SCROLL NEWS / DISCOVERY
Search headlines
Find the stories shaping the conversation.
Results for " zero " 542 found 🔀 AI-powered Shuffle
Politics: The Midterm Races You Aren’t Hearing About
In just a dozen days, voters across the country will cast their ballots in midterm elections that could shift the balance of power in Washington. A lot of focus has been devoted to the contested races for Congress, with billions of dollars being spent on critical statewide contests and referendums.From governors’ mansions to statehouses to local ballot initiatives, voters will make their voices heard.We gather a panel of experts and zero in on some midterm races that aren’t necessarily making national headlines.Find more of our programs online. Listen to 1A sponsor-free by signing up for 1A+ at plus.npr.org/the1a.
Zero Republican political science professors found at 4 major Pennsylvania universities, research shows
(The College Fix) — Professors registered as Democrats outnumber colleagues registered as Republicans by a 28 to one ratio in […]
Is It Wrong to Try to Pick Up Girls in a Dungeon? Frontier Zero announced for PC, iOS, and Android
enish has announced Is It Wrong to Try to Pick Up Girls in a Dungeon? Frontier Zero, a new game based on the anime series Is It Wrong to Try to Pick Up Girls in a Dungeon? It will launch for PC, iO…
Phantom Blade Zero launch trailer, PS5 physical editions launch in 2027
S-GAME has released the launch trailer for Wuxia action RPG Phantom Blade Zero, as well as announced that physical standard, Deluxe, and Ultimate editions will launch for PlayStation 5 in 2027.
Three Six Zero Launches Entertainment & New Media Division Led by Rebecca Dowell
Three Six Zero announced a new Entertainment & New Media division led by Rebecca Dowell, who will represent digital creators and emerging talent including…
Three Six Zero Launching Entertainment Biz Under Rebecca Dowell
Three Six Zero is launching a New Media and Entertainment biz under Rebecca Dowell, rep of Amelia Dimoldenberg and Munya Chawawa.
Zero Republican political science professors found at 4 major Pennsylvania universities, research finds
ANALYSIS: Democrat professors outnumber Republican colleagues 28 to 1 in Pennsylvania political science departments Professors registered as Democrats outnumber colleagues registered as Republicans by…
PHP RCE Zero-Days CVE-2026-109618, CVE-2026-109619 Leak Early
CVE-2026-109618 and CVE-2026-109619: watchTowr confirms leaked CVSS 10 PHP RCE zero-days. Patches due October 22; only PHP 8 gets fixes.
Sundar Pichai: Google’s Frontier Imperative, Compute Scarcity Dilemmas, Future AI Products and Businesses
Alphabet CEO Sundar Pichai has little patience for doubts about Google’s willingness or capacity to compete at the frontier of AI. His company’s latest Gemini 4 Argon model — long-promised and recently delivered — beats OpenAI and Anthropic’s latest public models on coding, math, science, and knowledge-work benchmarks and has placed Google’s AI back among the world’s best. Amid chatter that Google was fine with good-enough lightweight ‘Flash’ models, Pichai pushed back hard against any suggestion that company is limiting its research ambitions. “We won’t do it any other way,” he told me in a Big Technology Podcast interview recorded this week at Google’s Mountain View headquarters. Pursuing cutting-edge research is costly for Google, requiring compute and chips that otherwise could be sold at a high markup by its cloud business. But Pichai said allocating resources to frontier AI development is “ground zero” for Google. “Our cloud business is at any given moment in time forgoing short-term revenues because we are making compute trade-offs to support frontier AI development,” Pichai told me. In our conversation, we discussed Argon (which Pichai says processed three times as many tokens as Gemini 3 in the 72 hours after launch), AI’s role in developing better models, the AI interface of the future, competition with OpenAI, Google’s partnerships with Anthropic and Apple, data centers in space, and the nature of his optimism. You can read the full conversation, lightly edited for length and clarity, below, or listen to it on Apple Podcasts, Spotify, YouTube, or your podcast app of choice. Alex Kantrowitz: Google is back at the frontier with Gemini 4 Argon, which beats OpenAI’s and Anthropic’s public models on benchmarks spanning coding, math, science, and knowledge work. The conventional wisdom was that Google had given up on the frontier to focus on fast, good-enough models for its products. That was wrong? Sundar Pichai: Very wrong. Look, we are so committed to being at the frontier. It’s in our DNA. We won’t do it any other way. We’ve been thinking about the AI frontier for well over a decade now. The frontier is very dynamic. It’s very competitive. It’s very exciting. I couldn’t be more pleased with how Argon has come together. We took our time with it. We made sure we tested it safely and deployed it carefully. We took a bit longer than our other models, given its capabilities. Across the board, be it for agentic work, knowledge work, coding, creative tasks, Argon is one of the really good frontier models. The first 72 hours since we launched, it’s almost up to 25 trillion tokens used, which is about three times what Gemini 3.0 had in the first 72 hours. And we are capacity constrained as we are serving it. Excited, but there are always areas….. You know, the frontier is very dynamic, and you have to keep evolving the models. There are areas which we can see we need to improve as well. We’ll be at work on all of that as well. Which areas? On any given model, Argon would be across the board better at everything if you just look even two months back. But the space is so evolving. Even on coding, for example, there are pockets where we know we need to do better, and there’s clear feedback. We’ll make it better. The argument for standard models is that they’re already good enough to build great products, with the frontier valuable only for certain tasks. Meta built Muse with good models rather than the best, and Google could put standard models into Maps, Gmail, and Search. Why isn’t that sufficient? It’s definitely not how I have thought about it at all. We are deeply committed to driving the state of the art and actually pushing where this technology can go, all the way from transformers. That’s a journey we’ve been on. You will see us continue to do that. When you’re at the frontier, you get an earlier look into the emerging capabilities, earlier than others, which helps you better prepare for the future. We are fortunate to have a wide set of opportunities across consumer and enterprise. Some of our cloud customers, for sure, want the model at the frontier. The frontier models will be better at developing the next set of models in an era of recursive self-improvement. I would argue it matters there. Internally, we have definitely seen a steep point of inflection by deploying Argon to our developers. It definitely tackles problems which our previous models could not do. Being at the frontier puts you at the innovation frontier and helps you drive that as well. There is room for models of all sizes and capabilities. That’s why we did Flash, Flash-Lite, etc. I agree with you that you can develop very good product experiences with not necessarily the most frontier model, because they tend to be slower and more expensive too. But we care about the Pareto frontier, and we’re committed to doing all of that. You said developing frontier models gives you a look at what’s coming. What do you see on the horizon? What makes this moment so exciting is you build something which you think is a culmination of months of work, and you put it out. You already see clear headroom ahead. We’ve all been working on these agentic, long-horizon capabilities, but clearly the frontier is going to have multiple of these agents working together to do things. That’s an example of an exciting paradigm underway. It’s also clear to me that users, both in consumer and enterprise, will have access to some multi-agent framework for them, working all the time doing things. It feels like a moment of infinite possibilities. Sergey Brin said one of Argon’s core innovations was relying more on the model to help develop the model. Yet incidents involving OpenAI’s bots breaking containment and Google’s bots accessing a site they shouldn’t have suggest these systems still aren’t