SCROLL NEWS / DISCOVERY
Search headlines
Find the stories shaping the conversation.
Results for " model " 3,155 found 🔀 AI-powered Shuffle
Why Anthropic freaked out the AI industry this week
Anthropic limits AI model use, citing safety concerns. But it's also trying to stop open-source models from competing.
How 30 Billion Pokémon Go Scans Became a Military AI Tool for Drone Warfare
30 billion images from Pokémon Go players have been used to train an AI model, now being adapted by Vantor for US military drone navigation.
He Went From $300,000 in Debt to Running a $1 Million Business. Now This Founder Is Taking His Model to America.
Brandon Willington is taking the lead generation system behind “Where U?” into the United States, following a breakout year that saw him named Marketing Entrepreneur of the Year.
Record AirPods Price Drops and a Rare Switch 2 Sale: This Week's Top Tech Deals
Multiple AirPods models hit record low prices this week, including the AirPods Pro 3 and AirPods Max 2. We're tracking these great discounts alongside an ultra rare discount on a new Switch 2 on Woot, plus a Summer sale at Sonos. Note: MacRumors is an affiliate partner with some of these vendors. When you click a link and make a purchase, we may receive a small payment, which helps us keep the site running.
Scientists Created Nuclear Fireballs in a Lab—on Purpose
When a nuclear accident happens, it’s vital that scientists have accurate fallout models.
MLB’s Rangers Sports Network joins pursuit of NBA teams’ local broadcasts
The Texas Rangers Sports Network is currently trying to acquire the local broadcast rights to various NBA teams that exited Main Street Sports Group, through a mostly direct-to-distributor model that somewhat mirrors Fubo’s failed attempt to do the same.
House committee votes to halt prior authorization pilot impacting radiology services
Lawmakers have opted against authorizing funding for the Wasteful and Inappropriate Service Reduction Model, citing concerns it will overburden patients and providers.
Apple’s third-generation Foundation Models explained
During the WWDC26 keynote, Apple announced its third generation of Apple Foundation Models (AFM). Here’s a breakdown of how each one works.
How ‘Stop! That! Train!’ Producer Unapologetic Projects Built a Business Model on Diverse Voices and Audiences
Unapologetic Projects co-CEOs Chris Quintos Cathcart and Tyler Boehm explain their business model that's built on diverse voices and audiences.
Gemini can now control your Google TV’s settings, but only on these models
Announced earlier this year, Google TV is now rolling out a Gemini update that lets you adjust TV settings with...
Apple Maps to Get These 10 New Features in iOS 27
Apple Maps is getting a range of new features in iOS 27, headlined by an upgraded Flyover experience that uses AI to improve the realism and detail of its aerial imagery. Flyover is a longstanding feature of Apple Maps and lets users explore more than 350 cities in 3D with detailed landmarks, roads, parks, and buildings. Apple described the enhanced Flyover in iOS 27 as combining aerial imagery with AI models to produce sharper, more lifelike visuals of cities around the world.
Exclusive | Nvidia Is Developing an AI Healthcare Model With Startup Abridge
The chip giant is joining with the maker of AI note-taking technology to train a model tailored for clinical conversations.
OpenAI and Anthropic keep warning about a future they're building
Both Anthropic and OpenAI have released papers warning that frontier models are being deployed before government regulation can catch up.
Anthropic says 'we made the wrong tradeoff' in new model guardrails
"We're changing Fable 5's safeguards for frontier LLM development to make them visible," an Anthropic spokesperson said.
Satya Nadella has a message for Microsoft's tokenmaxxers
Microsoft's CEO said workers should use the right AI model for the job as companies look for ways to control rising AI costs.
Apple’s iPadOS 27 beta downloads briefly included two unsupported iPad Pro models
Apple briefly listed iPadOS 27 beta 1 restore images for two older iPad Pro models before removing them from its developer downloads page.
A $1,500 foundation model that rivals larger LLMs
Sapient researchers trained a 1B reasoning model on just 40B tokens — scoring competitively with 2B-7B models at a fraction of typical pretraining cost.
iOS 27 gives FaceTime huge upgrade with this powerful new feature
FaceTime is getting one of iOS 27’s biggest upgrades thanks to a powerful new feature for compatible iPhone models: dual camera.
Xbox exploring ‘radically different’ console business models
Microsoft is re-evaluating its plans for Project Helix and considering new ways to lower Xbox console costs.
Xbox exploring ‘radically different’ console business models
Microsoft is re-evaluating its plans for Project Helix and considering new ways to lower Xbox console costs.
Xbox Boss Says a Mass Audience Can't Afford 'Thousands of Dollars' to Spend on a Next-Gen Console, So Microsoft Must 'Look at New Business Models'
The spiralling cost of new and existing consoles will require "radically different business models" over the coming generation, Xbox boss Asha Sharma has said.
Xbox CEO Asha Sharma teases "radically different business models", hints at potentially cheaper console to arrive this year
Xbox's newly-appointed CEO has hinted at what could be a major shakeup to the console model business that could help it navigate the RAM apocalypse
iPhone 17's 8GB Limit Costs It These Two Siri AI Features in iOS 27
Apple this week revealed what its most advanced on-device AI model does, and the feature list is shorter than the hardware requirements might suggest. In its Siri AI announcement during WWDC 2026, Apple confirmed that the model powers two things: more expressive Siri voices and a major accuracy gain for systemwide dictation. Both require 12GB of unified memory.
