, its $5 billion valuation, as I wrote earlier this year and last year. Sure enough, according to Business Insider, regarding its “booming” consumer products division_:_ “US sales volume grew 13% year-over-year in 2025, following a 33% leap the previous year, according to a presentation deck dated May 2026....A February 2025 investor deck viewed by Business Insider showed Feastables’ net revenue more than doubled in 2024 from 2023, to $215 million, and was forecast to grow 74% in 2025, to $375 million.” [Emphasis mine.] If you told investors that you projected 74% growth but delivered 13%,1 that’s not ideal. But it’s not just MrBeast. Logan Paul and KSI’s Prime had sales drop in half for their energy drinks. It’s one thing to build a successful entertainment company and an entirely different thing to succeed in the heavily-consolidated-but-also-very-competitive consumer packaged goods space! CPGs are dominated by a few firms with very wide moats compared to YouTube, which has few barriers to entry. These businesses require different skill sets and have completely different business environments. WRONG: Does the NBA Have an $8 Billion Bidder? Almost as soon as I published my article about the NBA’s finances, casting skepticism on NBA expansion, I saw this Front Office Sports headline: One group “claims to have already raised $8 billion” for the Vegas expansion team. I mean, wow. That’s just $2 billion less than what the Los Angeles Lakers were valued at when they were sold just last year. And they’re arguably the most popular team in the NBA. (Side note: the new owner is now under an investigation for fraud in how he went about financing the Lakers and the Dodgers purchases.) So much for the “dried up” language I used in that article. (This is why I usually stick to my “Entertainment Nuance Guy” tone in articles...) I went off old information—from April, when I read/heard about the NBA downplaying expansion efforts—instead of doing a quick double-check for new news. That said...if you read the article, the market for Seattle is “slow”, with just one reported bidder. So I wasn’t entirely off. This goof has annoyed me since I first published it, mainly because it undercut that article’s real point, which I stand by: NBA franchises are really overvalued right now. Now, I get why NBA teams are overvalued. When it comes to owning sports teams, it’s basically never wrong to assume that some rich person somewhere will care more about owning a team (or just being part of a group since they’re so expensive) than the actual financials. There are only so many teams to own, and opportunities to buy in preferred markets don’t come along very often. This explains why the teams sell for such high multiples (usually above a 10 to 12.5 revenue multiple, which is very high). You only need one person to overpay! Beyond the ultra-wealthy luxury good explanation, I worry NBA teams are becoming speculative assets. More and more private equity investors are buying into NBA teams. But the value, in owning a team, comes from selling it later, not any meaningful relationship to revenue or profit growth. In other words, the newer NBA owners are increasingly expecting someone later to buy their inflated asset. That’s not sound economics; that’s the “greater fool” theory at work.2 At some point, actually making money has to matter. Especially since revenue growth is now below the growth of the stock market. (And this matters because the NBA doesn’t have easy levers to pull to drastically lower their costs—half of their revenue has to go to players and then you actually have to run a team, including coaching, training, travel, media expenses, and more—or raise revenues aside from one-off expansions in America and overseas.) I just don’t see how investors make their money back buying NBA teams, except for selling at a higher price later on down the line. But if revenue/profit growth continues to slow, at some point, someone will be left holding the bag. RIGHT: Baseball and Basketball are about equally popular. Related to the above, I wanted to cite two quick pieces of evidence for one of my other theories: that the NBA and MLB are about as popular as each other, but the NBA gets much, much better media coverage. One YouGov survey has that fans prefer basketball to baseball by 52% to 48%, which I’m guessing is within the margin of error for that poll. That’s virtually even, meaning they’re about the same popularity. Over on NBC, a recent Red Sox-Yankees game averaged 4 million viewers, higher than Sunday Night Basketball’s average of 3.4 million on the same network. A Yankees-Dodgers game averaged 3.9 million two weeks ago. Not every game averages this much, but it shows you that baseball isn’t far off from basketball, depending on the matchup. Altogether, the NBA is just slightly more popular than baseball, but NBC is paying MLB a fraction of what they pay the NBA, a point I’ve made before. RIGHT: Angel, MUBI and Neon Updates... I’ve long been skeptical about many of Hollywood’s buzziest startups/smaller indie studios. It’s not that they don’t make great movies—films like Anora and Everything Everywhere All at Once are amazing—or that they’re not (rightfully) pushing the majors to make better films, but...they get a ton of hype, and that hype often doesn’t live up to the reality or difficulty in scaling up. Starting with MUBI, well, The Wall Street Journal wrote a feature on MUBI’s struggles last April, best summarized by the subheader “After Sequoia valued Mubi at $1 billion, a left-wing revolt sent the indie film company into a tailspin”. MUBI lost $7.3 million on $200 million in revenue in Q4 2025, but reportedly hit a record 1.7 million subscribers at the end of Q1. As for Neon, Department M purchased a “significant stake” in the company earlier this year. How much is that “stake” worth? According to Deadline, “No financial figures for the Neon-Department M deal were disclosed.” So...that’s not great! If it was a big valuation, we’d hear about. Also, I found this curious: “Both Neon and Department M already have separate partnerships with Qatar. The Qatar Film Committee has a slate deal in place with the former, and a biopic production pact with the latter.” Color me skeptical about much Middle Eastern investment dollars. See: LIV Golf. Finally, as for Angel (formerly Angel Studios), their stock very, very slightly popped last month after Young Washington ($20 million budget) which made $45 million in the US, but their stock is still way down from its opening. Smaller Updates RIGHT: The WGA should focus on antitrust concerns! I’ve been on this for years. Post-WGA/AMTPT deal in 2023, I wrote an article for the Ankler calling on the WGA to focus and organize against industry consolidation. Sure enough, the WGA filed a lawsuit against the Paramount-WBD merger, and people are organizing around this issue now. I wish this energy had coalesced earlier (specifically to lobby Washington much earlier), but better late than never! WRONG: Actually, the chip shortage might be due to market manipulation. In my last “Long Reads for the Long Summer” article, I shared an article about how LLMs are causing memory chip prices to spike, which will make all electronics more expensive. Well, turns out, three RAM manufacturers control 90% of the market, and now a private lawsuit accuses them of price-fixing. So yeah. WRONG/UPDATE: The Odyssey will hold! Two weeks ago, I wrote, “As of this moment, we haven’t had a super-blockbuster, though The Odyssey, Spider-Man: Brand New Day and Avengers: Doomsday could all cross the $500 million mark. (At this point, The Odyssey would need a very strong hold to get there.)” Well, it’s safe to say it’s going to have that hold! It only dropped 30% in weekend two and IMAX screenings are sold out all month. Spider-Man: Brand New Day is also off to a great start. RIGHT: People are rushing to get shows out yearly. I’ve long been bullish on yearly, weekly-released shows, and Leslie Goldberg at The Ankler wrote about how many people in town are now listening. WRONG: Mark Carney reverses Canada’s stance on homegrown production. After writing about Canada installing domestic production quotas, Canada has backed off. WRONG: I don’t trust Spider-Man: Brand New Day’s online “views”. When I originally wrote about the latest Spider-Man film’s record-breaking trailer, I thought this was just YouTube views, but after watching this Corridor Crew video and doing more research, well, it’s all social views of the trailer globally. Really? That stat includes TikTok, Reels and YouTube Shorts views. Basically, I don’t trust it anymore. (This film will probably set record, but I trust the pre-sales data more.) RIGHT: Celebrity Production companies are still struggling. I’ve written about this for years, but the company that I didn’t track was Kevin Hart’s HartBeat, which was valued at $650 million but is struggling now and might just end soon, especially after Kevin Hart pulled his endorsement deals. And Bad Robot is moving to NYC. CORRECTION: SpongeBob is a “second run” title. In my 2025 recap article, I had _SpongeBob SquarePants_labelled as a “library” title, when it is still a “second run” show, with new episodes still coming out on Nickelodeon. CORRECTION: In my last “What I Got Right, What I Got Wrong” update, I offered a correction from Doug Creutz and misspelled TD Cowen. Thanks, Grammarly! 1 Though he could have increased prices to make up for the gap, he couldn’t have increased them that much. 2 There’s a theory, which I think that Daryl Morey created and Bill Simmons popularized, that, to be successful building your team, you only need a handful of bad GMs in the league to take advantage of. To be clear, Daryl Morey’s theory was also for signing bad contracts, that you could always get out of bad contracts. Basically, there’s always someone who will make a bad trade. It’s not the same exact thing, but similar enough to the Greater Fool Theory that I really like it.
House ethics panel investigating Gooden, who rejected the allegations
Gooden said allegations related to spending are ‘baseless.’
Sen. Lindsey Graham's sudden death spurs false claims
Sen. Lindsey Graham's unexpected death on Saturday night fueled baseless speculation on social media that his passing was the result of a criminal action. Here's a closer look at the facts.
They're Building Half A Bridge to Save Science...
Today’s post is about science: what you can do now to protect it from unprecedented threats, and how your action is the first half of a long-term program that Congress must take up next. Let me start with describing the embryonic state of science advocacy. There are many well-meaning organizations: the American Association for the Advancement of Science, Research!America, and subject-specific organizations such as the Society for Neuroscience and the American Physical Society. Their general approach is to emphasize the common good: the remarkable value that science and innovation have provided to all Americans. However, these organizations are offering an incomplete solution. Before I start, a preliminary: If you already know about the threat, go ahead and weigh in with OMB at this link. Ideas about what to say come from Research!America and the American Physical Society. You have until Monday, July 13 at 11:59pm Eastern. Then come back here and read the rest of this post. Finally, get in touch with whoever is likely to be your next member of Congress in 2027. The threat: An end to merit-based science… The onslaught of attacks on science is accelerating. It’s not just cuts to grants, which was the last battle. That battle is not done yet, but there is a new and worse threat. Now comes a proposed public rule from the Office of Management and Budget. The rule would take decision-making power over grants away from experts and put it into the hands of political appointees. Former NIH official Elizabeth Ginexi wrote an excellent summary of what OMB is trying to do, which is codify the executive branch’s ability to award or terminate grants, for arbitrary cause or no stated cause. Since World War II, a necessary step in scientific grant-making has been consideration of scientific merit. This approach has made America into a dominant scientific superpower, the envy of the world. Of course, political considerations do come into the mix. The proposed rule tilts the balance dramatically away from merit. …but why? Many people have difficulty comprehending why Russell Vought and OMB would want to damage science in this way. If you’re a working scientist, you are likely to be particularly uncomprehending. Who would hurt an activity so self-evidently good? The answer is: this is part of the program to replace the rule of law with a personalist regime. OMB Director Russell Vought, a Christian nationalist, would like to implement what legal scholars call a “unitary executive,” and what I would call a dictatorship. Or if you like, a personalist authoritarian regime. I am a working scientist. This rule will debilitate my research. It may chill the speech of scientists for speaking freely in the public sphere. It is a form of personalizing power and decision-making within the executive branch. In short, it sacrifices science in the service of tyranny. To Russell Vought, a decrease in the quality of science might be an acceptable cost. He might even consider it an added benefit to weaken a pillar of soft U.S. power based on knowledge and skepticism. The professional societies are building half a bridge The nonprofits are asking you to make public comment. Public comment is like building the first half of a bridge. Before implementation, OMB is compelled by law to have a period of public comment. Rules have to be consistent with the law and the Constitution. They express the federal government’s stance toward how those laws will be implemented. When a presidential administration works in good faith, these public comments are used to inform their eventual action, and sometimes mitigate it. The comment period reflects our First Amendment right to petition our government. The executive branch is then supposed to take those comments into account - completing the bridge between us, the people, and our government. Project 2025 seeks to take away the second half of the bridge. We have to replace it. The first half of the bridge Organizations like AAAS, Research!America, and SfN are partly right. You should certainly comment. Their core model is earnest: if enough of us make noise, then we will achieve change. This idealistic point of view is offered in good faith. That point of view is also currently bullshit. The fundamental reason is that the federal rulemaking process has two parts: There has to be a period of public comment. OMB makes the final rule…but is free to disregard public comment. Oops. That’s a loophole. However, not all is lost. Congress can complete the bridge Congress has enormous power. That power suggests a roadmap for how to proceed. This lawless administration is still going to issue that rule. Nonetheless, your comments still play an essential role. Here are two steps you can take. Step 1: Send your comment to your Representative and Senator. Whether they are Democrats or Republicans, many of them support science. Whatever you wrote as public comment in the Federal Register, also send it to your member of Congress. You can include information about jobs in your district or state by looking at Joshua Weitz’s ScienceMatters interactive map. Call the U.S. Capitol switchboard at (202) 224-3121 and give your town, zip code, or state, then deliver your message. Step 2: Politicize the hell out of science. You heard me, politicize it as an issue. We hear a lot from the old scientific establishment about science being nonpartisan. This is obsolete, sentimental thinking. Under normal conditions, it was true. But we can no longer make that assumption. We can keep on pretending, but we’re putting science at risk. There is a way forward. Rules are subordinate to the law…and Congress makes laws. Therefore Congress can step in. In this fall’s election, it is likely that Democrats will retake the House of Representatives. That puts at least one chamber of Congress in the hands of Democrats. They are far more inclined than the current Congress to investigate and legislate. This opens a path to undoing the damage. It also creates a permission structure for pro-science Republicans to find a way back from the brink. We have to encourage them. In the House, every district is up every two years. In your district, contact both of the major party candidates. If your state has a Senate race, contact those candidates too. Specifically mention the proposed OMB rule. Ask for two things: For this year’s appropriations process in August and September, ask that amendments be added to stop the implementation of the new rule. Those amendments should support peer-reviewed science and stop political interference. Looking ahead to next year, ask the Congressman/Congresswoman/Senator to work to hold hearings in 2027 on the importance and benefits of peer-reviewed science for health and prosperity. What hearings and amendments will accomplish Congress can use its investigatory power to summon heads of agencies and the director of the OMB. That will allow fact-finding regarding the consequences of this rule. Congress can override the rule through the appropriations process. They can write into statute a requirement to follow peer review and best scientific practices. In this way, Congress can stop the attacks on science. Postscript: Moving forward Now I have a request for you, to help me. As many of you know, I ran for Congress in order to save democracy and science. Although I did not win, that battle continues by other means. I met and heard from thousands of supporters both here in New Jersey and nationally. I learned that pro-science, pro-democracy activists have the potential to become a much greater force. In the weeks and months ahead, I will be building ways to harness that energy. I will be announcing those plans soon. In the meantime, an important step is to wind down the Congressional campaign, which has a few remaining bills to pay. Please help me retire that last debt.