perfectly aligned. Isn’t there reasonable cause for concern about accelerating toward recursive self-improvement? There are three parts to your question. Let’s first talk about RSI itself. You’re right, Sergey mentioned RSI being a critical contributor as we are developing Argon. But to be very clear, today, in a continuum, we are pretty far from the end state of RSI or something like that. The right way to describe it is you have RSI loops along with AI researchers in various aspects of model development. You almost have agents helping AI researchers in some automated way work together to build these models. It’s good to at least understand where we are today, and it’s super exciting. I expect that continuum to keep making progress as we go into next year. The second question is around safety. Security comes from defense in depth. You’re talking about alignment, obviously. We all are doing a lot of work to make sure the models are as aligned as possible. But it’s an active area of research, and there are clear gaps. That’s not the only thing we rely on. You have the work you do on alignment, but beyond that, we put a lot of effort into real-time monitoring as these models are working. Making sure you have good telemetry into what’s happening so that you can intervene in real time if you see unexpected behaviors. How you do that, the architecture with which you do that, how you make sure it’s all sandboxed and isolated in a way that it can continue doing its job during real-time monitoring — these are all areas of investment. Obviously, when we deploy these models, giving customers, enterprises, control, both with real-time monitoring, but their ability to decide where they deploy, what type of data which models ever have access to, etc. All that work is very, very important, and we are doing a lot of that. As an industry, we need to constantly up our game and take the time to do all this stuff, including dedicating the compute required to make sure you can safely deploy these models. The third part is: Is it all enough? This is why there are a lot of conversations in the industry, and I think it’s reasonable to be worried. The right way to approach something you worry about is early on thinking about what more we can do. As an industry, you’re definitely seeing increased levels of work on alignment, testing for safety, etc., as we deploy these models. The next step would be developing better standards so there are cross-industry standards of how we do these things, including the ability to evaluate using independent evaluators. If you look at areas like financial controls, we have accounting standards, we have independent auditors. Those are all the next natural steps for us as an industry to work toward, and regulation will end up playing a role as we go through all of that as well. Google pursues both language models and world models, while some other AI companies like OpenAI have focused more heavily on LLMs as the path to AGI. Why is world-model research important to you? As humans, you obviously perceive the world around you, and you’re taking all those inputs. Making sure these models are able to understand the laws of physics, simulate the world in high fidelity. All that, I think, are key enabling blocks for intelligence. We are working across all possible applications of this. Take an area like robotics. It’s not an accident: If you look at embodied reasoning benchmarks, Argon is state of the art. That all comes from our experience building world models. Inherently, I would expect there will be architectures which will converge all of this in the future. But for us, we view this as, today, this jagged frontier that is intelligence. We want to be at the frontier of all the possible approaches, and that will help us both drive better research, better insights, and develop a much better AGI model, is how we think about it. Some people argue Google should focus on selling its computing resources rather than using them for AI research. Cloud wants to sell compute at a high markup, while researchers need a very big computer to build the best models. How do you balance those demands? Ground zero for us is making sure we are investing in the compute that’s needed to be at the AI frontier. In 2015, I set the company to be AI-first. AI is that key horizontal ingredient across everything we do, all the way from Search to Cloud to Waymo. That is ground zero for us, making sure both for research and developing the models that we have enough compute needed. Beyond that, we have these independent businesses which each have their own long-range plans. They’re using those long-range plans to project forward, and we plan for compute for all of them. The cloud team, they have a business and they have a projection and they have demand signals. They have committed contracts. With that help, we can plan on a long-term compute trajectory for them. We’re fortunate to have this wide set of opportunities. There’s no company in the world which isn’t making trade-offs at any given moment in time. But the first principles we operate on is making sure we have enough compute to invest in the frontier, and that’s non-negotiable, and everything else works on top. Remember, we are building our own compute, our own data centers. There are long lead times. The compute we have today is a function of decisions I made in 2024. Even if you think back to a year from today, last year the agentic coding takeoff hadn’t happened. Today, obviously, when you sit and look at anything, everyone is short—not just us, other companies—and everyone wishes you had two times what you had planned for. That comes with the territory. At any given moment in time, we are planning for what we would need two to three years out in the future. You’re going to find some decisions you make may look suboptimal two years out, depending on where the state of the art is. I don’t spend too much time thinking about it. The way I think about compute allocation is making sure Google DeepMind, both for research and frontier model development, is cutting edge in terms of compute access. After that, we plan for all our businesses. It’s great we can provide our compute to other people too, like we have done. It allows us to scale that up, get real tests for it in the widest application set possible, test it in the real world, and make sure we are at the cutting edge of technology at every layer of our AI stack as well. Ben Horowitz relayed on a podcast that Andy Jassy had said compute was sold out until 2028. Are you seeing the same thing? If today a customer is coming out and asking us for compute in ’27, that is incredibly difficult to do. Even last year we went out and leased some short-term compute to accommodate some customers and help them out. But yes, we are all dealing with similar timelines. Every major technology shift seems to involve overbuilding, from railroads to fiber. Do you think that’s happening here? It depends on the time frames you’re looking at. This level of build, if I take a long enough time frame, it makes complete sense to me. We are at early stages of the most consequential technology ever, and this is not like fiber. It’s not dark. We are using it. We are capacity constrained on a daily basis. The applications, the underlying technology is improving rapidly. But running a company, the way I work through the moment is you look at your demand signals. You look at your committed revenue opportunities that you’re getting in the form of contracts. You balance that with the underlying technical capability, or internally you balance how we are using it internally, the model trajectories. Based on all that, you take a disciplined approach—a confident, aggressive, but disciplined approach—to plan out compute. But when you talk about the broader economy as a whole, you’re having people operate with various levels of risk. I focus on what I can control as a company and plan through it. If your question is, can there be an overbuild? Yes, because you can have some constrained components. Remember, what matters is a set of everything you need to build out that compute. Memory, electricity, electricians…. Land. Like 2,000 components or more. You could get constrained at some point. There could be things which cause bottlenecks, and then you could have overbuilds. There could be external shocks to the economy. There are macro risks. All that is possible at any given time. But am I confident