Xbox CEO outlines strategy to prioritise cost-efficiency and exclusive titles over premium hardware
"We will continue to look at new business models – that's what is needed rather than premium, high-performance consoles," says Asha Sharma
Large language model helps radiology department boost patients’ report comprehension
Emory University has utilized OpenAI’s GPT-5 to create LLM-produced radiology report summaries, clickable terms with embedded educational content, and AI-generated videos.
P. Terry’s redefines its business model; initiates employee ownership, profit sharing
The Employee Ownership Trust (EOT) will benefit 1,800 employees across 38 locations, allowing employees to share in the long-term values of the business.
Researchers are furious over Anthropic's hidden AI limits
Anthropic faces backlash as Mythos-based models intentionally limit help for AI research, raising transparency and ethical concerns.
A classic brain test exposed AI's biggest weakness
Researchers gave top AI models a classic attention test used in psychology and found a major flaw. While the models could correctly name colors in short lists, their performance deteriorated sharply as the task became longer and more complex. Some leading systems fell from over 90% accuracy to nearly complete failure.
'We'll Start To See Radically Different Business Models' - Xbox Boss Talks Future Of Consoles
She says mass audiences can't afford thousands on a console generation
Why Model Flows Are the Key for Reproducibility in AI for Science
Reproducibility is absolutely critical in science, but it’s a troublesome characteristic when it comes to AI. Frontier models developed by Big AI may deliver superior accuracy and reasoning capabilities, but they do so largely as black boxes with little regard for reproducibility. If AI is going to turbo-charge scientific productivity, it must do so without […]
Google makes its AI Plus plan cheaper, doubles the cloud storage that comes with it
Previously, the Google AI Plus plan cost $8 a month and came with 200GB storage. Google scored a big victory yesterday as its Gemini model became the...
Sports Illustrated Swimsuit Runway Show models say every body is a target online
As “skinny” returns to mainstream fashion and wellness culture, several celebrities have expressed fatigue with constant criticism.
Verizon says these Pixel Watch models are about to get Wear OS 7
Google announced Wear OS 7 less than a month ago and, apparently, it’s already ready to roll out, as Verizon...
15 technologies that seemed impossible 20 years ago
Smartphones, mRNA vaccines, reusable rockets, language models — the past 20 years produced technologies that experts said were decades away or simply impossible
Truist says Meta has found its next $20 billion business, stock set to gain more than 40%
Analysts at Truist estimate Meta's new subscription model could generate over $20 billion in revenue by 2030.
Gemini 3.5 Live Translate rolling out to Google Meet & Translate with new ‘listening mode’
Google today announced Gemini 3.5 Live Translate as its latest model for live speech-to-speech translation.
Nigel Farage to headline Liz Truss’s UK CPAC conference after apparent snub
Reform had previously suggested Farage would be ‘steering clear’ of event, modelled on US conservative gathering
Apple's New AI Models Contain 'None' of Google's Gemini Assistant
Apple executives have detailed the architecture of the company's new Apple Foundation Models (AFM) and clarified exactly how Google's technology factored into their development.
A reporting checklist for large language models in behavioural science
Nature Human Behaviour - Large language models offer new opportunities for behavioural science, but their rapid evolution poses challenges for research rigour. We introduce a consensus-based...
tvOS 27 Drops Support for Two Apple TV Models
The upcoming tvOS 27 update that was previewed today drops support for two Apple TV models. The two Apple TV models that are compatible with tvOS 26 but not tvOS 27 are the Apple TV HD from 2015 and the Apple TV 4K (1st generation) from 2017. tvOS 27 is compatible with the Apple TV 4K (2nd generation) and newer, according to Apple's documentation. tvOS 27 features a redesigned Podcasts app, smoother app launches and animations, support for larger text across the interface, and more.
https://9to5mac.com/2026/06/08/watchos-27-apple-watch-update-details/
Although watchOS 27 drops support for several older Apple Watch models, Apple has confirmed to 9to5Mac that one compatible model was mistakenly left out of its official list.