Roma, Basel reach agreement over Calafiori’s resale clause
Riccardo Calafiori is one of the players who emerged from Roma’s youth system and has managed to establish himself on the big stage. The former Giallorossi defender won the Premier League with Arsen...
F.B.I. Assigns Scores of Analysts to Examine Election Records in Georgia
The expansion of an investigation into Fulton County’s election office reflects President Trump’s desire to prove his baseless claims that the 2020 election in Georgia was rigged.
The Lasting Consequences of Bad Climate Science
The retirement of flawed baseline emissions scenarios will initiate changes across energy policy, regulation, and law.
Heat forces yodelers at annual Swiss festival to sing in fountains
Yodelers at a festival in Basel, Switzerland, turned city fountains into rehearsal spaces during Europe’s June heat wave.
Efzimfotase alfa demonstrated improvements in bone health in treatment-naïve pediatric patients with hypophosphatasia achieving median difference of 1.67 vs placebo in RGI-C Score at week 25 in MULBERRY Phase III trial
Positive results from the MULBERRY Phase III trial showed that efzimfotase alfa (ALXN1850), an investigational enzyme replacement therapy, demonstrated a statistically significant and clinically meaningful improvement in bone health in children (2 to
New Media Event Jupiter Festival Planned from Art Basel Company
MCH Group, in which the Fox scion is the largest shareholder, is launching an event titled the Jupiter Festival with an eye toward becoming a stop on the thought leader circuit.
UFL viewership stays level from Year 2 to Year 3
It seems as if the UFL, the upstart spring football league that just completed its third season, is finding somewhat of a baseline when it comes to viewership. The league averaged 674,000 viewers in its third season across ABC, ESPN, ESPN2, Fox, FS1, and NFL Network, up 1% versus average viewership in its second season…
Business News: Basel Returns To Watch Calendar With April Basilia Watch & Jewelry Show
Baselworld is defunct, but Basilia aims to fill a gap for accessible watch and jewelry brands, but no details yet on who might participate among 400 exhibitors.
Newsom says DOJ conducting baseless investigation of him and his wife at Trump's direction
Newsom said that in recent days, "federal agents have knocked on the doors of family friends and former employees," and have been "demanding records" in a quest to find any kind of wrongdoing by him or his wife.
Is SMIC N+3’s Metal Pitch Smaller than Intel 18A’s?
Almost four years ago, we published that SMIC had started shipping 7 nm (N+1) chips. Now, SMIC is shipping its third-generation 7 nm (N+3) process in Huawei’s Kirin 9030, with a minimum metal pitch of 32.5 nm, about 10% tighter than the 36 nm minimum metal pitch shipping in Intel’s latest Panther Lake CPUs on 18A. The headline is true, but incomplete cherry picked metric. N+3 reaches the density of TSMC N6 through aggressive DUV multi-patterning and design-technology co-optimization (DTCO), but it pays for that in complexity, efficiency and process control. We found this and more in our reverse engineering and teardown where we cover SMIC’s N3 process technology, Huawei’s packaging, memory, architecture, and more. SemiAnalysis has been building a state-of-the-art teardown lab in Oregon capable of analyzing the world’s most advanced and important chips over the last year and half. We have already generated revenue on advanced datacenter chip teardowns including our recent reverse engineering of a major TSMC customer’s COUPE CPO optical engine + EIC 3D stack. This is the first public report from the SemiAnalysis Teardown Engineering & Evaluation Lab, or STEEL for short. The lab is aggressively scaling up and out and we’re excited to announce it publicly. This is a bit of inconvenient timing for TechInsights as they are private equity owned and currently being sold while having enjoyed virtually no credible competition for decades. This has led to TechInsights underinvesting in CAPEX. SemiAnalysis exceeds TechInsights in revenue despite no venture or private equity ownership and being founded only 6 years ago. Because we have no external investors and are founder led, we move faster, build faster, and we can release client chip teardowns for free regularly, while focusing on datacenter for our major clients. Here’s the first public image from our lab, the HiSilicon Kirin 9030 Pro SoC: This report will detail our teardown of the Kirin 9030 and our findings on SMIC’s N+3 process, the most advanced in China. For comparison, we’ll show our teardown of the MediaTek Helio G99, made on TSMC N6. Through this comparison, we can look at the effect of export controls – SMIC N+3 and TSMC N6 are comparable nodes, but one is heavily export-controlled, the other free to use the West’s most advanced equipment. Here we see both China’s progress and constraints. SMIC N+3 reaches TSMC N6-class logic density, but it requires far more aggressive DUV multi-patterning, so it does not match N6 on process maturity or cost. The Kirin 9030 Pro performs similarly to three-year-old Android flagships, and trails far behind the current flagship SoCs from Apple, Qualcomm, MediaTek, and Samsung. The efficiency gap is even wider. Export controls have not stopped Huawei and SMIC from shipping advanced silicon, but they have forced a different path. Without EUV, SMIC is leaning harder on DUV multi-patterning, DTCO, and increasingly complex integration. The roadmap continues forward through tighter design rules and backside power, but each step adds cost and process risk. Huawei’s τ scaling and LogicFolding show another path: stacking active logic and recovering density through advanced packaging and system-technology co-optimization (STCO). To understand the Kirin 9030, we must first understand Huawei’s SoC history. HiSilicon is Huawei’s chip design arm, responsible for the Kirin smartphone SoCs, Kunpeng server CPUs, Ascend AI accelerators, and switch/router networking silicon. Before export controls, Huawei was TSMC’s largest customer – the only customer on TSMC’s first EUV node, N7+, and among the first on N5, alongside Apple. That ended in late 2020. Huawei switched to Qualcomm SoCs in its flagship smartphones, though export controls limited them to 4G-only variants. In late 2023, Huawei returned to in-house silicon with the Kirin 9000s, a successor to the Kirin 9000, fabricated on SMIC N+2 instead of TSMC N5. In the following years, they released the Kirin 9010 and 9020 on the same N+2 process. These chips used Huawei’s in-house TaiShan CPU cores and Maleoon GPU. We have not torn down a Kirin 9020 ourselves, so the predecessor die shot is from Kurnal. The die shots show how Huawei spent its silicon budget: which functional blocks are where, and how their areas compare to the predecessor. First, a quick guide to the major blocks on the die. The total die area is nearly identical, but the 9030 uses that area more aggressively. A denser process lets Huawei fit an extra middle CPU core, more GPU and NPU cores, and larger caches into the same footprint. In contrast, the Helio G99 is a much smaller, low-cost SoC, built for budget smartphones rather than a flagship device. While the Kirin 9030 is ~140 mm², the G99 is only ~29 mm², roughly one-fifth the area. The underlying TSMC process technology, however, is directly comparable as a baseline for analyzing SMIC’s. The Kirin 9030 is an evolutionary refresh, not a clean-sheet design. Its CPU, GPU and NPU cores carry over the 9020’s families, and the gains come from three levers: the SMIC N+2-to-N+3 process step, DTCO and floorplan work, and incremental microarchitecture. Area is where the first two show up, and the 9030 scales well here. Performance and efficiency are the harder test. Huawei’s design holds up better than its node would suggest, but the chip still trails, both because N+3 sits behind the leading-edge nodes and because its cores, while competent, remain a few generations behind the newest designs. The new prime core is an incremental update. The main changes are a 10% frequency increase from 2.5 GHz to 2.75 GHz and a doubling in the L2 cache from 1 MiB to 2 MiB. Despite the increased cache, the core size decreased by 7.6%. Excluding the private L2 cache, the core size decreased by 21%. This is a large reduction for an incremental node. Compared with the TaiShan New V120 core in the Kirin 9020, the Kirin 9030’s middle core is almost unchanged architecturally, yet each core shrinks by ~22%. Most of that comes from the move from N+2 to N+3, with layout likely accounting for the rest. Visually, the most noticeable change is the increase from 3 middle cores to 4. There is also a 20% increase in the shared L3 cache of the big cluster. This helps to improve multi-core performance without sacrificing much in terms of area. Even with each core shrinking, the big CPU cluster’s total area is essentially unchanged. The per-core savings went back into an additional middle core and larger caches. The tiny cores shrank less than the prime core (excluding its L2 cache) and less than the middle cores. This is likely because fixed overhead is a larger share on a small core. We cannot resolve any architectural changes from the die shot alone, but the per-clock and efficiency gains shown below point to more than pure process and layout scaling. The area reduction was offset by a doubling of the shared L2 cache from 2 MiB to 4 MiB, leaving the total tiny CPU cluster area slightly larger. Area is the easiest improvement to see from a die shot, but it is only one part of PPA (power, performance, area). For modern logic, power and performance matter just as much, and often more. Since Dennard scaling broke down in the mid-2000s, voltage and frequency have not scaled in step with transistor dimensions, so each node has had to fight harder for gains in performance and efficiency. The starkest comparison is not Kirin 9020 versus Kirin 9030 Pro. Apple’s efficiency cores run circles around Huawei’s prime core. Apple’s low-power core delivers 20% higher integer performance while drawing only 1 W, compared with 4.5 W for Huawei’s prime core. N+3 matches TSMC N6, but N6 is several generations old. Apple and Qualcomm build on N4 and N3P, which are denser and sit on a better voltage-frequency curve, giving them a larger transistor budget and more performance per watt. The 9030’s own cores did improve. The middle and tiny cores gained 17% and 14% in per-clock integer performance over the 9020, with floating-point flat on the middle core and up 11% on the tiny. The tiny core improves cleanly, with performance rising while power falls and efficiency increases by 45% in integer and 24% in floating point. The middle core is mixed: integer performance rises but power rises faster, cutting integer efficiency by 7%, while lower power lifts floating-point efficiency 16%. Per-clock gains at the same or lower frequency are microarchitectural, so the cores are tuned, not just shrunk. Both also failed to hold their rated maximum frequencies, pointing to thermal, power, or stability limits. Per clock, the middle core sits around Arm Cortex-A720 and the tiny core near the Cortex-A520; absolute performance trails because Huawei clocks them much lower. The prime core is roughly Cortex-X2 class per clock, a 2021 design. Apple’s 2020 M1 Firestorm core is still 35% higher per clock and 57% faster in absolute integer performance at a similar 4.5 W. The current leading edge is further ahead again: the Apple M5 P-core is 60% higher per clock and 2.7× faster, the Arm C1 Ultra 45% higher and 2× faster. Matching older high-end cores per clock is a genuine design achievement. What Huawei cannot match is the voltage-frequency curve and transistor budget of leading-edge nodes, which let Apple, Qualcomm and others spend more transistors in the same area on wider cores, larger caches and deeper buffers while running at lower voltage. Huawei’s LogicFolding roadmap is one answer, stacking active logic to recover density and shorten signal paths. We return to it later. The GPU compute units (CUs) changed more visibly than the CPU cores, moving to a more rectangular layout for both the arithmetic logic unit (ALU) clusters and the CU overall. Even with ray-tracing support added, a CU shrank ~28%. However, that shrink is offset by the increase from 4 to 6 CUs and the area outside the CUs grew 33%. Overall, the GPU cluster is larger by ~10%. The GPU is where Huawei makes its biggest gains. The Maleoon 935 is not competitive with current flagships, but it is a large step up from the 920 and reaches older-flagship territory. In 3DMark it is 70% faster in Wild Life Extreme (WLE) and 79% faster in Steel Nomad Light (SNL) than the 920; with 11% higher clocks and 50% more CUs, the ~67% theoretical uplift roughly matches WLE and is beaten by SNL. It edges ahead of the Snapdragon 8+ Gen 1 in WLE and SNL, and the Dimensity 9200 and Apple A16 in WLE, but stays far behind newer parts: the Snapdragon 8 Elite Gen 5 and Dimensity 9500 are ~2.4–2.6× faster in WLE and ~3.2× faster in SNL. The Maleoon 935 is Huawei’s first GPU with hardware-accelerated ray tracing; there it lands slightly ahead of the Exynos 2200, and on par with the Apple A16, with current flagships up to 3.7× faster. The Neural Processing Unit (NPU) saw the largest structural changes of any block, moving from a Lite and a Tiny core in the Kirin 9020 to a Lite and two Tiny cores in the Kirin 9030. Both core types also show significant layout changes. This is a reversal in Huawei’s NPU design. The Kirin 9000 5G, its last flagship chip on TSMC N5, used two Lite and one Tiny core. The series of SoCs