that the compute we are planning for, we will have uses for it given enough time? Absolutely, yes. I was in the room when you said AI would be as profound as fire and electricity. If you believe that, shouldn’t you give researchers the biggest possible computer, even at the expense of today’s cloud profits? We’re all doing some form of that. Our cloud business is at any given moment in time forgoing short-term revenues because we are making compute trade-offs to support frontier AI development. What you’re seeing is true already. But you want to be thoughtful. Even our AI researchers have some constraints, which are also good. That’s what inspires creativity. That’s what helps you come up with a more efficient algorithmic breakthrough, maybe to use less memory. What we are committed to is to invest what’s needed to stay at the frontier. OpenAI President Greg Brockman told me a few months ago that the ideal AI product might have no interface at all—just as natural as conversation. What do you think the interface looks like in the age of AI? Computing historically has been humans adapting to computing. With AI, it’ll be the other way around. The technology will adapt seamlessly to how humans live and go about their day-to-day life. You want to be able to speak to that AI when you want to. It has context. It can be persistently running somewhere, give you continuity. You’re not dealing with a lot of complexity. It’s natural. All that makes sense to me. That doesn’t mean it won’t have any interface, however. Nowadays, even on my phone, Gemini voice models are so good. I can just press the microphone button. I can just ramble. That’s why we call it Rambler. I can ramble. It intelligently corrects with my personal context. Even my friends’ names, weird names, it gets it right, even if it first guesses it wrong. It’s amazing. I’m not going in, backspacing, editing once I finish saying it. It’s remarkable how good it’s gotten. But there are going to be moments you will need an interface. You may want to entertain yourself. You may want to watch a video. You may ask your AI to create something, and you want to watch it, consume it. The way I think about it is, you want it to be as seamless and as natural as possible, and that you are not doing the hard work adapting to whatever the interface is. The technology is adapting to how you’re working at any given moment in time. When I asked Sam Altman about competition from Google, he argued that OpenAI was building AI-native products and implied Google was bolting AI onto existing ones. Do you accept that framing, and can a “bolt-on” compete with an AI-native product? First of all, it’s a very dynamic moment. We are all pushing each other. OpenAI has pushed us as a company too. This is not just we are doing something to them. Far from it. Yes, Google went into its own version of a code red too. In the AI moment, all of us have to respond with varying levels of urgency at any given moment in time. We are constantly building new products as well. Gemini was built from the ground up as a product—the Gemini app I’m talking about. Or tomorrow, as all of us are adapting to agents, are you bolting agents onto ChatGPT, or are they building it natively? How would you think about that? You can always put these questions back. I don’t think of it that way. I think of it as you want to build world-class product experiences that are AI-native and they are winning. Some of it is adapting products, some of it is rethinking new experiences. They will all be part of how you approach this moment. How do you expect to make money from these AI interfaces? If an agent recommends a product and I buy it, should it take a cut of the sale, even if that’s disclosed? I’m going to put enterprise aside. Consumers are going to increasingly value this technology. Take software developers today. Developers are willing to pay for these coding models because it really helps make them more productive. Nobody would want to code the way they were coding three years ago, or even the same time last year. When you’re delivering value, there’s value to be had. I’ve never seen examples where consumer products really scale to billions of users, where people are using it intensely, and haven’t had monetization to go along with it. If you’re providing value to people, there is inherent value to be had. The nature of it may change. You’re asking a scenario in which agents are doing stuff for you. At the end of the day, you will have commercial needs. The agents need to fulfill those commercial needs. The providers on the other side are going to see value in that. There’s going to be some mechanism by which it gets fleshed out. That’s what an efficient system does. But what matters is that you’re building those world-class consumer experiences. I promised myself one product feature request: Can you make Gemini-powered search in Gmail as good as searching Gmail through ChatGPT? What’s going on here? One of the exciting things we announced recently is the unified Gemini agent for enterprise, and we’re bringing that to consumers. These are persistent agents which run 24/7 in the cloud. We’re spinning up a micro VM effectively for every query, and from the runtime, the harness, the model—it’s all a headless service. It’ll work everywhere, and it has your context with you. You can address this agent whether you’re in Workspace, in Search, in Gemini. Agents are going to drive this unification from a customer experience, which people are looking for. When you have a broad product portfolio and you’re moving forward fast, there are always areas where you’re moving forward faster than some other areas. That’s valid feedback. But we are putting a lot of time and effort into making sure, as we build out our agentic experience, that unification and that capability works well across all these services. People have asked where Google’s equivalent of Muse or Dots is. Do you anticipate creating a new product to compete with them? Our goal with Gemini is to be a personal, proactive, and powerful assistant. That’s what we are working toward. You saw us describe the early versions of it at Google I/O this year with Spark. Effectively, already, if you open the Gemini app as a Pro or an Ultra user, you see Chat and Task, and they start giving you that agentic experience. With Argon now, that capability is dramatically accelerated. What we announced recently in our enterprise event, we are bringing that to consumers as well. Soon in Gemini, you will be able to ask something, and it’s a continuum. You don’t need to think about whether you’re asking a chat or an agentic task. It’ll just automatically do the right thing, and it can work persistently in the background securely and get things done for you. Already, empirically, when I look at what’s coming into Gemini in our app, last year you would have tagged them and most of them would have been chats. But a steadily growing percentage is already agentic tasks in Gemini. Yes, we are going to bring personal agents to our consumers. It’ll work in the Gemini app. It’ll work in Google Search. It’ll work in Gmail. You can address it, call it, etc. We want to make sure we’re getting it right. It’ll be powered by Argon, and we are very excited to bring it to our users. Google has invested and committed tens of billions of dollars to Anthropic. What do you see in Dario Amodei that makes you want such a deep investment and partnership? Very early on, when we looked at the founding team of Anthropic, it was very clear to us this was a world-class team which was involved in some of the earliest GPT models. We were quickly able to assess that they would be at the frontier. We’ve always tried to identify and partner with companies which would be at the cutting edge of technology. We’ve done that with SpaceX. We’ve done that with Anthropic. That’s what we saw, and we had a chance to partner. Is your partnership with Apple to license Gemini for Siri expanding, contracting, or standing still? For both us and Apple, that partnership was about translating the capabilities of AI into better experiences for users. I thought Apple has done a phenomenal job of incorporating—they customize these models and build it into Siri—and it’s been great to see the reception. We are committed to doing more. You’re going to do more with them? There’s nothing specific to comment, but it’s a partnership. It’s working well. What have you learned so far about