We can create the future of science right now
Paul Litvak is the founder and Executive Director of the Robyn Dawes Institute and a Visiting Scholar at UC Berkeley. He has a PhD in Behavioral Decision Research from Carnegie Mellon where he studied emotions and the sunk cost bias. Over 15 years in industry, he solved a wide range of challenging technical problems. At Meta he created models optimizing the ad review process. At Google he ran experiments to measure social influence. At Airbnb, he was a product manager leading machine learning teams optimizing search ranking and price suggestions. He also co-founded and led product at Rhythmic Health, a venture-backed biosensing startup, creating an accurate low cost system using a color changing strip and a smart phone to measure salivary lactate. The bottleneck At this point, it is uncontroversial to say that science needs to stop using the PDF article as the unit of knowledge and currency. The unbearable slowness of scientific publishing, the profit motive and margins: I’m not saying anything new. The PDF also sucks because it’s hard to extract structured information, which makes it hard to do evidence synthesis. As a result, we do much less evidence synthesis than is needed. And evidence synthesis ultimately undergirds most policy and medical decision-making. I can see second-by-second real-time odds for any sporting or newsworthy event, but a school board can’t see the best evidence on whether their 8th graders should be taught algebra. As a society, we don’t treat this as an important problem. Again, not controversial. Not only is the problem well understood, but the solution has already been laid out. What we need is AI-assisted living evidence synthesis - (1) an open knowledge graph of atomic claims, (2) claims linked to evidence, (3) assessment and synthesis of each piece of evidence, and (4) continuous updating with new data. What few realize (yet) is that the technical capacity to build this vision for a significant portion of science already exists. Not only that, scientists and startup teams are already building many of these components. I know this because I’ve been surveying the space and talking to many of the builders. There are some missing pieces: for example, evaluations of how well some of the components work. But at this point, most of what’s missing is a fully end-to-end working integration of all of these parts. In the rest of this essay, I’m going to lay out all the parts of a working living evidence layer for science and who is working on them, and propose concrete next steps for building this system. What’s now possible The diagram above outlines the components of a living evidence synthesis platform, including some of the teams working on each component1. Scientific PDFs are processed into claims with associated evidence. The evidence is subjected to a forensic audit, methodological evaluation and robustness and reproducibility checks. Finally it’s given a weight in a continuously updating synthesis. What follows is a description and status of each component and a few of the teams working on them. Document understanding The first thing you need to be able to do is turn an article into structured data. Mostly that means parsing PDFs. There are often multicolumn layouts that confuse non-specialized PDF to text processing libraries. For scientific papers, there is the added complexity of parsing formulas and tables and figures. This is a really hot area - there are startups offering APIs, and it seems like a new open source package gets posted to Github every few weeks. What follows isn’t exhaustive. A package called GROBID was the state of the art for a while, they didn’t update their package for nearly two years until very recently. In the meantime reducto.ai released an AI powered PDF extraction API, PaddleOCR became popular, IBM released a model called Docling, and both Mistral and Gemini created models and libraries. I also know of at least one other well-funded psychology research group working on a paper parser. By contrast, there are few open evals in this space, with no extensive evals for complex table comprehension in particular. Nonetheless, I’m confident this will be a solved problem soon, given the combination of LLM advances and developer interest. Hypothesis level extraction There has been increasing interest in comprehending the extracted text of papers and linking information to evidence for each hypothesis. A lot of work has already been done. Trialstreamer (Marshall et al. 2020) and RobotReviewer LIVE (Marshall et al. 2023) demonstrated automated extraction of trial population, intervention, and outcome at scale on clinical RCTs. PaperQA2 (Skarlinski et al. 2024) and Ai2 ScholarQA (2024) extended this to retrieval-augmented question answering with citation grounding. Elicit, Consensus, and SciSpace operationalized claim-level extraction for end users. OpenEval (Booeshaghi et al. 2026) is the most recent and most ambitious: 1.96 million atomic claims extracted from 16,087 eLife manuscripts using Claude Sonnet 4.5, grouped into ~299,000 results, with LLM evaluations showing 81% agreement with human peer review on a 2,487-paper subset. None of these solutions link claims to test-statistics, as you would need to evaluate randomized controlled trials. This is why I built the evidence.guide API - to extract hypotheses and associated test statistics from behavioral science papers. The best public eval of this kind of extraction I’m aware of comes from the recent SCORE project - they had humans code thousands of psychology papers to extract their claims by hand. It would be extremely helpful to the world if all scientific PDFs were available as structured open data. I’ve been working to make this happen, both directly at Berkeley and through coordination with large entities I can’t yet speak of; as hard as it is to do, I think it’s possible2. Forensic audit A lot of work has been done on forensic audit, but some gaps remain. Of course, for biology papers that rely on images for evidence, there are a variety of tools (notably Proofig and ImageTwin) to spot anomalies. These are still well short of what sleuths like Elizabeth Bik can do on her own, but these tools are constant companions among fraud analysts. There’s someone working on auditing Excel files for anomalies, and a number of teams are automating numerical checks like GRIM and SPRITE, including Scrutiny project, the INSPECT-SR team as well as statcheck. The regcheck team is building a way to use AI to compare preregistrations to analyses in papers, to ensure there aren’t significant deviations. Nonetheless, there are many other kinds of anomalies to screen for, both public and less publicly known. And there are no formal evals for anomaly detection that I’m aware of. Still, there’s a lot to draw from in this space and I’m pretty certain we will be able to scan papers for most kinds of obvious anomalies in the near future. Methodological review This area has been white hot, though I fear for many of the startups in this space, because this capability may become commoditized. There are at least six different AI peer review companies, including Refine.ink, Reviewer3, ReviewerZero.ai, Q.E.D. Science, Paper Wizard, and Isitcredible. Coarse (a pun on refine) was also recently created as an open source alternative. These systems provide qualitative feedback on the content of papers, spotting methodological weaknesses and mathematical errors. They seem to work pretty well, and many academics report bitterly that they exceed the average quality of typical peer reviewers. But there are few evals here either. What evals exist so far involve using LLM-as-judge (circularity problems abound) or comparing against human reviews of questionable quality. What you’d ideally want is an eval that measures capturing known errors in papers3. Reproducibility and robustness Another active area has been using AI agents to automate computational reproducibility4 and robustness5 checks in papers that report numerical results. For more recent papers where data and code are available, AI agents can see whether they can re-run the analyses and produce the numbers reported in the published paper. In addition to a handful of individual academics who have been experimenting using Claude Code for this, the Institute for Replication is a leading group working on building an end to end system. The evals related to this problem are the most mature, with CORE-Bench (Siegel, Kapoor, Narayanan 2024) and PaperBench (OpenAI 2025) available to benchmark agents on this task. There is also work on getting AI agents to test alternative ways of analyzing the data to ensure the results are robust to small analytic design choices. Synthesis This is the most underdeveloped area where significant investment is required. Although some automated