on SMIC N+2 moved to one Lite and one Tiny core, likely for area savings. With the Kirin 9030, Huawei has shifted back toward a larger multi-core NPU cluster, but with the additional area going to a Tiny core rather than a Lite core. We’re diving deep into the most advanced datacenter and AI hardware hitting the market. To learn more about what’s in the pipeline or to commission a custom teardown, contact sales@semianalysis.com. Interested in joining us on this ride and think you can be a difference maker? Check out our Careers page. Before diving into the process stack, the package and memory are worth separating from the SoC itself. The Pro variant of the Kirin 9030 carries 12 GB of Samsung DRAM, with two stacks of four dies each. The dies were identified as the K4L2E165YD, a 12 Gb LPDDR5X-9600 device fabricated on Samsung’s 1a node, the fourth generation of its 10 nm-class DRAM after 1x, 1y and 1z. 1a has shipped in volume since 2022, so this is current memory rather than older-node inventory. The 16 GB Pro Max variants we obtained were found with both CXMT and Samsung packages. The CXMT package is marked CXDD7JEDM, with two stacks of four dies, packaged in week 45 of 2025. The inferred die dimensions from X-ray computed tomography (CT) are consistent with a known density of ~0.3 Gib/mm² for the CXMT G4 process, roughly equivalent to other manufacturers’ 1z processes. The Kirin 9030 uses a typical integrated package-on-package (iPoP) stack: multiple DRAM dies in a memory package sit above an organic redistribution layer (RDL) interposer, which sits above the SoC and package substrate. The full package is then mounted to the printed circuit board (PCB) through ball-grid array (BGA) solder bumps. The memory package substrate is a thin bismaleimide-triazine (BT) laminate carrying the LPDDR5X stack. The organic RDL interposer over the SoC routes the PoP signals around the die and carries possible dummy thermal copper pillars. The package substrate, a thicker Ajinomoto Build-up Film (ABF) build-up over a BT core, fans the flip-chip bumps out to BGA pitch and embeds the power planes. The whole stack is organic. The only silicon is the SoC and the LPDDR5X dies; there is no silicon interposer. Keeping it all-organic brings the package’s coefficient of thermal expansion (CTE) close to the PCB’s, reducing board-level warpage, and avoids the cost of a silicon interposer the SoC’s bandwidth does not need. In an iPoP stack, the memory package connects to the organic RDL interposer through an array of solder bumps. Underfill fills the gap around those bumps, adding stiffness and protecting the joints from mechanical stress. The Pro and Pro Max variants differ here, which we cover behind the paywall. The die shot and architecture tell us how Huawei allocated its silicon budget. The process tells us what SMIC can manufacture. We use the Helio G99 as the process reference for TSMC N6. Both SMIC N+3 and TSMC N6 are evolutions of previous 7 nm-class nodes. We used targeted TEM cross-sections through logic and memory regions, imaged in both fin-cut and gate-cut directions. Each cross-section caption gives its horizontal field width (HFW), the real width of the imaged area. We start at the transistor fins, then move up through standard cells, local interconnect, and SRAM. SMIC has not overtaken Intel or TSMC. It uses aggressive DUV scaling and DTCO to reach N6-class density, but that density doesn’t translate into comparable performance and efficiency, for two reasons: the node gap to leading-edge nodes, and Huawei’s core designs. Fin Profile One of the most important knobs in a FinFET process is the fin profile: the shape of an individual fin and the channel where current passes from source to drain. The ideal fin is tall, narrow, and nearly vertical. A taller fin increases effective channel width, while a narrower fin improves electrostatic control by thinning the body the gate must control. Push either too far, and the process pays for it: weaker drive current, fragile fins, taper, footing and line-edge variation that hit yield and device variability. The Intel 22 nm, 14 nm, and 10 nm fin cross-sections show how FinFET nodes have improved over time. 22 nm fins were a first-generation structure, relatively short, wide and strongly tapered. The shape limits current density and reduces gate control uniformity across the height of the fin. At 14 nm and 10 nm, Intel pushed the fins taller and narrower while also making the sidewalls more vertical. Rather than shrinking the device, these changes increase the effective channel width per fin and improve electrostatic control. The trade-off is that taller fins at tighter pitches are much more difficult to manufacture. Now, let’s compare the Helio G99 on TSMC N6 with the Kirin 9030 on SMIC N+3. Both processes are in the same class, with fin pitch of 30-32 nm on N+3 and 34 nm in our N6 cross-section. The pitch for N6 is especially interesting as N7’s HD library is generally listed with a 33 nm fin pitch, and N6 did not shrink pitches directly. Its density gains came from DTCO instead of tighter pitches. The 34 nm pitch was stable across our sampled region and serves more as a comparison against the SMIC N+3 we have not investigated further. Pinning down N+3’s fin patterning scheme takes more than one core unit. The CPU cores show a dense ~32 nm pitch, with the pitch between N-P fin pairs alternating between 78 and 88 nm. Logic alone may be consistent with dual-pitch mandrels of 120 and 110 nm, but this is a complex and unusual approach. Combining the pitch from the the 8T SRAM, which has more complex repeat unit, with the CPU core sequence allows us to reverse engineer the patterning steps with more confidence. As both the logic and SRAM should share the same base grid, a single CD mandrel lithography pattern with 128 nm pitch undergoing SAQP produces a die-wide ~32 nm grid (128 nm/4), which supports the pitch sequencing seen in both logic and SRAM cells. In the sampled cross-sections, N+3 shows a taller, narrower, higher-aspect-ratio fin than N6. The measured fin aspect ratio is ~9.5:1 on N+3 versus 7.8:1 on N6. N+3 also shows less top rounding, with an estimated radius of ~2 nm, compared with 2.8 nm on N6. Even though the fin widths differ, the ratio of top rounding to fin width tells the same story, with N+3 at 0.37 and N6 at 0.44. In a geometric sense, lower is better; a perfectly rectangular fin would have no top-rounding penalty. These are single-digit-nanometer features measured from a handful of cuts, so treat the absolute numbers as approximate. The important result is the relative gap: N+3’s fins are consistently taller, narrower and less rounded than N6’s. We’re diving deep into the most advanced datacenter and AI hardware hitting the market. To learn more about what’s in the pipeline or to commission a custom teardown, contact sales@semianalysis.com. Interested in joining us on this ride and think you can be a difference maker? Check out our Careers page. Standard Cell A standard cell is the basic building block of chip layout: a fixed-height row pairing one NMOS and one PMOS transistor that share a gate, tiled in a grid to build logic blocks. The key dimensions are contacted gate pitch (CGP), cell height (CH), fin count, and the lower-metal routing grid. To measure density, we use the Bohr metric: a weighted average of NAND2 gate area (60%) and scan flip-flop area (40%). This represents a realistic mix of combinational and sequential logic. This metric has its limitations, especially for complex cell layouts like TSMC’s FinFLEX, which alternates cells with different fin counts. Even so, it is the best metric for a pure process-level comparison. Another important measurement is the fin pitch; it refers to the distance between two fins of the same transistor. In a FinFET process, multiple fins are used in each transistor to increase the drive current and thus performance. TSMC N6 ships both a high-density (HD) library with 2 PMOS and 2 NMOS fins per cell, and a high-performance (HP) library with 3 of each. More fins under the shared gate mean more effective channel width. HP cells switch harder at the cost of area. Designers mix the two on a die, primarily spending HP cells on timing-critical paths, and matching their PPA targets. In the Cortex-A55 core of the Helio G99, we found a cell height of 240 nm for the HD cell. MediaTek has used HD cells in the G99 to minimize die size and thus cost. As an SoC for budget smartphones at ~$100, this is essential. By contrast, we found only one library in the Kirin 9030, with 2 NMOS and 2 PMOS fins. This suggests a narrower library strategy than TSMC N6, where both HD and HP libraries are widely used. This likely reflects the smaller customer base and the more constrained domestic design and electronic design automation (EDA) ecosystem. In all three CPU cores of the Kirin 9030, we found cell heights of 228 nm, 5% smaller than on N6. This is also a reduction of 9.5% over SMIC N+2’s cell height of 252 nm. SMIC N+3 and TSMC N6’s HD library both feature a CGP of 57 nm. For SMIC, this is a 9.5% shrink over N+2. In the past, CGP and cell height alone may have been enough to compare transistor density. Now, however, we must consider scaling boosters and DTCO as well. SMIC’s density gain does not come from EUV. It comes from using every available DTCO booster aggressively. First is fin depopulation: reducing the number of NMOS and PMOS fins in each cell. The first FinFET nodes started with 3 or 4 fins for each transistor. SMIC N+3 and TSMC N6 HD both use only 2 fins per transistor, trading drive strength for density. Next is contact over active gate (COAG). By landing the gate contact directly over the active gate, instead of out over the isolation region, the cell height drops. N+3 integrates COAG while N6 does not. Our N+3 gate-cut cross-sections indicate COAG, with the gate contact sitting over the active region, while N6 shows an off-gate contact. Last is single diffusion break (SDB). Diffusion breaks are inserted between cells in the same row to provide electrical isolation, but they also introduce local layout effects (LLE), layout-dependent shifts in electrical characteristics. In the past, a double diffusion break was used, consuming the space of two CGPs. SMIC N+3 and TSMC N6 instead use SDB, saving area but increasing LLE sensitivity. This must be controlled at the process level and accurately modeled in the process design kit (PDK) so that EDA tools can account for it. Overall, SMIC N+3 has a transistor density of 113.4 MTr/mm², slightly above TSMC N6 at 107.7 MTr/mm². Even without EUV, SMIC has achieved density beyond TSMC’s mature N6 node which utilizes EUV. Metal Stack The smallest critical dimension in the teardown is M0; SMIC N+3 uses a 32.5 nm local metal pitch. That is smaller than the 36 nm M0 pitch on Intel 18A in Panther Lake. However, this does not mean that SMIC has a better process than Intel 18A or TSMC N3P. M0 is a local intra-cell routing layer. Its usefulness depends on the full interconnect stack: M1 and M2 pitch, track count, via and line resistance, design rules, mask count, overlay control, and routing flexibility. The 32.5 nm M0 is consistent with self-aligned quadruple patterning (SAQP), whose four-population line-width loading we read coarsely as alternating widths of 21.5 to 24 nm; M1 and M2, at 38 and 40 nm, are consistent with self-aligned double patterning (SADP), a single A/B split. On TSMC N6, M0, M2, and M3 sit at a relaxed ~40 nm and are consistent with SADP-class double patterning, with no need for quadruple patterning. That said, we measure M2 for example at ~43 nm, likely inflated by sparse routing. We do not assign any specific layer to EUV from our cross-sections; the distinction we can draw is double versus quadruple patterning, not lithography wavelength. Transistor-level density in the front-end-of-line (FEOL) sets an upper bound, but the design is ultimately limited by what the interconnect stack can route. The lowest metals are the most important for standard-cell density, but the semi-global and global layers determine how usable that density is at the block and chip level. Two axes are commonly used for chip cross-sections: the fin-cut and the gate-cut. The micrograph above is a fin-cut and shows metals 0 through 3. This axis lets us see the even-numbered metals, with M0 right above the fins. There are two kinds of M0 lines. The first are the power rails; these are wide wires for the VDD and VSS running horizontally at the top and bottom edges of each standard cell. The wide wires measure 55 nm across, more than double the other M0 lines. Their width minimizes resistance and reduces IR drop. The second kind are intra-cell wires, short segments within the cell that connect terminals to M1. These have alternating widths between 21.5 and 24 nm. The M0 pitch is 32.5 nm, a 19% reduction versus N+2 and N6. At this pitch, DUV patterning requires more aggressive multi-patterning, increasing mask count, overlay sensitivity, process complexity and cost. M0 is below what a single DUV-defined spacer (SADP) can resolve, so SMIC cascades a second spacer step (SAQP). The cross-section reflects the cost: the M0 trenches are visibly more re-entrant (narrower at the bottom than the top) than M1 or M2 on the same chip and carry a bright barrier-rich foot where the trench meets the etch-stop layer. That shape is partly the intended damascene profile, as a slightly narrow bottom helps void-free copper fill, but its magnitude at M0 is driven by the tight pitch and the higher trench aspect ratio. Intel 18A supports an M0 pitch of 32 nm, although Panther Lake has only shipped with a looser 36 nm pitch. This is