the feasibility of data centers in space? It’s a very exciting project. Projects like this remind me—I still recall that moment in the desert for the DARPA Grand Challenge, watching the Stanley car compete in it, which ended up being the basis for Waymo. With all these technologies, you’re at a moment in time where it looks a bit like sci-fi, but 10 to 20 years down the line, it’s actually working in a meaningful way. I feel the same in space. We’ve always believed in betting big on these technologies. Quantum has been a great example of it. We’re probably 10 years into the journey. Waymo, we are almost 20 years into the journey. Space, we are two to three years into the journey. It’s clear to me that it’s an approach which will make a lot of sense in the future. Being in sub-synchronous orbits with 24/7 access to the most abundant form of energy we have access to will make a lot of sense for this exponential increase in compute that we are going to need. Having said that, this is a challenging project. The team has disciplined milestones. It was a super exciting milestone to send the TPUs up to space. Sometime, as we are speaking, we are expecting a “hello world” moment from them. I look forward to seeing that happen, and we’ll continue investing here too. You previously said 25% of the code written within Google was written by AI. What’s the number today? That number in our world today is meaningless in the sense that every part of what we now touch is influenced by AI deeply. What we are tracking with Argon, for example, it’s been a point of inflection. People are using Argon to really go tackle complex projects, including projects which affect millions of lines of code. In the four weeks since we have launched Argon, we have seen a 30% increase in CLs, and more importantly, almost a similar increase—a one-third decrease in rollbacks. People are writing code more productively and better code with it. We are tracking it with newer metrics now. How are you dealing with opposition to data centers? If you can’t build in the U.S. the way you want to, where else do you go? We’re very committed to building our data centers very responsibly. When I look at our largest data centers in some of the communities where we have been the longest, I visit them, I talk to people. There are people who come and tell us their kids have gone to college because of our investments in these communities, and we’ve grown in those communities. To me, this is about doing it right. Engaging with the communities early, doing it transparently. We are committed to making sure we are bringing incremental power, and we cover 100% of the cost of our power needs for these projects. It’s understanding the impact on neighborhoods and making sure you’re working thoroughly to address them. Done rightly, you can continue making progress. But we are constantly taking feedback into account. I do worry not everyone operates the same way. People have been surprised by data centers showing up in their backyard. Those are all important. As an industry, we should continue to do better. But part of the reason, in a 20-year time frame, we are thinking about things like data centers in space is making sure we are able to put these data centers where maybe the trade-offs will be different. What unproven area of AI is Google working hardest to crack? Is it continual learning—getting models to keep updating after release—or something else? There are multiple parallel paths. You saw with Argon we have a million output tokens. With our Gemini models early on, we had a one-million context window on the input side. These are all examples of breakthroughs we are constantly working on. How do you think about synthetic data? How do you think about continual learning? How do you think about more efficient architectures? We do dedicate a meaningful percentage of our compute to these research bets too. It always goes hand in hand, including our world models. Those are all bets we are placing. Sometimes, as part of these bets, there are ideas that emerge which just make their way into the mainline models. Maybe we don’t fully talk about it, but we have multiple steps in these areas. Working in technology requires optimism that building something will improve the world, but those bets can leave real consequences when they fail. I recently returned to a town in El Salvador where enthusiasm for Bitcoin had given way to unfinished construction and something resembling a ghost town. How do you maintain the optimism you spoke about in your Stanford commencement address while accounting for those risks? I genuinely have a deep conviction that improving technology and improving access to technology makes people’s lives better. It’s been my personal experience in my life, out of which comes that deep conviction, optimism. But with that, I’ve always felt humanity’s journey is always about working hard to harness that technology to make it work well for society. The optimism is not an optimism that you just build whatever you want, whenever. You work hard to make sure you’re building technology in the service of humanity, to make the human condition better. That’s got to be the defining goal as we are approaching it. With that framework, I’m very, very optimistic. We are in very early stages of the kind of progress that we are going to see on many areas. We’re all talking about areas like healthcare. A lot of the progress is ahead. The ability to make things in a fundamentally cheaper way on many things where things have been getting expensive for a long, long time, discovering new materials, etc., which really bend the curve—it’s tough to step today and understand the impact of all that, but they’re going to be very profound. I’m very optimistic, but we have to be rightfully concerned about all the ways this technology can go wrong. Sundar, I know you don’t do many of these. Thank you. Thanks, Alex. Appreciate it.
GM's core profit rises 22% on strength of pickup-truck business
GM shares fell about 6% in premarket trading.
UFC 333 Embedded, Episode 1: 'I just take their zeros'
The first episode of UFC 333 Embedded was released on Monday.
The RTX 5090 Isn’t 'Alive' if It Costs $10,000, No Matter What Nvidia Says
Nvidia's flagship GPU may still be in production, but that isn't saying much for the state of Nividia's RTX 50-series lineup.
The modal numeber of citations is zero (in political science)
I was told when I was a graduate student that the modal published research paper was cited zero times. They then when on to say that the median was one or two, while the mean depended on how good t…
Champion Teal Knight FNCS Cup Zero Build in NA West: Session 1
Answer the call. Fight for glory. Compete in the Champion Teal Knight FNCS Cup for a chance to earn Champion Teal Knight cosmetic rewards. Put your skills to the test, rise to the challenge, and forge your path to victory. Good luck and have fun!
Is Melco Resorts & Entertainment (MLCO) Undervalued Following Its REM Hotel Opening And Buyback?
Texas Capital Securities has zeroed in on Melco Resorts & Entertainment (MLCO) after highlighting drivers such as the REM hotel completion at City of Dreams Macau and a larger share repurchase program. Recent trading shows mixed momentum for Melco Resorts & Entertainment, with a 1-day share price return of 2.86% and a 7-day gain of 3.60%. These gains are offset by a 30-day share price decline of 6.90% and a 90-day drop of 20.73%, as the market weighs the impact of property disruptions against...
‘Close to zero’: 4 AI chiefs say no companies have truly changed in response to AI
At a Fortune event, executives from Accenture, AXA, Mastercard and Mistral AI said the technology isn't the main obstacle to AI payoffs.
Bills and Pills: How Congress and Science Could Make the Future Less Lonely
A Bill for Loneliness? Proposed legislation would fund more research and social support—and create an office to advise the President A decade ago, official Washington wasn’t thinking about America’s loneliness epidemic. “Congress was paying zero attention to this issue, the executive branch was paying zero attention to this issue,” says Andrew MacPherson, chair of the
Walmart to expand health benefits for employees, including $0 prescriptions, rare disease care
Walmart is expanding benefits for its 1.6 million U.S. | The new offerings include “zero cost” prescriptions, expanded Centers of Excellence for rare conditions and paid NICU leave.