evidence synthesis systems exist — for example, otto-sr is building an AI agent to write systematic reviews — none of these incorporate the full range of paper level signals to weight evidence appropriately. Nor is there anything like an eval or a gold standard for a good systematic review. Arguably Cochrane reviews are the closest we have to gold standard human systematic reviews, though I’ve heard academics in the know complain about their uneven quality. A key question for a synthesis platform is how to weight anomalies and methodological issues in assessing the quality of a piece of evidence. This is an unsolved problem and one I’m very keen to work on. Continuous updating There are many pieces of basic infrastructure available for monitoring for new research and initiating updates. OpenAlex is the current open citation graph. Retraction Watch integrated into Crossref in October 2023. Scite tracks how citations support, contrast, or mention prior claims. The Living Evidence Network demonstrated continuous-update workflows in clinical guidelines. Engineering this is a relatively straightforward task. When you look over this technical architecture and all the progress being made, it’s hard not to be optimistic that a living guide to scientific evidence will be built. The stakeholders are ready The social infrastructure for this is starting to coalesce — it’s not just a pie in the sky academic exercise to imagine this coming into existence. Institutions like the Center for Open Science, the Institute for Replication, the INSPECT-SR, the Living Evidence Network and more are all working on scaling work to improve research quality. Funders are also aligned. The Sloan Foundation has funded living evidence work through COS. Coefficient Giving supports the Institute for Replication and COS. The Astera Institute and the Institute for Progress have shown interest in this space. NIH has established an Office for Replication and Reproducibility. Although there are (very unfortunately) serious headwinds in science funding generally, there is an active group of funders interested in metascience. A brief word about what I’ve been doing at RDI. First, as a Visiting Scholar at Berkeley I’ve been actively figuring out how a non-profit and a public university can conduct and make public the results of large-scale academic article data mining. With some of the money I raised from donors, I commissioned a legal analysis of recent case law and publisher text data mining (TDM) agreements in order to understand whether a massive open data mining of academic articles is possible (with caveats, it is). I’ve also been working to bring together stakeholders in this space, and identify gaps. I’ve also been doing some software development in this space, with more to come. A pilot proposal The assumption undergirding all of this is that an AI, given all this information, would make the right judgment about a scientific claim with lots of conflicting evidence, weighing all the factors appropriately. That’s the hypothesis we need to test. Randomized control trial research is the best place to focus on first. RCTs are used to make many of the important decisions in society - from medical trials to public policy changes. And they use a relatively uniform set of inferential statistics with lots of known and available diagnostics. Behavioral science experiments, within the broader realm of RCTs, should be first used as a testbed whose results can be generalized. Because behavioral science is at the vanguard of open science practices, replications abound (there are thousands of them) to serve as ground truth training data. Key Hypothesis Therefore the pilot would test, in behavioral RCTs that have been replicated, whether the quality of evidence for a claim can be used to accurately predict whether that claim will replicate. Secondary Hypotheses Compared to claims that replicated, non-replicated claims demonstrate a greater share of forensic anomalies in their source literatures. Hypothesis level claim and statistic extraction is accurate enough to scale living evidence without onerous human review costs. Replication prediction is more accurate than prediction markets6 or journal prestige. If all the different quality signals we gather do accurately predict which studies will replicate, then we can use that model to score evidence to power the living evidence layer. Why this is informative regardless of outcome If the pilot succeeds, the architecture extends to medical RCTs (where Living Evidence already operates and integration is mostly about claim representation), then to slices of basic biology with stable replication structure. If it fails, the field learns which quality signals are load-bearing and which ones metascience has oversold. Either result is a contribution to knowing what the literature supports. Implications for funders Because this burgeoning ecosystem of builders already exists, a well-informed philanthropic or government funder could play a crucial catalyzing role in bringing this future about. They could play at least three roles: creating open structured datasets, publishing open benchmarks and incentivizing the solving of key technical challenges. First, open archives of papers that the government maintains, like PubMed, could be turned into structured data amenable to large scale metascience and claim aggregation. I know the US government already has an interest in doing this, though some key open questions remain unanswered. How do you determine the best models and systems for accurately extracting information from papers? How can you establish a robust way to allow researchers to flag errors and correct them? And finally, how do you create a legal regime in working with publishers to maximize the scope of available papers? For the latter, a university or a private philanthropy may be better positioned to make structured data publicly available under journal subscription terms or fair use. Philanthropists or government funders could also coordinate to create or commission benchmarks that evaluate whether important problems have been solved. For example, an open benchmark for claim extraction from a range of different scientific article types would be extremely helpful. Ensuring the underlying data are accurate is vital for creating these evaluations. I’ve discussed opportunistically using various human-created datasets for this purpose, but a consistent problem is that errors in human data make it difficult for them to serve as a gold standard. Finally, with structured data and benchmarks available, the government or private philanthropy could use them to incentivize groups to develop machine learning systems that meet these benchmarks. Prizes are one potentially valuable tool for this. For example, you could establish a prize for a system that accurately updates a living meta-analysis for a small set of claims. Prizes are particularly useful as signals of problem importance, and can help create vibrant ecosystems of public and private research—see, for example the role the government played in kickstarting the current work on nuclear fusion. Conclusion The drawbacks of the current scientific publishing system are known. Scientists agree, metascientists agree, philanthropists agree: the published PDF plus citation graph isn’t the right substrate for maintaining a representation of the evidence base in science. The pieces needed to build the alternative either already exist or are rapidly taking shape. The community is forming around exactly this problem, with concrete partnerships and shared infrastructure. A pilot should start on behavioral science RCTs because that’s the slice of empirical science most amenable to legibility, where replication ground truth is richest, and where the failure modes are best documented. What’s been missing is the galvanizing mission to assemble these pieces into something that works. That’s what I’m proposing to build. *** 1 I have a broader field map that I’ll release publicly soon. This is me, building in public! 2 If this is something that you are excited about, please reach out and talk to me. 3 More on this very soon too! 4 This tests whether, given the code and the data, you can get the same statistics as reported in the published paper. 5 This tests whether the results are the same as a paper’s given alternative analytical decisions in conducting the analysis (like outlier omission). Closely related is the idea of a “multiverse” where you come up with many different ways of answering the same underlying research question with the same data, and test whether the results hold in all those alternative methods. There’s been work on the latter as well. 6 Some of the replications, e.g. those from the SCORE project, had paired the experiments with forecasts from prediction markets. So we get to look at this for free.