due to Intel’s heavy usage of HP libraries. Among leading-edge nodes, 18A has the loosest M0 pitches due to PowerVia. With power routing moving to the backside, congestion is reduced, and the entire front-side metal stack can be used for signal routing. M2 is the first true inter-cell routing layer. It runs horizontally like M0 but spans across multiple cells to carry block-level signals. The M2 pitch sets the cell’s track height – the number of M2 tracks that fit between the VDD and VSS rails, defining what the library calls a 6-track or 7.5-track cell. This layer is the most important, limiting the routing of entire blocks. SMIC N+3 features a 5.7-track cell. The M2 pitch is 40 nm, a 5% decrease over N+2 and the same as N6. This shrink keeps the pitch at the edge of what is possible with double patterning. Future nodes will need to increase the number of masks for M2 as reducing the number of tracks is much harder due to the limitations in routing. The micrograph above is in the perpendicular direction, the gate-cut, and shows metals 0 through 4. This allows us to see and measure the odd-numbered vertical metal layers. The M1 pitch is 38 nm, 9.5% less than N+2 and 33% less than N6. The M1-to-gate ratio matters because it sets local routing flexibility. N+2 and N+3 use a 3:2 ratio, while N6 uses a 1:1 ratio, explaining the huge differences in M1 pitch. The more M1 lines there are compared with the gates, the more flexibility there is for power and signal crossing within the cell. Routing flexibility enables more complex and better cells. Clean fractional ratios are also preferred as a grid is periodic and improves layouts. The 3:2 ratio gives SMIC more local routing flexibility than a strict 1:1 grid, but it also complicates layout and patterning. This is a DTCO choice, with SMIC increasing process complexity to recover density and routability without EUV. This 3:2 ratio is not very popular in the leading-edge nodes. TSMC has only used it on N7+, the N5 family, and the short-lived N3(B). They have switched back to a 1:1 ratio for N3E. Intel has only used it on the 10 nm/Intel 7 family, with Intel 4, 3 and 18A all using a 1:1 ratio. Samsung is the only one still using a 3:2 ratio at the leading edge, using it in the SF4 and SF3 families. It remains to be seen if SMIC will remain at a 3:2 ratio or move to a 1:1 ratio with its future nodes. The industry is still actively exploring these local-routing ratios. At VLSI 2026, imec will be presenting work on even higher ratios, including a 2:1 scheme that can reduce area by up to 14%. We will be covering the conference in a future newsletter article. Subscribe We’re diving deep into the most advanced datacenter and AI hardware hitting the market. To learn more about what’s in the pipeline or to commission a custom teardown, contact sales@semianalysis.com. Interested in joining us on this ride and think you can be a difference maker? Check out our Careers page. The final local interconnect layer for N+3 is M3, with a pitch of 44 nm. The M3 pitch is the same as on N+2 and 10% larger than on N6. The semi-global layers carry the majority of block-level signal routing. They have a coarser pitch than the lower local layers. On leading-edge nodes, they are designed to sit at the limit of DUV single patterning. M4 through M11 pitches were found divided between 80–82 nm (M4–M6), 128 nm (M7–M10), and 148 nm (M11). Given limited sampling, it is possible these are divided further in dense routing areas. At the top are two giant metal layers, M12 and M13. These have kept the same pitches as N+2 at 1920 nm and 4600 nm respectively. While the lower layers’ pitches are generally fixed by the process and library, the upper layers vary much more in pitch and count, depending on the design. Even two smartphone SoCs on the same process can have wildly different metal stacks. The Helio G99 carries fewer routing layers, reaching coarse metal pitches of 850 nm by M9, while the larger and higher-performance Kirin 9030 keeps fine pitches until M11. SRAM At the leading edge, SRAM is much more difficult to scale than logic. TSMC’s latest nodes have seen little to no bitcell scaling, while logic still has more DTCO levers to pull. While looking for other logic libraries in the GPU compute units, we stumbled upon the SRAM. The most common type of SRAM has 6 transistors (6T), but this cell had 8 transistors (8T) instead. 8T SRAM adds two transistors to form a dedicated read port. Unlike a 6T cell, where reading disturbs the storage, the decoupled read port removes read-disturb, improving read stability and letting the cell be pushed harder for performance. At first glance, the cut looked like an unusual logic library, with each cell row having 3 fins of one polarity and 5 fins of another. The rows also alternated in orientation. Energy-dispersive X-ray spectroscopy (EDS) resolved our confusion. The cut had not landed on the GPU logic, but on the SRAM macro beside it. The unusual fin pattern was due to the SRAM library. We return to EDS in the process flow analysis behind the paywall. SRAM libraries are not like traditional logic libraries. Due to the unequal number of PMOS and NMOS transistors, they require specialized rules and layout libraries. They do not need the flexibility of logic libraries, so they are hyper-optimized for one purpose: dense, reliable memory. The SRAM cell we found is a 1:2:2-2:2 cell. This means there is 1 fin per pull-up (PU) PMOS transistor, and 2 fins per pull-down (PD) and pass-gate (PG) NMOS transistor. These 2 PU, 2 PD and 2 PG transistors would usually form a single 6T high-current cell (HCC). An 8T HCC adds a read-pull-down (RPD) and a read-pass-gate (RPG) NMOS transistor, each with two fins. We measured a cell height of 406 nm, which brings the bitcell size to 0.0463 µm². That is a theoretical peak density of 21.6 Mib/mm². We estimate that a 6T HCC would have a cell height of 292 nm and a size of 0.0337 µm². This is ~12% larger than a 6T HCC on Intel 3 and 4. We also estimate the 6T high-density cell (HDC) to have a cell height of 228 nm and a size of 0.0260 µm². This is coincidentally the same as the logic standard-cell height measured earlier. The estimate puts the cell near Samsung 7LPP/5LPP and slightly below TSMC N7/N6. That is a theoretical peak density of 38.5 Mib/mm². 6T HDC is arguably the most important cell as it is used for the largest caches in a chip, the L3 caches and system-level cache (SLC). Both the Kirin 9020 and 9030 have split the SLC into 4 banks to raise total SLC bandwidth. In the Kirin 9030, the SLC increased from 2 MiB to 3 MiB per bank. Correspondingly, the number of arrays within the bank also increased by 50%, from 16 to 24. Each array can store 128 KiB and forms an orderly pattern on the die shot. From the Kirin 9020 to the Kirin 9030, the area of a 128 KiB SLC array decreased from 0.0477 mm² to 0.0392 mm², an 18% shrink. The achieved density is 25.5 Mib/mm², 66% of the theoretical maximum. While the SLC was quite similar across both chips, the L3 has seen some major changes, particularly in terms of its layout. The total capacity also went up from 10 MiB to 12 MiB. Much like the SLC, the L3 is also split into 4 banks. In the Kirin 9020, an L3 bank consisted of 16× 128 KiB arrays and 16× 32 KiB arrays. However, an L3 bank in the Kirin 9030 instead consists of 48× 64 KiB arrays. In the Kirin 9020 L3, a 128 KiB array was 0.0513 mm² and a 32 KiB array was 0.0154 mm². The size of the 128 KiB array is different on the L3 and SLC as the assist circuitry for the two arrays differs depending on their purpose. In the Kirin 9030 L3, a 64 KiB array is 0.0210 mm². Although not a like-for-like comparison, normalized for capacity, it is 18% smaller than the 9020’s 128 KiB L3 array and 31% smaller than its 32 KiB L3 array. The achieved density is slightly lower than the SLC, at 23.8 Mib/mm², 62% of the theoretical maximum. Unlike the L3 and SLC, the prime cores’ private L2 cache uses a 2-bank design. As the prime cores’ L2 is latency-critical, it likely uses 6T HCC instead of 6T HDC. The 9020 has 16 arrays in each bank while the 9030 has 32. Each array has a capacity of 32 KiB. A 32 KiB array in the L2 shrank from 0.0171 mm² to 0.0142 mm², ~17% smaller. The density is 17.6 Mib/mm², ~59% of the theoretical maximum for 6T HCC. SRAM scaled well from N+2 to N+3, shrinking by ~19%, close to the theoretical logic shrink. The caveat is that N+2’s bitcells were unusually large, bigger than comparable 7 nm-class nodes, so part of the gain is catch-up rather than true scaling. With the insights from STEEL’s teardowns, we will be doing a deep dive into SRAM in a future newsletter article. Subscribe Everything above came out of a single STEEL teardown: die annotation, block-level area analysis and TEM cross-sections through logic and SRAM. We’re diving deep into the most advanced datacenter and AI hardware hitting the market. To learn more about what’s in the pipeline or to commission a custom teardown, contact sales@semianalysis.com. Future Roadmap The same cross-sections that pin down N+3 also show where SMIC can go next. Although N+3 is already close to the practical limits of DUV multi-patterning in several layers, SMIC still has a few scaling levers left. A theoretical N+4 would likely start with cell height. N+3 uses 5 M0 tracks between its power rails. Moving to 4 M0 tracks, as on SMIC N+2 and TSMC N6, could reduce cell height by roughly 15%. The routing grid is only one side of the shrink; the front end also must fit into the smaller cell. One possible FEOL lever is reducing the p-to-n isolation spacing from two diffusion grid units to one. Intel used this scaling booster on Intel 4, and TSMC did so on its N3 family. This path trades layout flexibility for density. Fewer M0 tracks reduce local routing resources, while tighter p-to-n spacing raises integration and design-rule difficulty. M2 is also constrained by the cell height shrink. For SMIC to maintain a ~5.7-track cell, M2 would need to move toward ~35 nm. That would move another layer into SAQP territory. SMIC could also reduce the CGP from 57 nm to 54 nm. Intel reached a similar CGP on Intel 10 nm/Intel 7 without EUV. The local interconnect is also tougher. If SMIC keeps the 3:2 M1-to-gate ratio, M1 would need to shrink to 36 nm and would likely require SAQP as well. If SMIC moves to a 1:1 ratio, M1 could relax to 54 nm, but it would give up routing flexibility. Under this theoretical path, we estimate that SMIC N+4 could reach a cell height of 198 nm and a CGP of 54 nm, implying a Bohr density of 137.8 MTr/mm², on par with TSMC N5 or Samsung SF4. However, the difficulty is cumulative. Each step is individually plausible, but together they make N+4 harder than the transition from N+2 to N+3. It will likely take longer, cost more, and carry less process margin. A theoretical N+5 would require a larger integration shift. One possible path is backside contacts (BSCon), moving power routing and source/drain contacts to the backside, which would reduce front-side routing pressure and enable another reduction in cell height. The front-side metal pitches could relax to reduce the process complexity. M0 would likely relax slightly to ~34 nm, and M2 and M4 pitches could relax further. CGP is unlikely to shrink much further. Even with EUV, 48 nm has been the practical limit for yield and process control. This approach would allow N+5’s cell height to fall to 170 nm and its CGP to 53 nm. This implies a Bohr density of 163.6 MTr/mm², on par with Intel 18A’s HP library. However, this would not make N+5 cost-competitive with the leading edge. It would reach a similar density through a much more expensive route. The integration difficulty rises sharply, with new process flows for backside alignment, wafer thinning, contact reveal, and backside metallization. Past this point, standard density and interconnect scaling become increasingly unattractive. That is where Huawei’s roadmap stops looking like a normal foundry roadmap and starts looking like a packaging roadmap. At ISCAS 2026, Huawei unveiled its tau (τ) scaling law, reframing process scaling in the time domain. τ is the time cost of data movement and processing: switching delays in transistors, RC signal propagation delays in circuits, compute, memory, and networking latency. Outside Huawei’s terminology, this is called system-technology co-optimization. This is Huawei’s answer to its lack of EUV lithography. Without EUV, planar density cannot keep pace with TSMC, Intel, or Samsung. If transistor density cannot shrink further, Huawei’s alternative is to shorten wires, reduce buffering, and stack logic vertically. “LogicFolding”, Huawei’s implementation of this new scaling idea, is, in practice, an aggressive 3D stacking approach. AMD V-Cache places SRAM above or below a CPU die. AMD’s MI350X places active interposer dies (AIDs) underneath accelerator and compute dies (XCDs), with the AIDs handling cache, IO interfaces, the network-on-chip (NoC) and embedded metal-insulator-metal (MIM) capacitors. With LogicFolding, parts of the same logic block are split across multiple active dies bonded face-to-face at ultra-fine pitch. This allows Huawei to shorten some critical paths and reduce buffer overhead, not merely add cache capacity or offload the IO and interconnect. Shortening wires is where the higher clocks come from. A large share of a modern core’s delay and energy budget goes into driving long interconnects and the repeater buffers along them. LogicFolding distributes a block’s critical-path gates across multiple stacked tiers bonded at a very fine pitch, so the bond interface behaves like an additional metal layer and the longest paths get