Zero Escape And Hundred Line Director's New Game Canceled Before He Could Announce New Studio
Kotaro Uchikoshi planned to announce new studio Crash Keys alongside its first game
Kotaro Uchikoshi establishes new studio CRASH KEYS
Zero Escape and AI: The Somnium Files creator Kotaro Uchikoshi has announced the establishment of CRASH KEYS, a new game studio headquartered in Shibuya, Tokyo, Japan.
Ossoff again said he won’t run for president. Such denials don’t always stick.
Sen. Jon Ossoff (D-Georgia) said in a debate that he has “zero interest” in running for president. Plenty of other politicians said that and ran anyway.
Windows update patches a whopping 1,400 security bugs - including 2 zero days
This month’s update also marks the first official one for Windows 11 26H2 and the end of support for Windows 11 24H2 Home and Pro.
Microsoft October 2026 Patch Tuesday fixes 1,411 flaws, 3 zero-days
Today is Microsoft's October 2026 Patch Tuesday, with security updates addressing a record-breaking 1,411 flaws, including one actively exploited and two publicly disclosed zero-day vulnerabilities.
Northwestern QB Aidan Chiles Deserves More Respect
Aidan Chiles has 1,172 passing yards, 15 total touchdowns and zero turnovers through five games. Northwestern's quarterback is proving he belongs among the Big Ten's best.
New Doctoral Training Programme will tackle key polar science questions with a reduced carbon footprint
The Net Zero Polar Science (NZPS) Doctoral Training Programme (DTP) has welcomed its first cohort of PhD researchers, marking the launch of a new UK-based...
Jared Goff Makes NFL History During Detroit Lions’ Win Over Cardinals
Detroit Lions QB Jared Goff joins Russell Wilson as the only NFL quarterbacks with 12+ touchdowns and zero interceptions through five games.
Street Fighter Cup Zero Build Reload in NA Central: Session 1
Ready? Fight! Throw down in the Street Fighter Cup and battle for a chance to earn cosmetic rewards. Let’s go, World Warrior!
20 Minutes of New PC Gameplay Footage From Phantom Blade Zero
YouTube's Geek Culture Plays shared a video showcasing 20 minutes of gameplay from the PC version of Phantom Blade Zero.
Phantom Blade Zero's Fast-Paced Action Took 15 Years, Two RPG-Maker Games, And A Whole Lot Of Money To Bring To Life
S-Game is finally nearing the release of Phantom Blade Zero, and as part of previewing it one last time, I discovered this game has a whole lot of history
After 4 Hours, Phantom Blade Zero Just Became My Most-Anticipated Action Game
Get ready for the ‘kungfupunk’ era
Phantom Blade Zero hands-on preview
Six hours with Phantom Blade Zero gave us a deep look at its combat, weapons, exploration, story, and PC performance ahead of release.
I Played 6 Hours of Phantom Blade Zero and Can’t Stop Thinking About It
After half a day playing Phantom Blade Zero, it’s gone from mostly not on my radar to to the top of my wishlist this fall, and I truly cannot wait to dive back in later this month to see if it can deliver beyond those first handful of hours. At least for now, though, just about everything I’ve seen and played has made me more convinced that this is going to be one of the coolest action games of 2026.
Study finds simulation game helps students grasp complex health economics concepts
In Phillip Zane's classroom, students weigh an invitation to jump out of an airplane. The skydiving lesson and jump are free, and the exercise assumes there is zero risk of death. But an injury could ...
La Liga's new market values 🤑💸
That's a lot of zeros.Between the World Cup, the start of the season and the transfer windows, players' market values have been shaken up this summer.From new arrivals to dramatic drops and massive ri...
Star Wars Zero Company Studio Tries To Bring Back As Many Furloughed Devs 'As Possible' Following 'Unexpected Business Developments Outside Of Our Control'
Bit Reactor says it will have more to share on the XCOM-like strategy game's future 'soon'
Some furloughed Star Wars Zero Company developers are returning to work, studio says
Bit Reactor, the Maryland based company, behind this summer’s well-received tactical combat game Star Wars Zero Company says that “more” of its team is returning to work after being furloughed ahead of the game’s release. This is the studio’s first statement on the matter since Game File’s August report that “most” of the development team had been furloughed—technically kept as employees, but without pay or benefits—a few weeks prior to launch. Some of those furloughs ended this week. “This week, we started the process of returning more of our team back to work at Bit Reactor, an important step forward in our continued support for Zero Company, “ Bit Reactor CEO Greg Foertsch said in a statement. He continued: “Having to ask team members to step away after we navigated unexpected business developments outside of our control and so close to launch was an incredibly difficult decision, but one that was necessary to help us eventually bring as many people back as possible. “This team means a tremendous amount to me, and I am incredibly proud of what we have accomplished together. Thank you to our fans for the incredible response to the game and for supporting us along this journey. We look forward to sharing more updates regarding Bit Reactor and Zero Company soon.” Game File reported in August that Bit Reactor had retained a small team to work on updates for Zero Company. One person familiar with the situation told Game File that the returns announced today bring the studio’s pay-rolled headcount to more than half of what it had been ahead of the furloughs. That means a considerable number of people have been brought back, but that a fair number also remain without pay. While companies in the U.S. are often legally required to publicly disclose how many people they’ve laid off, furloughs can often be kept private, for a time. The United States Department of Labor requires companies to publicly report mass furloughs if they meet specific criteria, including any expectation that they will last more than six months. For those game makers who remain furloughed by Bit Reactor, it’s now been about two months. For more about Bit Reactor, the furloughs, the studio’s back-story and the ongoing lawsuit that brought this to light, read my August report. [ ‘Most’ of Star Wars Zero Company studio Bit Reactor was furloughed ahead of launch ](https://www.gamefile.news/p/star-wars-zero-company-bit-reactor-furloughs-lawsuit) Stephen Totilo · Sep 1 Last week was meant to be a moment of triumph for Bit Reactor, the Maryland-based game studio founded in 2021 with ambitions to create a great new tactical combat game. [ Read full story ](https://www.gamefile.news/p/star-wars-zero-company-bit-reactor-furloughs-lawsuit)
MLB playoff bracket: Games today, full playoff schedule
MLB's division series round is almost complete, with anywhere from zero to four teams able to advance with wins today. Print out your playoff bracket.
Telefónica is embedding climate action across its business to accelerate its path to net-zero emissions by 2040
Telefónica is embedding climate action across its business to accelerate its path to net-zero emissions by 2040
2026’s Biggest Blockbuster Games Just Received Their First Major Discounts in Amazon’s Prime Sale, but These 48-Hour Deals Won’t Last Long
Amazon's latest Prime Day sale features massive discounts on PS5, Switch 2, and Xbox titles, including deals on 2026's biggest games like The Blood of Dawnwalker, Star Wars: Zero Company, Pragmata, LEGO Batman, and more.