7 of the best back braces for easing lower back pain, recommended by healthcare experts
To find the best back brace for lower back pain, I asked healthcare professionals to share the models they recommend for delivering comfort, support and stability to the spine.
tvOS 27 drops support for two Apple TV models
Two Apple TV models currently supported by tvOS 26 are not compatible with the new update, tvOS 27. Here are the details.
Apple drops support for a long list of iPads and Apple Watches with latest OS updates
iPadOS 27 drops support for M1-powered iPads, and watchOS 27 won’t even support Apple Watch Series 9 models.
watchOS 27 supported by just five Apple Watch models
Apple has hammered the Apple Watch compatibility list for watchOS 27, with just five models making the cut this year.
watchOS 27 Drops Support for Apple Watch Series 9, Ultra 1, SE 2, and Older
Apple today confirmed that watchOS 27 will not support the Apple Watch Series 9, Apple Watch Ultra (first generation), or Apple Watch SE (second generation), effectively drawing a line at devices equipped with the S9 or S10 chip. The only Apple Watch models compatible with watchOS 27 are the Apple Watch Series 10, Series 11, Ultra 2, Ultra 3, and SE 3. Despite the fact that the Apple Watch Ultra 2 contains the S9 chip, the Series 9 which contains the same chip will no longer be supported.
Will your Apple Watch run watchOS 27? Here’s the full list of compatible models
watchOS 27 drops support for several older models of the Apple Watch. To check out whether your Apple Watch will run the new system, check out the full compatibility list below.
iOS 27’s most powerful on-device AI requires iPhone 17 Pro, iPhone Air
iOS 27’s most powerful on-device AI model will be restricted to some of Apple’s newest products. Here are the details
NotebookLM's new update will help you build source repository from chat
Google is making Gemini 3.5 the default model in NotebookLM
New York businesses cash in as Knicks' historic NBA Finals run sparks economic boom
Former Modell's Sporting Goods CEO Mitch Modell joins 'Varney & Co.' to discuss how the Knicks' NBA Finals run is fueling a major economic boost for New York City businesses.
She Paid $20,000 For Her Own Pregnancy, Then Built The Fix For Medicaid Moms
Malama Health's doula-led model cuts preterm births 45% for Medicaid moms. Mika Eddy is building the maternal care infrastructure America never had.
$2 million gene therapy cures require a financing model
“We do not lack cures. We lack the infrastructure to pay for and deliver them,” writes William Padula.
Why You Need to Prepare Your iPhone Before Downloading iOS 27
Prepare for the iOS 27 beta release at WWDC 2026. Learn about new AI features, Siri 2.0, and which older iPhone models will lose support.
New Mayo–Microsoft collaboration may mean much for AI in healthcare
When Mayo Clinic and Microsoft announced last week that they’re partnering to develop a frontier AI model for healthcare, observers could see where Mayo’s expertise in advanced digital medicine would interest Microsoft. The Big Tech behemoth has not been coy about its healthcare ambitions.
How Six Flags’ Expanded Membership Model At Six Flags Entertainment (FUN) Has Changed Its Investment Story
In recent weeks, Six Flags Entertainment Corporation expanded its flexible Membership program to six additional parks, introducing year-round, multi-park access with regional tiered options and low monthly payments to strengthen guest loyalty and recurring revenue. This shift toward a subscription-style model is material because it increases revenue predictability, deepens customer engagement, and may reshape how investors assess the resilience of Six Flags’ business. We’ll now examine how...
I Tested All 4 of Microsoft's New AI Models. Here's the Brutal Truth
Microsoft says its new MAI models revealed at Build 2026 are the future. After testing them, I'm not convinced they're ready for that spotlight.
Ebola spread in central Africa could match 2014 record outbreak, US health officials say
Modelling from US CDC shows Ebola spread could be on ‘dangerous trajectory’, but experts warn outbreaks can be very hard to predict
Microsoft AI chief says company was “set free” from OpenAI to pursue superintelligence
Microsoft AI chief Mustafa Suleyman says the company was “set free” from OpenAI as it unveils seven in-house models and pursues superintelligence.
CDC report: Ebola outbreak could rival the worst on record unless world acts
New modeling from the CDC shows that if measures aren't taken immediately, this outbreak could sicken more than 20,000 people in the next three months.