shorter. That is how Huawei expects to recover frequency and efficiency it cannot get from the process alone. Huawei’s roadmap shows its intent. Prime core frequency is targeted to rise from 2.75 GHz in the Kirin 9030 to roughly 5 GHz by 2031, far beyond what planar scaling alone could deliver. Prime cores with 3.1 and 3.39 GHz clocks are being tested in its labs, although their power consumption is unknown. Beyond that, chips are in the design, simulation or pathfinding phase, meaning the frequencies are targets. However, the direction matters more: LogicFolding also helps with performance, not just density. The catch is that Huawei’s density claim is not directly comparable to foundry densities. A stacked design can report more transistors per package footprint by adding active layers, even if each patterned die remains well behind TSMC or Intel in front-end density. This is how Huawei can claim to reach foundry 14A-equivalent density by 2031. This is not a like-for-like foundry comparison, with Huawei using stacked logic and measuring density per package footprint. On a normalized Bohr-density basis, SMIC N+3 is ~114 MTr/mm², 38% less than Intel 18A’s HD library. Huawei’s 3D roadmap closes the gap by stacking active logic, reaching 215 MTr/mm² by 2030. In 2031, the roadmap density jumps to 295 MTr/mm², implying either a third active layer, partial EUV insertion or aggressive planar DUV scaling. Huawei’s methodology makes other foundries look much denser as well. Applying it to AMD’s MI450X with an N2 top die on an N3P base die yields a theoretical density of 460.2 MTr/mm² in 2026, compared with Huawei’s 295 MTr/mm² in 2031. This Kirin 9030 does not use LogicFolding, remaining in a conventional mobile SoC package. Instead, it forms the baseline for how far Huawei and SMIC can push planar scaling. Future teardowns of Kirin and Ascend chips will show both planar logic density and Huawei’s hybrid bonding solutions. Export controls changed China’s optimization problem rather than ending it. EUV restrictions raised the cost and complexity of leading-edge manufacturing without freezing it. SMIC reaches N6-class logic density through DUV immersion, SAQP and DTCO, while Huawei shifts more of the burden onto architecture, packaging and system-level integration. Future nodes will be tougher. N+3 still had room to tighten local metals and reduce cell height and CGP. Further scaling without EUV leaves fewer levers. More aggressive multi-patterning adds masks and overlay error. SMIC can keep pushing DUV, but each step will get more expensive and less forgiving. The design side is just as critical. Huawei had domestic EDA tools and flows before the Kirin 9030, with the Kirin 9000s, 9010 and 9020 making that clear. Huawei was able to ship multiple consumer SoCs on SMIC N+2 and N+3 while cut off from the Western EDA stack. US export controls restricted EDA tools for advanced chips in 2022 but did not target tools for more mature chips. In 2025, the US government briefly placed much broader restrictions on EDA software from Synopsys, Cadence, and others, before lifting them less than two months later as part of a trade deal tied to rare earths. Huawei has been unable to access those tools because it remains on a US trade blacklist. That forced Huawei, SMIC and Chinese academic institutions to build their own tools and flows. Researchers at Peking University recently announced a prototype EDA tool for Huawei’s LogicFolding architecture, which requires a new flow to handle the multilayer layout and floorplan. This is not the same as replacing the full Synopsys or Cadence stack, but it shows where domestic EDA is headed: toward tighter co-optimization between architecture, process and packaging. These advances are also diffusing into the Chinese ecosystem. SMIC is licensing its N+2 and N+3 processes to HLMC/Hua Hong at the government’s direction rather than by choice. If the same process learning feeds into Ascend accelerators for AI training and inference, the choke point shifts from one named fab to an ecosystem. Alibaba’s T-Head silicon arm and Cambricon, a Chinese AI chip designer that is expected to supply ByteDance, could also be major beneficiaries. Sanctions aimed at SMIC alone become less effective once the manufacturing knowledge has spread to other fabs and design houses. China is not closing the gap with Intel, Samsung and TSMC. The teardown shows the opposite in several places: no EUV, no backside power, higher process complexity, and visible trade-offs. But China is still advancing. If domestic chips become good enough for phones, inference, networking and security-sensitive workloads, they can matter strategically without matching TSMC at the leading edge. Behind the paywall, we show what else STEEL can do, with material and process flow analysis of SMIC N+3, and analysis of the Kirin 9030 package. We’re diving deep into the most advanced datacenter and AI hardware hitting the market. To learn more about what’s in the pipeline, access the full Kirin 9030 and SMIC N+3 analysis or to commission a custom teardown, contact sales@semianalysis.com.
The facts didn’t fit Trump’s California conspiracy theory. So he adjusted the conspiracy theory
If you’re pushing the baseless conspiracy theory that the results of last week’s California primary elections were rigged against Republicans like gubernatorial candidate Steve Hilton, it would seem highly inconvenient that Hilton has succeeded in qualifying for the November runoffs.
Social Security is Facing a Political Crisis
On Tuesday the Social Security Trustees released their latest report on the system’s finances. The numbers didn’t change much: Unless something is done, the Old Age Survivors and Disability Insurance (OASDI) program, Social Security’s official name, will be unable to pay full benefits starting in either 2032 or 2034, depending on some technical issues. That’s not far away: If the Trustees are right, the prospect of a Social Security crisis will loom over the next presidential administration. It’s important to understand, however, the nature of the looming crisis. It won’t be an economic crisis. It won’t even be a serious fiscal crisis. Whatever you may have heard, Social Security isn’t in danger of going bankrupt. What we’re facing, instead, is potential political crisis. Congress and the White House could easily take action to sustain America’s retirement system. But given the current state of our politics, there’s no guarantee that they will. There is a widespread misunderstanding of how Social Security works. While Social Security was designed to look like a pension fund, it isn’t. A pension fund pays benefits out of a stock of assets it has accumulated over time. In contrast, Social Security operates as a government transfer program, like food stamps or Medicaid. Now, unlike food stamps — but like the highway trust fund — Social Security is on paper supported by a dedicated tax, the payroll tax, that is assigned to that program. But I say “on paper” because from an economic point of view assigning the payroll tax to Social Security is just an accounting convention. What matters for the U.S. economy is the overall balance between government spending and government revenue, not the difference between one type of spending and one source of revenue. So there’s no inherent economic significance to the fact that by 2034 payroll tax receipts will be insufficient to cover promised benefits. There is, however, a legislative issue. As long as the Social Security Administration can pay benefits out of payroll taxes and its cash reserve, there’s no need for Congress to vote each year to authorize benefits — they just keep going out until further notice. However, once those resources become insufficient, benefits will fall —by 17 percent according to the Trustees — unless Congress passes new legislation that “tops up” Social Security’s finances. Yet the current administration and Republican party are such extremists that there is a real risk that Social Security will be held hostage on behalf of their goals. If this should come to pass, the hostage-takers will claim that shoring up Social Security is unaffordable. Right on cue, Mike Johnson, the Trump-sycophant Speaker of the House, declared on Monday that “entitlement programs” like Social Security “have to be adjusted and fixed,” and that Republicans will introduce a plan to that effect next year. But this is a ploy, because while the cost of maintaining Social Security benefits at their promised level isn’t trivial, it is in fact affordable. According to the Trustees’ report, the actuarial balance of OASDI up through 2050 — the amount of additional funds it would need to keep paying full benefits for the next 25 years — is 1.06 percent of GDP. To put that number in perspective, the Trump administration proposes increasing military spending next year by $420 billion, equivalent to about 1.4 percent of GDP – without any discussion of whether that’s affordable Yet how did we get to the point where Social Security will need to be topped up? The main answer is that we have an aging population, with a growing ratio of retirees collecting benefits to workers paying into the system: Trump’s anti-immigration policies are making this problem worse. According to the Trustees’ report, lower immigration will deepen Social Security’s financial hole because many immigrants are working-age adults who will pay into the system for decades before they collect benefits. In fact, this problem may be much bigger than the report acknowledges: The report’s baseline assumption is that we’ll have net immigration of almost 1.2 million people a year, and even the pessimistic case assumes 750,000 a year. Meanwhile actual net immigration has already been cut far below that — and may now be negative. Moreover, Social Security is being financially damaged by growing income inequality in America. Payroll taxes are levied only on wages up to $184,500, and they don’t touch capital income. With the distribution of income increasingly shifting from labor to capital, as well as becoming more unequal among wage-earners, revenue from the Social Security payroll tax has been falling as a share of national income. From 1990 to 2024, it fell from 5.02 percent of gross domestic income to 4.46 percent. Which brings me to the question that, these days, we ask about everything: How might Social Security be affected by the advent of AI? A short answer: If, as many of us fear, AI accelerates rising income inequality, it will further reduce the payroll tax receipts that currently pay for Social Security and further endanger its finances. On the other hand, if AI, as its advocates promise, leads to faster economic growth, it will increase the potential tax base that could and should be used to support Social Security and other social insurance programs. But to take advantage of that larger base, we’ll have to get serious about taxing wealth and capital income. Is Social Security in trouble? Yes, but only because of the way its financing is currently structured — a structure that no longer works well because our society and economy have become so unequal. Moreover, Trump’s immigration policies are further endangering its already deteriorating financial condition. So don’t believe Republicans’ gaslighting that it will be necessary to cut Social Security benefits. All that is necessary to preserve Social Security is political will to raise taxes on the wealthy and a sensible immigration policy. MUSICAL CODA Actually 67 under current law.
Trump Previews Fall Strategy With Baseless Claims of California Vote Fraud
The president is using the slow count of mail ballots in California to renew his effort to cast doubt on election outcomes he doesn’t like, despite a lack of evidence of any widespread fraud.
Trump doesn’t rule out giving Jan. 6 rioters who attacked police ‘anti-weaponization’ fund payouts
In an interview with NBC News’ “Meet the Press,” Trump also baselessly claimed that Tuesday’s elections in California were “rigged.”
DOJ debunks social media claim of discrepancy in LA mayor voting count
A Los Angeles-based Justice Department official late Friday debunked a baseless claim of a discrepancy in the vote count of the city’s mayoral race that has circulated on social media.
First on CNN: DOJ sends prosecutor to observe LA ballot counting amid Trump’s baseless ‘cheating’ claims
The Justice Department sent one of its attorneys to observe ballot processing in Los Angeles Friday morning, the county’s elections office told CNN, after President Donald Trump claimed earlier this week the US Attorney’s office there was investigating the vote counting.
AI-generated images are making it impossible to distinguish truth from fiction. We need laws and AI watermarks to protect our shared reality.
Generative AI is destroying the baseline assumption that photographs bear some causal connection to reality. That's bad news for democracy.
Smart, hybrid, integrated, and engineered geothermal (SHIEG) systems
npj Thermal Science and Engineering - Geothermal sources can supply clean, sustainable, baseload thermal or electrical energy to fulfill societal demands in diverse climates via various...
ENTERTAINMENT: FUNFEST set for River Market; bike rodeo ready on Baseline Road
FUN501FEST 2026, "a celebration of our community and the people who sustain it" marking 501 Day, which happened on May 1, features live entertainment, a Creators' Village marketplace of local vendors, interactive art installations, food trucks and "engaging activities designed for the entire family," 4-8 p.m. Friday and 11 a.m.-6 p.m. Saturday at 350 E. Second St. in Little Rock's River Market Entertainment District.