Fear is bigger than ever… and Zero Latency VR has the science to prove it
MELBOURNE, Australia — Zero Latency VR, the world’s largest and most innovative immersive VR network, is celebrating a double Halloween anniversary: three years since the launch of Outbreak, …
Star Wars Zero Company Has Received Its First Big Discount at Amazon Before Its October Prime Day Sale Begins, Dropping Below $50
Star Wars Zero Company has dropped to a new low price at Amazon ahead of the retailer's Prime Big Deal Days sale event, with a 20% discount across PS5 and Xbox Series X.
Ground Zero Blues Club
A Labor of Love for its Three Managers It is no mystery why Ground Zero Blues Club (GZBC) is one of the most famous blues clubs in the country. It is a genuine juke joint where customers are treated to great live music along with genuine hospitality serving up Southern food favorites like barbecue […]
How Americans’ Financial Health Is Faring In The “K-Shaped” Economy
Hey all, Jason here. Money20/20 is still two weeks away, but I’ve already started packing (not procrastinating for once!). My calendar is already filling up with various sessions I want to catch, meetings, and of course happy hours, dinners, and other side events. If you’ll be in Vegas and want to catch up, let me know by replying to this email, and we can try to find a time amidst the chaos of the Venetian. A big if not totally unexpected piece of news dropped on Friday: the Independent Community Bankers of America, a trade group that represents smaller U.S. banks, filed a lawsuit against the OCC and Comptroller Jonathan Gould, arguing that the regulator’s move to grant trust bank charters to firms seeking to use them to conduct substantial non-fiduciary activities exceeds the authorities granted to the OCC by Congress. I haven’t had time to fully digest the legal filing or speak to folks in my network about it, but you can expect to see coverage and analysis on it in next week’s newsletter. Subscribe or Support by Upgrading The NerdCon agenda just dropped. It’s stacked. Partner content: Nubank built an AI-first bank. Chime built its own banking core. Figure and Valon are rebuilding the mortgage stack. At NerdCon, you’ll hear from the people behind those bets: what they chose, what they learned, and what they’re still figuring out. The agenda is live, with leaders from Nubank, Chime, OpenAI, Remitly, Mercury, Figure, Valon and more. Follow the mainstage conversations, bring the problem your team is wrestling with to a hands-on workshop, or pull up a chair at a roundtable. Pick your quests. Meet us in San Diego, November 18–20. Fintech Business Weekly readers save 20% with code FBW20. Explore the Agenda Things To Know & Other Good Reads Federal Reserve Board finalizes changes to enhance the transparency and public accountability of its stress test and reduce volatility in its stress test-related capital requirements (Federal Reserve Board of Governors) FDIC Announces Conclusion of Independent Monitorship (FDIC) Modernizing Financial Regulation: Initial Observations from eSLR (Fed Vice Chair for Supervision Michelle Bowman) The Data Version of Godzilla versus Kong: FRED Takes on AI (Fed Governor Christopher Waller) Financing the AI buildout (Brookings) Q3’26: Rules for Banks but Not for Crypto (Fintech Takes Banking) The Inevitability of Local Stablecoins (Lombard Notes) Delusions of AI Governance: The Human-in-the-Loop Comfort Blanket (Fintech Snark Tank) Stripe agrees to acquire Parafin to expand revenue opportunities for platforms and help small businesses grow (Stripe) Listen: Fighting Fraud in the Age of AI, with SEON’s Tamas Kadar (Fintech Business Weekly) How Americans’ Financial Health Is Faring In The “K-Shaped” Economy If there is one overarching theme across the economy and politics in the U.S. at the moment, it’s “affordability.” The term is vague enough to encompass a panoply of factors shaping Americans’ day-to-day lives: rising costs for the hallmarks of being “middle class,” including healthcare, housing, childcare, and education, inflation and rising interest rates, and a job market with the specter of AI hovering over it. At the same time, policy shifts under the Trump administration have resulted in reduced subsidies to those who get health insurance under the Affordable Care Act and new restrictions on qualifying for Medicaid and SNAP, the impacts of which have yet to be fully felt. These factors are contributing to declining consumer confidence and general dissatisfaction with the economy, despite continuing GDP growth and stock market records. This divergence is encapsulated in the idea of the “K-shaped” economy, in which a small proportion of Americans have seen their wealth balloon, while the majority of American households struggle to preserve the lifestyle they have. Perhaps the greatest determinant of which branch of the “K” a household is on is whether their wealth and income are primarily derived from employment vs. from assets. Nearly two decades of low interest rates and, more recently, elevated rates of inflation have benefited asset owners, while those whose income is primarily or solely derived from labor have largely seen real purchasing power stagnate or decline. This is reflected in recent Bureau of Labor Statistics data showing that labor share of U.S. GDP — the fraction of economic output that accrues to workers as compensation in exchange for their labor — dropped to just 52.8% in Q2 2026, the lowest since the BLS began keeping records. In 1947, approximately 2/3rds of economic output accrued to workers; even as recently as 2001, labor’s share of GDP was 64.1%. Against this backdrop, we’ve seen the rise of more credit and credit-like products: cash advance apps, “no-fee” overdraft, earned wage access, and buy now pay later, which are often used as small-dollar short-term borrowing mechanisms to meet immediate consumption needs or pay other bills (eg utilities, cell phone, other debt payments). With the markers of a middle class existence and, increasingly, basic financial stability seemingly out of reach for many, it should be no surprise there’s been an increase in “financial nihilism,” something industry peers like Alex Johnson and Frank Rotman have discussed and analyzed the roots of. The explosion of gambling and gambling-adjacent products and services — often marketed under the guise of being an “investment” — is inextricably intertwined with the rise of financial nihilism. Crypto, sports betting, and, more recently, prediction markets offer an escape or even hope of sorts, like a contemporary, digital version of a scratch-off lottery ticket, while leaving the overwhelming number of people who use such products worse off. Share of Households That Are Financially “Vulnerable” Ticks Up The Financial Health Network’s 2026 Financial Health Pulse® report adds context on how American households are faring. (I linked to this report in last week’s newsletter, but wanted to take time to further unpack the data in the report this week.) The Financial Health Network publishes the report annually, providing insight into how Americans’ financial circumstances are changing over time. The 2026 report is derived from a survey fielded in April and May 2026. The report leverages survey responses to assess indicators of financial health across spending, saving, borrowing, and planning/protecting and to determine a zero to 100 “FinHealth” score. Those with scores between 0 and 39 are considered “Financially Vulnerable,” consumers with scores ranging between 40 and 79 are defined as “Financially Coping,” and those with scores of 80 to 100 are “Financially Healthy.” The report found that moderate improvements in 2025 were