Touch ID coming to iPhone Ultra this fall, here’s why
iPhone Ultra is coming this fall as Apple’s first foldable model, and it will use Touch ID instead of Face ID for authentication. Here’s why.
HERMES: HiERarchical Modelling for Exoplanet Science
ESA's Ariel Space Mission will characterise the atmospheres of approximately 1000 exoplanets to quantify population-level trends.
Meta AI Chief Sees Opportunity in Models’ Giving Health Advice
Meta Platforms Inc. Chief AI Officer Alexandr Wang said the company’s future artificial intelligence models will differentiate themselves from competitors through their consumer health capabilities.
2027 Dodge Charger Daytona Is $12,500 More Expensive Than It Is In 2026 For No Apparent Reason
Usually, when pricing changes this drastically between model years, it's because of some major updates, but that definitely isn't the case here.
Marshall Milton ANC review: Making the rare case for premium on-ear headphones
Marshall's latest headphones are like its Major and Monitor models had a baby.
NSF renews support for MIT-led AI and physics institute, expanding a new model for discovery
The MIT-led Institute for Artificial Intelligence and Fundamental Interactions (IAIFI) has received renewed support from the National Science Foundation for an additional five years.
Amazon, Microsoft and Google are quietly morphing their businesses — and Wall Street is missing the big picture
Tech giants are shifting from solely providing computing power to distributing AI models, unlocking a lucrative new revenue stream.
Run Google's Gemini LLMs right on your Mac with the new AI Edge Gallery
Google has released its AI Edge Gallery app on the Mac for the first time, allowing AI fans to run its Gemini large language models (LLMs) locally, without the need for an internet connection.
Apple's Overhauled Siri Will Reportedly Run on Nvidia's Blackwell Chips
Apple will rely on Google's fleet of Nvidia chips to power its overhauled version of Siri when it launches in September, according to a new report from The Information. Last week, the outlet reported that Apple plans to highlight the on-device AI capabilities of its devices at WWDC next week, but queries that require cloud-based processing will still fall back on one of Google's large Gemini models, as per an agreement between the two companies.
Google AI Edge Gallery launches to macOS
In addition to Google AI Edge Gallery, the company also released the Gemma 4 12B model and the Google AI Edge Eloquent dictation app for the Mac.
Alphabet's capital raise signals the end of low capex, high earnings business model: Expert
Blind Squirrel Macro's Rupert Mitchell hints at how investors may be underestimating a key transition underway when Big Tech moves to capital-intensive AI infrastructure operations. This move, the veteran capital market banker suggests could reshape the market dynamics that helped fuel U.S. equity outperformance over the past decade.
Researchers decode hidden patterns in World Cup data
Northeastern’s Network Science Institute is using data to decode the 2026 World Cup, from passing networks to predictive models.
Google's new Gemma 4 open AI model is sized for your laptop
Gemma 4 12B uses a new encoding scheme and token prediction to punch above its weight.
Why climate scientists have dropped a worst-case scenario
International scientists decided it is no longer useful to study the most extreme model of runaway global warming. So why are some people upset?
‘Galaxy Z Fold 8 Ultra’ confirmed, likely leaving Samsung’s most interesting foldable with a boring name
Samsung is planning two Galaxy Z Fold 8 models for launch this year, and one is confirmed to have “Ultra”...
These iPhones, iPads, Macs may not support Apple’s new software
iOS 27, iPadOS 27, and macOS 27 will be unveiled next week, but certain iPhone, Mac, and iPad models might not support the new software.
Microsoft and Mayo Clinic partner to build healthcare AI model
Mayo Clinic will own the model, which will be trained on de-identified patient data and initially deployed for use by its clinicians
Michigan’s rural health investment must be built to scale
Michigan’s rural health investment should not simply fund isolated success stories. It should accelerate the movement of proven models from one community to the next.
Pioneering AI Founders Join to Accelerate Roblox Reality Vision
Founders from Morpheus AI, Dynamics Lab, Lucid AI Make Playable Photorealistic Roblox World Models Possible
2026 Memorial Tournament odds, picks: Surprising predictions by golf model that's nailed 17 majors
Mike McClure's model has simulated the Memorial Tournament 2026 10,000 times and revealed PGA picks this week
Franklin Templeton says Wall Street fears blockchain because it threatens its profits
Jenny Johnson, Franklin Templeton's CEO, said blockchain and crypto threaten a huge number of business models that exist today in traditional finance.
Trump signs an executive order to vet top AI models for national security risks
The order establishes a framework for the federal government to vet the national security risks of the most advanced AI systems for up to a month before their public release.
Trump signs executive order to vet top AI models for national security risks
The move comes less than two weeks after Trump postponed a White House ceremony over concerns that a similar policy could dull America's edge on AI technology
Trump signs executive order to vet top AI models for national security risks
The move comes less than two weeks after Trump postponed a White House ceremony over concerns that a similar policy could dull America's edge on AI technology
Here’s how MacBook Neo sales compare to M5 MacBook Air and Pro
A new report explores MacBook Neo sales in its first few weeks compared to the also-new M5 MacBook Air and MacBook Pro models.
Motorola Edge (2026) for US and Canada unveiled with smaller 6.3" display
The 2024 and 2025 models had 6.6" and 6.7" displays, respectively. The new model has a 50MP 1/1.56" main camera and a 10MP 3x telephoto module.