‘The reality, for better or worse’: Columbia comp sci students and faculty grapple with AI’s disruption of the field
‘The reality, for better or worse’: Columbia comp sci students and faculty grapple with AI’s disruption of the field ‘The reality, for better or worse’: Columbia comp sci students and faculty grapple with AI’s disruption of the field A field that once seemed like a direct path to stable, lucrative work is becoming less certain under AI, students and faculty told Spectator. By Ria Vasishtha and Arjun Menon May 3, 2026 Shanying Liu / Deputy Illustrations Editor Asia Genawi, SEAS ’29, first encountered artificial intelligence during her sophomore year at her high school in Indiana. Back in 2023, policies on AI varied widely—even within the same district—and some schools had no guidance at all. As Genawi began to see the effects of AI trickle into her classroom, she turned to guidance from her home state’s Department of Education, which she said she found “outdated and inaccessible.” Genawi spent the rest of her high school career dedicated to the issue, including writing a 15-page memo exploring varying AI policies across Indiana schools, which she sent to every Indiana state senator and representative whose contact information she was able to find. She later helped draft a bill that would have required public high schools to establish and share their own AI policies with students in an accessible format. But with so many other bills already awaiting referral, hers never made it to committee. When Genawi arrived at Columbia to study computer science, she found that the gaps she had spent years trying to address in Indiana had followed her there. Many students and professors were also grappling with what AI meant for the field. Across Columbia’s undergraduate schools, computer science has become the most popular major—accounting for around 11 percent of degrees awarded in Columbia College, 32 percent in the School of Engineering and Applied Science, and 12 percent in the School of General Studies in 2024. While the number of computer science majors in SEAS grew from 166 to 173 during the 2024-25 academic year, the total number across the three schools fell around 4.5 percent from 393 to 375, driven by declines at Columbia College and General Studies. At Barnard, it has grown to become the second most popular major as of 2025, up from third the year prior. The number of degrees awarded have grown fivefold since 2016. Faculty members and students in the department told Spectator that the surge in the major’s popularity has been driven in large part by its apparent promise of a secure, well-paid career path. However, they noted that the recent rise of generative AI has compromised the sense of stability that once defined the field, even as interest in technical skills among students persists. According to data from the Federal Reserve Bank of New York in 2024, out of 74 majors tracked, computer science had the fifth highest unemployment rate among majors at 7 percent, while computer engineering had the second highest at 7.8 percent. “Anecdotally, people are moving away from CS as a major,” Daniel Bauer, a senior lecturer in the computer science department, said. “They’re still taking some of our classes, right, because they think they need that, but they’re instead majoring in other things.” Bauer described a field that had shifted from what he called a “nerdy outlier discipline” to one that “everyone needs,” a sign that technical skills have become a baseline expectation across industries with the emergence of AI. “A lot of people thought, ‘Oh, it’s a guaranteed path to a stable income. You get a six-figure job right out of your undergrad,’” Bauer said. Now, companies may see AI as a way to reduce demand for entry-level programmers, especially in what he calls “code monkey jobs,” roles where a large number of employees are “just writing out code nine-to-five.” Chris Murphy, a senior lecturer in the computer science department, described the current moment as the convergence of two forces: the downturn of the job market and generative AI coming onto the scene. “Things kind of started to really pick up in 2014, 16,” he said, recalling the years after the weaker 2010 market. Murphy added that this was also a period where “students were getting more experience outside of class than they were inside of class” through internships, jobs, and side projects. Murphy suggested that the earlier momentum helped make computer science feel like both an academic program and a reliable pathway into work. That feeling is harder to sustain now as the industry reconfigures, which is a shift that students are seeing firsthand. Rebecca Yu, SEAS ’27, said that during a recent internship in Big Tech, she saw that “every single tech company was racing to integrate” large language models. She added that her manager said her intern class was the first to complete their assigned projects, which she attributed to the increased efficiency that comes with AI. Frank Liu, SEAS ’29, said that with AI lowering the barrier to coding, companies now expect everyone to write code. “A lot of companies outright say if you’re not using coding in your job, you might get fired for not being productive enough,” Liu said. Yu similarly expressed that students increasingly have to accept AI being a part of industry workflow, adding that many also “make use of it to maximize their productivity.” Richard Li, CC ’28, described a “franticness” among his peers as the pace of change accelerates. He explained how “some new tool” is released regularly, noting that tools rise and fall in prominence within weeks. He has come to believe that he doesn’t need to “match that speed all the time” and keep up with every development. Li has learned that AI tools allow him to “take a breath” and “get back into the loop much easier than before” when he needs to. While some companies specifically target Columbia when hiring, sending representatives to campus career fairs and recruiting large numbers from Columbia, a computer science degree cannot be treated as a static credential, Yu described. “You can’t just get your degree and stop learning after that. You have to always be learning,” Yu said. “That is something you basically have to expect for your job.” AI has not only affected students’ decisions to major in computer science, but also reshaped how the subject is taught. Across departments, faculty members no longer treat work done outside the classroom as a reliable measure of understanding, shifting instead toward in-person exams, quizzes, and attendance checks. While homework once made up as much as 60 to 70 percent of a student’s grade in computer science classes, Murphy said that breakdown has since flipped, with exams and quizzes now accounting for the majority. “The assumption—especially among pessimists—is that they’re using it to do their homework,” Murphy said. “They take the homework assignment. They put it into generative AI. Out comes the solution. Submit solution. The end.” Murphy said that faculty members ranged in their responses from fully prohibiting AI use to actively teaching students how to effectively work with AI. He said he shifted his approach this year after his teaching assistants called his belief that students weren’t using the technology “naïve,” and now permits students to use AI on assignments. His surveys of students in his introductory classes have told a more complicated story than the one so-called pessimists offer: Around 60 to 70 percent of students who reported using AI said they used it to have concepts explained to them, to clarify assignment instructions, to generate test cases, or to create practice questions. “Those are all ways that you might have had a tutor in the past to do that, and now you have this AI essentially as your tutor,” he said. Bauer has taken a different approach, teaching students how to use AI productively without sacrificing the core principles of the discipline. “People who are trained to actually use AI tools productively, while being in charge of designing the overall project, will be very much in demand in the next couple of years,” he said. Over the past year, a working group of computer science faculty has been redesigning the introductory and intermediate programming sequence to integrate generative AI as both a subject and a tool, representatives from SEAS wrote in a statement to Spectator. The effort is “coordinated” across the department, “not course-by-course,” with a systematic assessment of how AI integration is affecting student outcomes underway to shape the next round of changes. The goal, the department explained, is for students to be able to “read, verify, and reason about code, whether they or a model wrote it.” Tony Dear, senior lecturer in the computer science department, is piloting individualized AI feedback in a discrete mathematics course, while a machine learning course is also testing a Socratic AI dialogue system designed to “guide inquiry rather than supply answers,” the statement reads. Beyond individual courses, SEAS funded seven cross-departmental faculty projects in generative AI this past academic year, spanning biomedical engineering, chemical engineering, industrial engineering, and computer science, according to the statement. This phenomenon extends beyond the computer science department. SEAS’ new AI minor, which it announced in November 2025, reflects the University’s attempt to adapt. The minor is specifically designed for noncomputer science majors and aims to provide “essential AI knowledge” while emphasizing practical and responsible use across disciplines, according to the SEAS website. At the center of the minor is AI in Context, a course taught by a team of faculty from the departments of computer science, applied mathematics, philosophy, music, literature, and the Writing Center, the representatives told Spectator. Bauer described it as targeted at students who want to bring AI into fields like economics or biology, rather than as a path into software engineering itself. At the graduate level, SEAS announced a new Master of Science in artificial intelligence in February, with the inaugural cohort starting in the fall. While recognizing how AI has helped them navigate coursework, students also described its rise as isolating, even among members of the Application Development Initiative, one of Columbia’s largest computer science clubs. “As people start to ask AI for questions that they would normally ask students and members, it starts to isolate people” Li said. Alexa Kafka, BC ’27, echoed Li’s sentiment, pointing out that computer science has always been a collaborative discipline with code reviews and pull requests, even as AI makes independent work easier. Despite this, Kafka noted that she hasn’t been to office hours “in at least a year.” Bauer added that faculty are exploring long-term solutions to the assessment problem. “One of the things that we’re trying to do at the department or school level is to maybe move towards computerized testing facilities,” Bauer said, referring to ideas in the department to revitalize computer labs as an alternative to pen-and-paper exams. “We have to accept that this is the reality, for better or worse.” But more exams, students say, does not mean better learning, nor does it reflect how computer science is actually done. “Code is so variable and can be approached in so many directions, there’s never just going to be one correct answer. There can be multiple optimal solutions,” Kafka said. Students also expressed concern that the shift toward exam-heavy grading pulls focus away from the applied, project-based work that better reflects both discipline and industry expectations. ADI has tried to respond more quickly than the curriculum can. Through a vibe coding workshop this past semester, club members said they wanted to help students understand not just that AI exists, but how to use it deliberately. “Vibe coding” refers to the practice of prompting AI to write code for a task or project, and has become widespread enough that ADI felt the gap in formal instruction was worth addressing, according to Sophia Huang, BC ’29. The workshop introduced students to the concept of context engineering, which Shaina Sahu, SEAS ’27, described as not just writing a prompt, but optimizing “the rest of the relevant information that you would feed into the model.” Without that, students said that AI-generated code can fail at scale or introduce difficult-to-detect bugs. Genawi said the shift toward AI has changed what it means to be good at computer science. “I think the shift has kind of gone towards what can you create and what can you do, instead of, what do you know,” she said. Indeed, during ADI DevFest, its annual hackathon, Li said more teams are trying to complete their projects than in years past. While sign-ups have historically outpaced submissions, Li said this is beginning to change because teams “have a tool that completes the vision for them.” While many students see AI as an equalizer that lowers the barrier to entry for coding and technical work, Murphy said it also creates new equity concerns inside the classroom around who can afford stronger tools and who already knows how to use them. “The equity issues I worry about are students who have access to paid versions of things versus free versions of things,” he said. Although popular AI models like ChatGPT and Claude have free versions available, both have paid models that can reach over $100 per month, which output different qualities of code. Murphy also pointed to differences in prior exposure to AI literacy itself. “Students who go to high schools or K-12 education, where they’re taught AI literacy, prompt engineering, context engineering, they understand how to do this, as opposed to students who go to schools who don’t teach that,” he said. “Then they come in, and then they’re somehow expected to know how to use these tools.” For Genawi, those disparities begin long before students arrive on campus. “It’s very interesting to consider how lucky you were geographically,” she said. “To be born where you were is influencing your ability to get access to these spaces.” Students like Yu, though broadly apprehensive about the future of computer science, are not without hope. “I’m optimistic, because I think if you’re truly passionate about CS, there will definitely be a place in the field to do something you want to do,” she said. Senior Staff Writer Ria Vasishtha can be contacted at ria.vasishtha@columbiaspectator.com. Follow Spectator on X @ColumbiaSpec. Staff Writer Arjun Menon can be contacted at arjun.menon@columbiaspectator.com. Follow Spectator on X @ColumbiaSpec. Want to keep up with breaking news? Subscribe to our email newsletter and like Spectator on Facebook. Produced with Spectate by Emily Ta
Queens man says family business, home robbed on wedding day
“They stole my mom’s jewelry, approximately $500,000-$600,000,” Basel Bakr said.
We spoke to over 30 CEOs and business leaders. Here's what worries them most.
Business leaders across the world are confronting a new operating reality: one where war, inflation and artificial intelligence are part of the baseline.
GK Digital Ventures and Harris Blitzer Sports & Entertainment Announce Return of “The Global Game: The Future of Soccer, Tech & Media Summit”
NEWARK, N.J.--(BUSINESS WIRE)--Apr 24, 2026--GK Digital Ventures, in partnership with Harris Blitzer Sports & Entertainment (HBSE), today announced the return of “The Global Game: The Future of …
DOJ demands 865K Detroit ballots, threatening possible legal action
Michigan and the Trump DOJ clash over a demand for 865,000 ballots from Wayne County, with state officials calling the request baseless and political.
DOJ demands Detroit-area 2024 ballots, escalating election scrutiny
DOJ official Harmeet K. Dhillon wrote that she is seeking ballots to investigate election law compliance. Democratic state officials call the request baseless.
Loneliness may be silently eroding your memory, new research reveals
New research suggests loneliness affects older adults' baseline memory performance but may not directly accelerate cognitive decline over time.
ICARUS experiment marks major milestone in first neutrino science results
The milestone by ICARUS paves the way to probe sterile‑neutrino oscillations for the full Short-Baseline Neutrino Program at Fermilab. The first results demonstrate the data’s excellent quality and suitability for physics analyses, as well as the maturity of the software tools for event selection, fitting and detector simulation.
How American, Delta, and United's newest business class suites compare
Doors in business class are now the baseline of premium flying as deep-pocket travelers keep paying up for more privacy and comfort.
‘Smart Underwear’ Measures Gut Gas for Microbiome Science
Device snapped into underwear monitors flatus frequency in real time to set a new baseline for what’s typical and gain insight into microbiome activity.
Zeldin tells climate skeptics to 'celebrate vindication' after repeal of baseline climate rule
The head of the Environmental Protection Agency on Wednesday defended his decision to repeal the legal determination that serves as the basis for federal rules to slow climate change.