reversed, with the share of respondents considered financially “vulnerable” rebounding to 17%. The longitudinal data reflect the impact of pandemic-era programs, like expanded unemployment, cash stimulus payments, and pauses of federal student loan payments. The positive impacts of those programs, as reflected in the Financial Health Pulse reports, had largely disappeared by 2023. While pandemic-era inflation — to be fair, in part caused by the various support and stimulus measures — had come down from as high as 9% in 2022, it has rebounded since the start of Trump’s second term, with tariffs and energy market disruptions owing to the war in Iran pushing prices back up. Specific indicators in the survey that deteriorated from 2025 to 2026 include the share of respondents spending less than their income over the past 12 months, the share paying all bills on time over the past 12 months, the share that have a manageable amount or no debt, the share that are “moderately” or “very” confident their insurance is adequate to cover them in an emergency, and the share that “somewhat” or “strongly” agree that their household plans ahead financially. Unsurprisingly, lower- and moderate-income households are more likely to struggle to pay their bills on time and have any funds leftover to save. The share of low-income households whose income exceeded their spending dropped from 35% in 2025 to 31% in 2026, while the share of upper-income households whose income exceeded expenses remained unchanged at 63%. Households with student loans or revolving credit card debt reported having “a bit more” or “far more” debt than was manageable at rates far higher than those not carrying these types of debts. The share of those with student loan debt indicating their debt load was too high to manage increased from 50% in 2025 to 55% in 2026, while the share of those with revolving credit card debt saying the same ticked up from 51% in 2025 to 54% in 2026. Households that had student loans were markedly more likely to be financially “coping” or “vulnerable” vs. those without student loans. In 2026, 27% of households with student loans were considered financially vulnerable, a jump of 6% points vs. 2025. Households without student loans saw just a 1% point increase (not statistically significant) in those considered financially vulnerable. Households’ perception of the affordability of categories of goods and services paints a stark picture, with less than one-fourth considering higher education affordable, about a third deeming childcare affordable, and less than half viewing healthcare as affordable. Fintech Business Weekly is made possible by the generous support of paying subscribers — bringing you independent analysis of banking, fintech, and crypto without fear or favor. You can support my work by becoming a paying subscriber if you aren’t already. Paying subscribers enjoy: access to the entire archive of nearly six years of newsletters extended versions of the weekly newsletter, with additional content and analysis and (for founding member tier) quarterly personal 1:1 fintech Q&A / strategy calls with me, tapping into my unparalleled knowledge of the intersection of banking and fintech and experience working in the sector, including helping to launch Goldman Sachs’ retail bank Marcus Support Fintech Business Weekly You can also support Fintech Business Weekly by sponsoring a newsletter or podcast, putting you in front of 93,000+ of the most influential decision makers in banking, fintech, and crypto. Learn more about sponsor opportunities or request a media kit by dropping me an email. Democratic Senate Report Highlights How Iran Uses Tether to Evade U.S. Sanctions Iran uses stablecoins, specifically Tether, to evade sanctions, fund regional proxies, and to purchase military drones, a report released last week by the Senate Permanent Subcommittee on Investigations says. The report was authored by committee ranking member Senator Richard Blumenthal (D-CT) and the minority staff. Crypto and stablecoin proponents will often push back on criticism that criminals make use of these assets by arguing that, on an absolute dollar basis, far more illicit funds flow through traditional banking systems and payment rails than via crypto and stablecoins. And while that is likely true, it elides the reality presented in news story after news story: whether state actors — Russia, Iran, North Korea — or criminal groups, those looking to engage in illicit transactions increasingly favor stablecoins and, specifically, Tether (USDT). The report from the Senate Permanent Subcommittee on Investigations released last week adds context to how Iran uses Tether has a lynchpin in its shadow banking system. The report analyzed 846 wallets that have been sanctioned or targeted for seizure, finding that 84% of them transacted exclusively or primarily in Tether. The ability to create near limitless, anonymous wallets and to move funds effectively instantaneously and irreversibly have made crypto a favored financial mechanism for bad actors. Stablecoins solved key drawbacks of bitcoin, namely, the original cryptocurrency’s highly volatile value in dollars. And while Tether (the company) has the technical ability to “freeze” or destroy funds, the company is often limited and reactive in its approach to doing so, the Senate report argues. Tether, which is nominally based in crypto-friendly El Salvador, “has stated that its compliance with OFAC sanctions is ‘voluntary’ and that it follows ‘OFAC guidelines,’” the Senate report says. Owing to these favorable attributes, Tether “became a primary cryptocurrency for Iran, Hamas, Hizballah, and the Houthis beginning in 2023 and has expanded in scale since,” according to the report. Tether’s role in Iran’s shadow banking system, the report says, is enhanced by crypto exchanges like Bybit, Kyrrex, OKX, Gate, and Binance, and through “over the counter” exchanges and hawala networks. The consequences of these financial flows aren’t hypothetical. The Senate report links Tether as a funds transfer mechanism to Iranian proxies in the region, including Hizballah, the Houthis, and Hamas. Tether has also been used to make payments to secure components necessary to manufacture drones, the report says. The report concludes by arguing that “[s]tablecoin issuers with a significant nexus to the United States, particularly those that offer dollar-denominated stablecoins, should be subject to American sanctions law rather than being allowed to hide behind foreign jurisdictions.” Stablecoin issuers like Tether must be held accountable for repeated failures to prevent illicit finance and sanctions violations by law enforcement, the Department of Justice, the Securities and Exchange Commission, and OFAC, the report argues. [Paying Subscriber Exclusive] OUSD Goes Live, Make Your Own Neobank, U.S. Sanctions Russia’s A7 Network As “Transnational Criminal Organization” OpenUSD, the stablecoin issued by the Open Standard consortium that includes Stripe, Visa, Mastercard, core providers, crypto firms, and numerous banks, went live last week. Social commerce platform Whop raised eyebrows in fintech by offering creators on its platform the ability to launch their own neobanks, which the company describes as “great businesses that are easy to run,” in just 15 minutes. And OFAC and FinCEN target Russia’s “shadow banking” A7 Network — more on these stories after the paywall.
'Godzilla Minus Zero' review: The epic monster melodrama fans have been roaring for
Now with R-rated mayhem!
Patient-zero drill put health facilities to the test—40% of them failed
Actors with makeup and fake symptoms staggered into 73 healthcare facilities.