How Science Works
The 2nd Weather and Climate Livestream is now underway. It is organized by a group of early-career researchers concerned about declining funding for the field and eager to demonstrate how weather and climate research makes a difference in our lives. Last year’s event was a great success, drawing terrific press coverage including a piece in the NY Times. This year features a strong lineup of talks aimed at a general audience — I highly recommend tuning in. Share I gave a talk yesterday about How Science Works in Three Steps. You can see the short presentation here (I’ve appended a transcript below): And please tune into the livestream here. It runs through tomorrow at 6 pm ET. Thanks for reading The Climate Brink! Subscribe for free to receive new posts and support our work. Subscribe Cleaned up transcript: Thank you, and thanks to everyone tuning in. I want to talk about something I think a lot of people don’t really understand: how science actually works. Arguments over any scientific topic — climate change, vaccines, you name it — revolve around competing claims. For climate change, the claims are: Is the Earth warming? Are humans to blame? What will the impacts be? Over the last 150 years, the scientific community has developed a remarkably robust method for sorting out which claims are true and which aren’t. That method is what I want to walk through. So how does science generate knowledge? If you ask a high schooler, they’ll give you the version they learned in ninth grade: an individual scientist forms a hypothesis, runs an experiment, and draws a conclusion. That’s more or less what happens at the individual level, though in practice it’s more iterative — you design, test, and refine all at once, landing on a slightly different idea each time until you reach a conclusion about whatever hypothesis you’ve ended up with. But this isn’t science. It’s only step one. Any individual scientist can make mistakes — scientists are human. Maybe there’s a bug in the code, maybe there’s an error in the experiment. Just because someone ran an experiment and reached a conclusion doesn’t mean it’s right, and scientists don’t treat a single result as settled truth. Step two is peer review, the first layer of quality control. The scientist writes up what they did and submits it to a journal, which sends the paper to other experts in the field. Those experts judge whether the methods are sound, whether it properly references prior work, and whether the conclusions actually follow from the data. Ask any scientist and they’ll tell you peer review is miserable — the comments come back brutal, often harsh, and you have to revise and resubmit, sometimes multiple times, before an editor finally accepts the paper. Peer review catches an immense amount of bad science, but it can’t catch all of it. Reviewers get sloppy or overworked, they miss obvious mistakes, and they can’t catch hidden errors — a bug in the code, a misread instrument. So a peer-reviewed paper making it into a journal doesn’t mean the paper is right. Nobody treats a single paper as the final word. That’s where step three comes in: the crucible of science. Important results get retested by the broader community. If you publish something interesting, it gets dissected. Sometimes people try to reproduce your work directly but more often they test its implications. If someone claims X is true, and that implies that Y is also true, then scientists will go check Y. If a claim survives this gauntlet long enough, it becomes accepted. That’s how we generate knowledge. Let me give an example: the ozone hole. In 1985 — and I doubt many people watching were around then, but I remember it well — Joe Farman and his colleagues published observations from Antarctica showing that a huge chunk of ozone was disappearing every year. This was shocking because most atmospheric chemists had expected ozone loss to occur at mid-latitudes and high altitudes, not over the South Pole. When scientists see something they don’t understand, they rush to explain it, and several competing hypotheses quickly appeared in the peer-reviewed literature. One pointed to chlorine chemistry from chlorofluorocarbons. A second blamed nitrogen chemistry, possibly driven by solar proton events — meaning the sun was responsible. A third argued it wasn’t chemistry at all but dynamics: winds were carrying naturally low-ozone air from the lower atmosphere up into the stratosphere. Then the community went to work testing which theory could survive the crucible. Aircraft flew into the ozone hole and found no nitrogen, ruling out the nitrogen theory. Measurements showed the air in the polar vortex was descending, not rising — which would bring high-ozone air down, eliminating the dynamics theory. Meanwhile, those same flights measured enormous amounts of chlorine, strongly correlated with exactly the regions where ozone was being destroyed. That became known as the “smoking gun” paper, and it established that chlorine chemistry was the cause. Through this process, a consensus emerged. And it’s worth being clear that this consensus is organic. There’s no vote, no poll, no meeting where it gets decided. Scientists go to conferences, talk to one another, and independently arrived at the same answer. It’s the ultimate free market of ideas. After the consensus was established, the peer-reviewed literature doesn’t even mention the rival theories — it simply takes chlorine chemistry as given. That’s what a consensus looks like. It doesn’t mean every single scientist agrees, but it does mean there’s no longer any legitimate debate. So where can you see what scientists actually think? It’s all in the peer-reviewed literature, which contains the entire historical record of the argument: the discovery, the competing theories, the debate, the testing, and finally the community coalescing around one answer. But there’s a lot of room for misrepresentation. You can dig up an old paper claiming the ozone hole was caused by dynamics and wave it around as if there still was a debate about the cause of the ozone hole — without mentioning that the idea was disproven long ago and that even its own authors no longer believe it. There’s a bigger problem, though: the peer-reviewed literature is written in a language I call “nerd.” It’s extremely hard to read — honestly, some of it is hard for me to follow. It takes graduate students years to work through the literature in a single field, so we can’t expect policymakers or ordinary people to do it. This is where scientific assessments come in. The most famous is the IPCC — the Intergovernmental Panel on Climate Change. The job of an assessment is to read and summarize the peer-reviewed literature, synthesizing what’s settled, what’s been disproven, and what remains genuinely debated. To ensure the assessments accurately summarize the peer-reviewed literature, assessments are written by large teams, which are far less prone to bias than individuals and which dilute the influence of any one scientist with an axe to grind. The teams are deliberately assembled to avoid stacking the deck, and the reports themselves go through multiple rounds of peer review. That rigorous process is exactly why you can trust them and why people consider the IPCC to be the gold standard of what we know about the climate system. An assessment produced through a different process may well not be credible. You can even watch the consensus evolve across assessments. Take the attribution of warming. The 1990 IPCC report said the observed warming was broadly consistent with model predictions but also comparable to natural variability — in other words, we don’t know whether humans are responsible. By 1995, it concluded that the balance of evidence suggested a discernible human influence — a sentence only a committee could write, and one that barely commits to anything. By 2001, most of the warming over the previous 50 years was judged “likely” due to greenhouse gases; by 2007, “very likely”; by 2013, “extremely likely.” The 2023 report says essentially the same as 2013 — there’s not much stronger language available. So, to wrap up: science is our ultimate tool for deciphering the universe — the most successful method we’ve ever developed for determining what’s true and what isn’t. Our understanding is written into the peer-reviewed literature, so if you want to know what scientists think, that’s where to look. And for those of us who don’t speak nerd, scientific assessments are the Rosetta Stone — they translate nerd into English. Let me open the floor to questions. Thank you.