Judge halts requirement that colleges show race isn't considered in admissions
The states argue the data collection effort threats student privacy and could lead to baseless investigations of colleges and universities
Judge halts requirement that colleges show race isn't considered in admissions
The states argue the data collection effort threats student privacy and could lead to baseless investigations of colleges and universities
Judge halts requirement that colleges show race isn't considered in admissions
The states argue the data collection effort threats student privacy and could lead to baseless investigations of colleges and universities
Extra entertainment, nightmare scenario, worthy of protest, baseline expectation, SAVE ACT distraction
Today's writers are Roger D. Eversole, Michael B. Shaffer, Terry Doran and Doug Johnston of Fort Wayne
Navigating the social media minefield
Scientists and researchers seeking to announce promising findings, exchange ideas or contribute to policy discussions outside of academic circles may find social media an expedient tool.But increasingly, scientists and other professionals are discovering that online platforms can be a double-edged sword: while sharing their work and expertise is crucial for visibility and meeting funder expectations, it also exposes them to baseless criticism and even accusations of spreading false information."It is getting harder and harder for scientists, and researchers in general, to convey their findings without facing attacks on their legitimacy and accuracy of their research," says Valère Ndior, a professor of public law at the University of Western Brittany in Brest, France, who has spent more than a decade studying misinformation on social media. Researchers attempting to challenge deceptive online information may themselves be accused of political bias or not being objective in their research, he explains, adding that some scientists disengage because they feel overwhelmed and unable to sustain meaningful dialogue. "It is quite sad to see in some instances that researchers end up leaving social media altogether because they feel like the pressure is getting too intense or feel threatened," he says.Ndior, a past president of the Francophone Network of International Law, is also involved in European efforts to make science and research findings more accessible to the public. As an ambassador for the Science Comes to Town (SCTT) project, he has participated in recent discussions on public perceptions of science and online misinformation. The European Commission-funded initiative connects Brest and two other cities — Kiel in Germany and Split in Croatia — and aims to strengthen ties between communities, universities and research institutions to promote science awareness and cooperation across Europe.Coordinated campaigns around contentious topicsNdior draws a clear distinction between critical thinking and thoughtless criticism. Yet he sees online discussions being increasingly shaped by suspicious behaviour and coordinated information campaigns — including "raids" aimed at dominating discussions and "astroturfing," where small groups act as larger movements to sway debates. Tracking these patterns has become harder because some platforms have restricted data or charge for it, despite provisions of the European Union's Digital Securities Act aimed at ensuring researchers can have access to such data. According to Ndior, "platforms became more reluctant from the moment access to data became a legal requirement rather than a voluntary initiative."Unfounded claims and scepticism about science existed long before Internet chat groups. But the noticeable rise in online hostility has intensified around contentious topics, such as the coronavirus pandemic and climate change, alongside the weakening of gatekeepers to moderate content.For Ndior, dubious claims about pandemic countermeasures and vaccine safety highlighted the struggle between "balancing freedom of expression and the protection of public health and order." Efforts to remove social media content deemed false or risky for public health triggered accusations of censorship and the silencing of legitimate discussion, and in some cases, legal challenges.Concerns about attacks on the credibility of scientists and researchers have generated a flurry of academic papers in recent years and prompted professional communities to address ways to counter online attacks. Such unease extends to the generation that has grown up with digital media. Ndior points out that his students, some of whom research artificial intelligence regulation, hate speech, online surveillance and other controversial topics, hesitate to share their work or viewpoints online. "Many of them don't want to expose themselves too much because they don't feel ready to be faced with the backlash of other people criticising their research," he explains.Safer harbours for engagementLegal challenges against false or defamatory posts are not necessarily a solution — especially given the trans-jurisdictional nature of the Internet. Lawsuits also take time and money, and if they fail, can be counterproductive. Instead, Ndior advises scientists and researchers to be selective in how and where they communicate. Professional online platforms and moderated discussion spaces offer a better environment for engagement. He also highlights the value of more traditional formats, such as radio interviews and in-person discussions at schools, libraries and science fairs, which encourage dialogue with the public. "Personal contact makes a big difference," he says, while acknowledging that these approaches lack the global reach of social media.Ndior’s interest in digital law and online misinformation stems from his time as a doctoral student in international law in the early days of social media. He began analysing posts from governments and heads of state to determine if they could be considered official and thus used as evidence in litigation. His work attracted media attention and led to conference and speaking invitations. "This was not part of my plan," he says, "but this is what happens as a researcher — your interest is sparked by a certain event or phenomenon, and it suddenly becomes your focus."Ndior believes SCTT provides "a double benefit" for researchers, enabling them to connect with peers and interact with the public. "There is a wide community of people who are not in academia who are critical thinkers looking for opportunities to ask questions, to get information and contribute their ideas,” he says. "I really value that kind of feedback."Science|Business, which connects policymakers, academics and industry through its news reporting, analysis and events, is an SCTT partner.
Navigating the social media minefield
Scientists and researchers seeking to announce promising findings, exchange ideas or contribute to policy discussions outside of academic circles may find social media an expedient tool.But increasingly, scientists and other professionals are discovering that online platforms can be a double-edged sword: while sharing their work and expertise is crucial for visibility and meeting funder expectations, it also exposes them to baseless criticism and even accusations of spreading false information."It is getting harder and harder for scientists, and researchers in general, to convey their findings without facing attacks on their legitimacy and accuracy of their research," says Valère Ndior, a professor of public law at the University of Western Brittany in Brest, France, who has spent more than a decade studying misinformation on social media. Researchers attempting to challenge deceptive online information may themselves be accused of political bias or not being objective in their research, he explains, adding that some scientists disengage because they feel overwhelmed and unable to sustain meaningful dialogue. "It is quite sad to see in some instances that researchers end up leaving social media altogether because they feel like the pressure is getting too intense or feel threatened," he says.Ndior, a past president of the Francophone Network of International Law, is also involved in European efforts to make science and research findings more accessible to the public. As an ambassador for the Science Comes to Town (SCTT) project, he has participated in recent discussions on public perceptions of science and online misinformation. The European Commission-funded initiative connects Brest and two other cities — Kiel in Germany and Split in Croatia — and aims to strengthen ties between communities, universities and research institutions to promote science awareness and cooperation across Europe.Coordinated campaigns around contentious topicsNdior draws a clear distinction between critical thinking and thoughtless criticism. Yet he sees online discussions being increasingly shaped by suspicious behaviour and coordinated information campaigns — including "raids" aimed at dominating discussions and "astroturfing," where small groups act as larger movements to sway debates. Tracking these patterns has become harder because some platforms have restricted data or charge for it, despite provisions of the European Union's Digital Securities Act aimed at ensuring researchers can have access to such data. According to Ndior, "platforms became more reluctant from the moment access to data became a legal requirement rather than a voluntary initiative."Unfounded claims and scepticism about science existed long before Internet chat groups. But the noticeable rise in online hostility has intensified around contentious topics, such as the coronavirus pandemic and climate change, alongside the weakening of gatekeepers to moderate content.For Ndior, dubious claims about pandemic countermeasures and vaccine safety highlighted the struggle between "balancing freedom of expression and the protection of public health and order." Efforts to remove social media content deemed false or risky for public health triggered accusations of censorship and the silencing of legitimate discussion, and in some cases, legal challenges.Concerns about attacks on the credibility of scientists and researchers have generated a flurry of academic papers in recent years and prompted professional communities to address ways to counter online attacks. Such unease extends to the generation that has grown up with digital media. Ndior points out that his students, some of whom research artificial intelligence regulation, hate speech, online surveillance and other controversial topics, hesitate to share their work or viewpoints online. "Many of them don't want to expose themselves too much because they don't feel ready to be faced with the backlash of other people criticising their research," he explains.Safer harbours for engagementLegal challenges against false or defamatory posts are not necessarily a solution — especially given the trans-jurisdictional nature of the Internet. Lawsuits also take time and money, and if they fail, can be counterproductive. Instead, Ndior advises scientists and researchers to be selective in how and where they communicate. Professional online platforms and moderated discussion spaces offer a better environment for engagement. He also highlights the value of more traditional formats, such as radio interviews and in-person discussions at schools, libraries and science fairs, which encourage dialogue with the public. "Personal contact makes a big difference," he says, while acknowledging that these approaches lack the global reach of social media.Ndior’s interest in digital law and online misinformation stems from his time as a doctoral student in international law in the early days of social media. He began analysing posts from governments and heads of state to determine if they could be considered official and thus used as evidence in litigation. His work attracted media attention and led to conference and speaking invitations. "This was not part of my plan," he says, "but this is what happens as a researcher — your interest is sparked by a certain event or phenomenon, and it suddenly becomes your focus."Ndior believes SCTT provides "a double benefit" for researchers, enabling them to connect with peers and interact with the public. "There is a wide community of people who are not in academia who are critical thinkers looking for opportunities to ask questions, to get information and contribute their ideas,” he says. "I really value that kind of feedback."Science|Business, which connects policymakers, academics and industry through its news reporting, analysis and events, is an SCTT partner.
Trump, Who Calls Mail-in Voting ‘Cheating,’ Just Voted by Mail
President Trump has long fixated on mail-in-voting to bolster his baseless claims of widespread voter fraud. But he recently used the method in a Florida special election.
In California, a Republican Sheriff Seizes Ballots, Prompting Concerns
Riverside County Sheriff Chad Bianco, a candidate for governor, recently took possession of more than 650,000 ballots as part of a fraud probe. Election officials say his investigation is baseless.
Your Thymus and Your Healthspan
Our thymus gland plays a central role in the development or our immune system, specifically for supporting T cell development (how these cells got their name, maturation and differentiation in the thymus) and discriminating between self and foreign, non-self antigen proteins with production of dendritic cells. As we age, our thymus gland shrinks—the process known as involution—with progressive change from spongy to fatty tissue, with loss of functionality. But this process, with respect to timeline and extent, markedly varies from one person to the next. And to make things even more complicated, our thymus gland anatomy and precise location in the chest also is highly variable. So the famous Frank Netter diagrams from the 1960s (such as at left below) don’t capture the remarkable heterogeneity that a random set of 12 chest CT scans (below at right). What I consider as 2 landmark papers in Nature this week (here and here) used AI to quantify health of the thymus in 2 large cohorts, and then correlate the metric to a broad array of health outcomes. In this edition of Ground Truths, I’m going to cover three questions: (1) How was thymus health determined by AI?; (2) How is thymus health linked to key clinical outcomes? and (3) Are there ways we can promote a healthy (aka rejuvenate) thymus gland in our later years? Subscribe 1. How was thymus health determined by AI? This was accomplished by extensive work developing and validating an AI pipeline for 2 stages, the first for localization and segmentation of the thymus bed, and the second to quantify a digital marker, termed the thymic health score. There’s a lot to this algorithmic development, so I won’t go into all the details, but just provide a rudimentary outline of what was done. The 1st stage used supervised learning with 2 radiologists reviewing 2,461 chest CT scans. That work trained a 3D U-Net model to automatically identify the thymic bed’s 3D cropped region using center of mass coordinates, which reached 99.8% accuracy. The 2nd stage used a foundation model that was pre-trained with SwAV (which stands for Swapping Assignment between multiple Views), a self-supervised model, that generated a high-dimensional thymus representation of 4,096 features. That compares to the radiology reductionist and subjective scoring of 0 to 3 where 0 means the thymus is fully degenerated and fatty, and 3 is considered dense, intact glandular soft tissue. Notably, the self-supervised learning (SSL) performance was superior to supervised (AUC ~0.75 vs 0.55, respectively) reflecting the shallowness of the 0-3 classifier vs SSL’s holistic, self-taught AI. The thymic health score ranged from 0 to 100, with the highest number indexed to fully preserved thymic health. The algorithmic work extended to explainability with both Shapely value distribution, meaning the model was holistic not relying on any “magic” pixel, and occlusion sensitivity, demonstrating the model’s performance was not affected by ribs, sternum, lung tissue, but rather focused on the thymic bed directly. Below are a few saliency maps from the occlusion sensitivity that indicate the thymic bed focus, not affected by neighboring structures, using the jet color scale. Now the AI was ready for processing ~25,000 CT scans, each over 30 MB files, from the 2 large participant cohorts to provide a thymic health score, it did so in less than 14 hours, which is less than 2 seconds per scan! Share 2. How is thymus health linked to key clinical outcomes? The two cohorts were the National Lung Screening Trial, NLST (N=25,031) and the Framingham Heart Study, FHS (N=2,581) each with baseline demographics and long term follow up of >12 years for health outcomes. From the NSLT, you can see below that increased age and body-mass index were correlated to reduced thymic health. And, overall, men had reduced thymic health scores compared with women. There are a lot of graphs in the paper for outcomes, for both the NLST and FHS cohorts. To simplify the major outcomes, below is a composite Figure with striking reduction of all-cause mortality and for the different causes of death. The results were consistent between the 2 cohorts. For example, the cardiovascular mortality hazard ratio average was 0.57 in NLST and 0.38 in FHS (weighted to be 0.53 for the aggregate). The reduction of mortality also extended to digestive diseases and pulmonary disease (data not shown below). Lung cancer incidence was reduced by 36% for high vs low thymic health (Figure below, with a panel -incidence and b panel-mortality, adjusted for age, sex, BMI, smoking status). Notably, there was a smoker’s paradox: lower incidence and improved survival for lung cancer in smokers who had a high thymus health score. Smoking was associated with lower thymic scores whereas alcohol was not. Higher thymic health was linked to higher HDL cholesterol, lower triglycerides, lower fasting blood glucose, and lower systolic and diastolic blood pressure. Inflammation biomarkers, such as C-reactive protein and interleukin-6, were elevated in the people with lower thymic health scores. In their second paper, the same team looked at the relationship between thymic score and outcomes for 3,476 patients receiving cancer immunotherapy. The non-small cell lung cancer progression-free survival was reduced in patients with low thymic health score. A parallel relationship for survival was seen for other types of cancer including melanoma, breast, and kidney with a 44% lower risk among high thymic scores (Figure). A similar pattern was seen with different immune checkpoint inhibitors. The thymus health score outperformed programmed death ligand-1 (PD-L1)and tumor mutation burden (TMB) assay of the tumor and was an independent prognostic marker of progression-free survival. In this dataset, a direct connect with thymus score and adaptive immune function was noted with a correlation before cancer treatment for both T-cell diversity and thymus T-cell production (a metric known as T cell receptor excision circles). Thymus Removal Study There is a highly relevant citation about thymus function before moving onto ways to rejuvenate it. In 2023, a very important study of thymus removal during cardiothoracic surgery was published, with 1146 patients undergoing thymus removal vs 1146 matched controls (mean age 55 years). The adverse outcomes for thymus removal were striking, with a 2.9 higher risk of all-cause mortality and doubling of cancer incidence, 1.5 fold increase in cancer mortality, and 1.5-fold increase in autoimmune diseases. The immune function showed marked compromise after thymus removal, as reflected by CD4+ and CD8+ T cells (Figure below) Share Ground Truths 3. Are there ways we can promote a healthy (aka rejuvenate) thymus gland in our later years? In recent months we’ve learned a lot about the process of thymus involution and the pathways by which this may be modulated. Thymus involution is primarily due to loss to thymus epithelial cells (TECs) and there are 2 subtypes with different functions: the cortical, responsible for positive selection of T cells, and the medullary, for negative selection. When adipose tissue infiltrates the thymus, thymic adipocytes are pro-inflammatory, knocking out T-cell output. Likewise, systemic inflammation, or “inflammaging,” from smoking, obesity, and chronic stress promote thymic adipocytes. Now, from preclinical studies, we know about key pathways that account for these processes. Historically, back in 2014, FOXN1 was recognized as a single master transcription factor essential for TEC function, with studies in mice with forced FOXN1 upregulation leading to thymus regeneration. That laid the foundation for potential regrowth of quiescent thymus cells. It took awhile to find pathways to accomplish this goal and to even refute its role as the master regulator. The Liver-Thymus Axis Feng Zhang and his team recently showed that injection of an mRNA vaccine to the liver encoding DLL1, IL-7 and FLT3-L in combination, factors known to support thymopoiesis (Figure), successfully enhanced immune function in aged mice. However, the effect was transient. The FGF21 Story Two reports in 2025 addressed the thymus gland’s production of FGF21, a growth factor (fibroblast growth factor 21) and its control of thymus involution. This growth factor is also produced by the liver, but it doesn’t have impact on the thymus in aged rodent models. In contrast, The liver’s hepatocyte growth factor (HGF) has been shown to reverse senescent TEC structural changes. Growth hormone stimulates production of insulin growth factor-1 (IGF-1) in the liver and the thymus. Two small studies in human participants (TRIIM, for Thymus Regeneration, Immunorestoration, and Insulin Mitigation) of 6 and 50 people, respectively, treated with combined human growth hormone, metformin and DHEA, suggested the potential of slowing epigenetic aging and improving thymus mass and function. But the combination of drugs, small sample, and lack of controls make conclusions murky. Increased thymus FGF21 led to increased CD8 T cells in old mice and extended healthspan and improved physical performance. Building on the previous FOXN1 data, ablation of β-klotho, the obligatory co-receptor for FOXN1, accelerated thymic aging. The benefit to FGF21 was confirmed in a companion paper using an FGF21 knock-in mouse model showing thymus enlargement and increased TECs throughout the lifespan. (schematic Figure below). The accompanying editorial for these 2 papers speculated: “The proposed link between age-associated thymus involution and organismal aging is intriguing, and suggests we might seek to achieve systemic rejuvenation by counteracting thymic involution.” Other Factors Lessons from the Axolotl (Mexican Salamander) In an elegant set of experiments of the axolotl (Figure), known for its complete thymus regeneration after total surgical removal, mediators were identified. Surprisingly FOXN1 was dispensable. But midkine, a growth factor, appeared to be the driver, the initiator of thymus cell growth, survival, and repair. Other components that contributed were Postn+ (periostin-expressing mesenchymal niches) and Ccl19, a chemokine protein from the bloodstream. RANK and RANKL RANK (receptor activator of nuclear factor κB its ligand (RANKL) have been considered key regulators of medullary TECs. In a study of both aged mice and human thymus cell cultures, their key role was confirmed, with multiple benefits of resorting TEC function, recruitment of progenitor cells, and T cell development. Subscribe Summing Up This body of work strongly supports the health of the thymus gland as a critical regulator of human healthspan, not just a correlate or link. The new landmark studies are reinforced by the regenerative biology experimental models that have shown, via an array of mediators, that healthspan of aged mice is at least in part dependent on thymus gland function and its critical role for maintaining adaptive immunity. It sure seems that we’d be better off having an immune reservoir of naive T cells instead of a profile in older adults of exhausted, low quality, memory T cells, with a senescent phenotype. It is striking that we ignored the importance of the thymus gland for many decades. A wake-up call was the results of the thymectomy matched control study reviewed here, with the opposite of high thymic health score improved health outcomes. The convergence of AI and healthspan here is notable. The exceptional and laborious work for developing and validating (no less explaining) the AI to quantify thymus health set the foundation for probing the connect the major health outcomes in two cohorts with extensive follow-up. We’ve never had a way to meaningfully quantify the thymus gland before, so this represents important leverage of supervised, self-supervised, and transformer models. The study wouldn’t have been possible without AI. And it begs the question as to whether there are ways we can maintain our thymus at a high health score. But it’s not so simple to rejuvenate our thymus. There are multiple risks including the induction of autoimmunity, increasing the risk of cancer, and inducing a pro-inflammatory state. For example, factors that increase TEC proliferation could compromise the medulllary negative selection process and leak out auto-reactive (self-attacking) T cells. Thymic cells could be induced to proliferate, but we know, for example, that growth hormone and IGF-1 are associated with an increase risk of cancer. If thymus rejuvenation doesn’t clear out the accumulated senescent cells in older adults, then inflammation within the gland could block the production of TECs. Keeping these risks in mind, we now have identified many ways to keep our thymus healthy as we age that undoubtedly will be tested in clinical trials going forward. Ultimately, the benefit-to-risk tradeoffs will be defined. In the meantime, we have desperately needed an immunome. As I’ve stressed in multiple prior Ground Truths, we have no test in the clinic to assess a patient’s immune status! It was very encouraging for me to get an email from Hugo Aerts, the senior author of the landmark papers in the days after they were published: “I also wanted to say that your Substack on “Why We Need an Immunome” was inspiring for this work. We found it a very compelling perspective and greatly enjoyed reading it.” Just having an assessment of a person’s immune status via a chest CT scan, with millions of these performed per year, might help in guiding the right immunotherapy for cancer (e.g. more intensive, combinations for those with low thymic health scores) and help define people who are increased risk for age-related diseases of cancer, cardiovascular, and neurodegenerative. The main theme of Super Agers is that we now have many layers of data (genetics, proteins, biomarkers, organ clocks that can be integrated via multimodal AI) for us to define a person’s specific risk decades before one of these age-related diseases leads to symptoms, enabling prevention. From the new studies this week, we can add thymus health score as yet another way to understand a person’s risk, a window into their adaptive immune system and healthspan. And perhaps someday we will be able to safely rejuvenate the thymus, or maintain its health through one’s life, and use the thymus health score to monitor. NB: I wrote this post (No A.I.) A Quick Poll Loading... ********************************************************************* Thanks to Ground Truths subscribers (now > 200,000) from every US state and 212 countries. Your subscription to these free essays and podcasts makes my work in putting them together worthwhile. Please join! If you found this interesting PLEASE share it! Share Ground Truths Paid subscriptions are voluntary and all proceeds from them go to support Scripps Research. They do allow for posting comments and questions, which I do my best to respond to. Please don’t hesitate to post comments and give me feedback. Let me know topics that you would like to see covered. Leave a comment Many thanks to those who have contributed—they have greatly helped fund our summer internship programs for the past two years. It enabled us to accept and support 47 summer interns in 2025! We aim to accept even more of the several thousand who will apply for summer 2026.
Smartmatic says Trump's 'campaign of retribution' is driving criminal prosecution
Voting technology firm Smartmatic asked a federal judge to dismiss a criminal money laundering indictment saying the case is political retaliation tied to President Donald Trump’s baseless claims about the 2020 election.
Iran after 48 hours: Tactical success, strategic uncertainty
During a crisis like what is now unfolding in the Middle East, a president receives a net assessment prepared from intelligence and military sources. It might evaluate the balance of forces, leadership dynamics and strategic calculations — without making policy recommendations. Its purpose is clarity: a baseline from which decisions can be made.
Justice Department Sues 5 States, Most Republican-Led, for Voter Data
The Trump administration sued Utah, Oklahoma and West Virginia, as well as Kentucky and West Virginia, to obtain unredacted voter registration databases in pursuit of baseless claims of voter fraud.
Sarah McCammon joins Third Way as Senior Fellow
Third Way welcomes journalist Sarah McCammon to the organization as a Senior Fellow, adding a fresh voice to its Moderate Power Project, the largest new center-left effort to bolster moderates since the 2024 election.
Codex Labs to Attend Health.tech Global Summit to Advance Root Cause Dermatology Innovation
BASEL, Switzerland - February 24, 2026 - Codex Labs, a biotech skincare company focused on integrative dermatology and
A Case Against 6 Democrats Lacked Urgency. Then Came a Swift Bid for an Indictment.
Prosecutors have been repeatedly caught between the president’s insistence that they undertake weak or baseless cases and the necessity of having to go to court.
A dozen 2026 tax-planning moves to help your business owner clients
The OBBBA reshaped several baseline assumptions for 2026 planning around state and local tax, above-the-line deductions and updated inflation adjustments.
Trump amplifies conspiracies on killing of Minnesota lawmaker
Trump shared a video baselessly linking Minnesota Gov. Tim Walz to conspiracies over brazen political killings in the state this summer.
Take The USITT Entertainment Design & Technology Workforce Demographics Study
Help USITT establish a baseline measurement of the demographics in the live entertainment industry by taking part in the Entertainment Design & Techno | Help USITT establish a baseline measurement of the demographics in the live entertainment industry by taking part in study administered by SMU DataArts.
Parkinson’s breakthrough changes what we know about dopamine
A new study shows dopamine isn’t the brain’s movement “gas pedal” after all. Instead of setting speed or strength, it quietly enables movement in the background, much like oil in an engine. When scientists manipulated dopamine during movement, nothing changed—but restoring baseline dopamine levels made a big difference. The finding could reshape how Parkinson’s disease is treated.
New plaques on Trump’s ‘Presidential Walk of Fame’ offer pointed descriptions of predecessors
The White House has installed plaques along President Donald Trump’s “Presidential Walk of Fame,” offering descriptions of his predecessors, often written in the style of his social media posts — including insults, baseless claims and random capitalization.
Wave Life Sciences Announces Positive Interim Data from Phase 1 INLIGHT Trial of WVE-007 (INHBE) for Obesity; Single Dose Resulted in Improvement in Body Composition With Fat Loss Similar to GLP-1 at Three Months Without Muscle Loss
Following a single subcutaneous 240 mg dose, WVE-007 (INHBE GalNAc-siRNA) improved body composition at three months compared to baseline, with a 9.4% reduction in visceral fat (p=0.02), a 4.5% reduction in total body fat (3.5 lbs; p=0.07), and a 3.2% increase in lean mass (4.0 lbs; p=0.01), with no statistically significant changes in the placebo group Sustained and robust suppression of serum Activin E supports expectations for continued improvements in body composition, further fat loss, and p
Robot Musk Goes Full ‘Poop Mode’ in Unhinged Art Freakshow
Robotic quadrupeds with hyperrealistic billionaire heads churned out artwork in the most unsettling manner at Art Basel Miami.