How Nespresso Builds Sustainability Into Its Business Model | MIT Sloan Management Review
Nespresso North America’s CEO explains how responsible agricultural practices help ensure coffee quality.
Are plants the key to solving energy and food crises worldwide?
Costas Maranas, Robert V. & Gloria H. Waltemeyer Chair and Donald B. Broughton Professor of Chemical Engineering at Penn State, is developing computer models and algorithms to help scientists better understand, analyze and redesign biological organisms, including plants. In this Q&A, Maranas discussed how recent breakthroughs in modeling, artificial intelligence and systems biology are accelerating their work, helping plant scientists better understand and repurpose plants into fuel, plastics and more.
Science and AI-enabled research platforms
In an Editorial, Vaughan C. Turekian, Executive Director of the Office of International Networks, Cooperation, and Security at the National Academies of Sciences, Engineering, and Medicine argues that science is entering a new phase. In an era of vast computational power, automated experiments, and AI, science will increasingly be built around integrated platforms that link data, models, computing, and instrumentation at scale.
Turbo-charging battery research with AI: An ambitious vision
Argonne battery researchers and data scientists have come together to review emerging large language model applications in the battery field, outlining an ambitious technical roadmap to accelerate battery breakthroughs.
iPhone 18 Pro Battery Capacities Allegedly Leaked
Battery capacities for Apple's upcoming iPhone 18 Pro have allegedly surfaced, and the numbers suggest only a modest increase over the iPhone 17 Pro. According to prolific Weibo-based leaker Digital Chat Station, Apple is testing the iPhone 18 Pro with different battery capacities for the China and U.S. versions of the device, similar to last year's iPhone 17 Pro models. The Chinese model is said to have a roughly 4,056 mAh battery, while the U.S.
New iPhone 18 Pro battery leak just changed the story for non-US buyers
iPhone 18 Pro's battery capacity has been leaked at 4,056 mAh for physical SIM models and 4,288 mAh for eSIM variants. The bigger story: Apple's eSIM expansion to Europe means millions of buyers will get the larger cell for the first time.
AI Boosting Business Formation
The surge in new US business formation is being fueled by AI and large language models, which are dramatically reducing the cost and complexity of launching a company, see chart below. Subscribe for daily updates.
These are the best pill organizers for every need, according to healthcare experts
I spoke with healthcare experts and editors to find the best pill organizers. These models are designed for travel, seniors, supplements and prescriptions.
Gigabyte Introduces the AORUS ELITE Series Gaming Monitors with OLED or Mini LED Panels
Gigabyte, the world's leading computer brand, announced its AORUS ELITE Series monitor lineup, bringing together next-generation Tandem OLED technology and the world's first 5K Multi Mode Mini LED glossy gaming monitor across 27 to 32-inch models. The new series combines premium visual quality, excl...
Business News June 2026
• The Cornell School celebrates 10th anniversary! • “School’s Out for Summer “ market at Bethany Village • Cedar Mill Farmers Market • End of School Block Party • Speech & Debate summer camp • BodElite Introduces Emsella for pelvic health • Buying a Home? What to Know Before You Remodel • How CWS sets rates for your community • Sunset Credit Union news • New Bottle Drop at Target store
Apple Releases iOS 26.5.1 to Fix Charging Issue on iPhone Air and iPhone 17 Models
Apple today released iOS 26.5.1, a minor update to iOS 26. The software is available three weeks after iOS 26.5 came out, and appears to only be available for the iPhone Air and all models in the iPhone 17 lineup. The new software can be downloaded on eligible iPhones over-the-air by going to Settings > General > Software Update. According to Apple's release notes, the update fixes a previously documented charging issue with iPhone Air and iPhone 17 models.
Apple releases iOS 26.5.1 for iPhone, here’s what’s new
Apple has just released a new software update for iPhone users: iOS 26.5.1, which fixes a bug impacting models in the iPhone 17 family.
New AI Inference Service Now Ready for Science at Argonne
AI can help accelerate scientific discovery, but setting up and running a foundation model is not a simple task. Thanks to the work of the Argonne Leadership Computing Facility, scientists affiliated with DOE National Labs and the Genesis Mission can now tap into a new AI inference service running on ALCF supercomputers. Dubbed the ALCF […]