Bishop’s computer science chair says AI is eliminating ‘grunt work’ and reshaping digital jobs
Dr. Russell Butler, chair of Bishop’s University’s Department of Computer Science, says artificial intelligence is increasingly taking over routine digital “grunt work,” changing the skills students and workers may need in the years ahead. Photo courtesy of Bishop’s University. Click here to read our platform! Get local news like this every morning. Subscribe free below. Subscribe Pulse Newsroom Artificial intelligence is already taking over some of the routine digital work once performed by entry-level employees, according to Bishop’s University Computer Science chair Dr. Russell Butler — a shift he believes will make expertise, judgement and the ability to evaluate AI-generated work increasingly important. Butler described the change as automation reaching what he calls the “bottom of the cognitive pyramid” during an interview with The Pulse about AI, university education and the future of digital work. His argument is not that humans are disappearing from the workplace. Instead, he sees their role changing as AI becomes capable of handling more of the repetitive tasks involved in computer-based work. “Everything’s converging on just AI basically doing all the grunt work,” Butler said. The bottom of the cognitive pyramid Butler is from Lennoxville, completed his undergraduate studies at Bishop’s and later earned his PhD at the Université de Sherbrooke. He joined Bishop’s in 2019 and now chairs its Department of Computer Science. He uses AI in research, teaching and other creative work, and said the technology is increasingly capable of handling routine digital tasks that once required substantial human time. “You’ve got the grunt work, like moving things around in Excel, doing basic programming tasks,” Butler said. “That’s like the bottom grunt work that the entry-level worker would have done two or three years ago.” Butler went further, saying in his view that much of this type of work has already been absorbed by AI tools. He connected that change to a shift he sees taking place in computer science, where entry-level roles have declined while experienced workers can use AI to increase what they can produce. “A senior developer now has access to essentially an unlimited amount of junior developers on demand,” Butler said. He was describing AI systems metaphorically, not literal employees. His point was that one experienced developer can now ask AI tools to produce code or other technical work that might previously have been assigned to more junior workers. That raises a broader question for universities and employers. If some of the simpler work traditionally given to newcomers can now be done by machines, how do people gain the experience needed to become senior workers themselves? Students doing work that once required years of programming Butler said he is already seeing the change inside his classroom. He created a Bishop’s course called CS208, Coding in Plain English, to show students — including those without traditional programming backgrounds — what they can accomplish using AI. Some come from disciplines such as psychology or neuroscience with no previous programming experience. “What I’m finding is that students who have zero programming experience, they come in from another discipline like psychology or neuroscience or something, they are now able to get up and running with real data sets,” Butler said. Butler’s own doctoral work involved neuroimaging, and he said he has assigned students work he considers comparable to the kinds of analysis he performed during his PhD. He described giving students a real dataset, an expected result and instructions to use AI to complete the analysis. Butler said those students have been able to carry out work at a level he associates with tasks he once performed after years of programming experience. That does not mean a student using AI instantly acquires the same knowledge as an experienced computer scientist. But Butler sees it as evidence that AI can dramatically lower the technical barrier between having a problem to solve and producing a working result. The role of the person using the system then changes. They need to know what to ask for, understand enough about the subject to judge the answer and take responsibility for what is ultimately produced. What about AI hallucinations? One persistent concern surrounding generative AI is hallucination — when a system produces information that sounds convincing but is false or unsupported. Butler said that problem has improved substantially in newer systems and is not as severe as it was with earlier models. He pointed in particular to newer AI systems’ ability to search the web, consult scientific databases and return source links rather than simply generating a citation from their internal model. Butler nevertheless said users should verify what AI produces. “You should verify,” he said while discussing citations generated with AI. His position is that hallucinations have become less common in the systems he uses, not that incorrect AI output is impossible. Human checking remains part of the process. Hours of work compressed Butler offered another example from his administrative work at Bishop’s. The university has been comparing CEGEP courses with first-year Bishop’s courses to determine where students may qualify for advanced standing. Traditionally, Butler said, that means reviewing two sets of course descriptions and manually matching one institution’s courses with another’s. As department chair, that could require hours of reading and comparison. He said AI can now be given both sets of material and asked to generate an initial mapping automatically. “It saves me hours and hours of work,” Butler said. Butler said the AI-generated mapping can also produce a better result than he would produce himself. He sees similar opportunities in businesses. A company may receive specifications, drawings, emails or other digital material from customers and then have employees convert that information into an internal format before anything can be manufactured or delivered. AI can increasingly perform parts of that intermediate processing. In Butler’s view, the employee does not necessarily disappear. Instead, the job becomes more about checking the result. “The human is still there,” he said. Judgement could become more valuable That change is central to Butler’s view of where digital work may be headed. If AI can produce a first draft, write basic code, compare documents or process large amounts of information, he argues that knowledge and judgement become increasingly important. “The real scarce resource going forward is going to be domain knowledge, expertise, and… credentials, authority, the ability to work with other people,” Butler said. He is therefore encouraging students not only to learn AI tools but also to broaden their education. He mentioned acting, fine art and photography as examples of courses that can help develop what he described as taste and intuition — the ability to recognize what is good, what is not and what should be changed. “And that’s the skill that they need now because the grunt work has been automated and now it’s about evaluating what the AI produces and telling it how to improve or expand the output,” Butler said. That does not remove the need for technical knowledge. Butler’s argument is that the emphasis may shift away from manually completing every step of a task and toward directing, checking and improving work produced with AI assistance. He also stressed that humans remain responsible for final decisions. “At the end of the day, the human has to make the final call,” Butler said, pointing to the question of accountability. Physical work may be different Butler draws a distinction between digital work and jobs requiring people to manipulate the physical world. He expects AI to continue advancing quickly in tasks performed almost entirely on computers. He is more cautious about claims that humanoid robots will soon be able to replace workers such as plumbers or electricians. Butler said those trades would not be replaced “anytime soon” in his view and described some current robotics demonstrations as overhyped. He described the situation as counterintuitive. “It’s almost like we solved the highest level first,” he said, contrasting advances in digital AI with the harder problem of building robots capable of operating reliably in the physical world. Butler’s argument is that some digital occupations could face automation sooner than trades requiring complex physical work. The wider AI shift in the Townships Butler’s comments come as local institutions are increasingly grappling with how AI should be used. The federal government recently announced up to $162 million over five years for 10,000 AI-related placements through Mitacs. The Université de Sherbrooke told The Pulse that the number of its collaborative research and innovation projects supported through Mitacs has approximately doubled since 2018-19, while the number of placements has approximately tripled. UdeS also said it already has projects involving both AI adoption and the application of AI to other fields. At Bishop’s, Butler said part of his role involves building the Computer Science department’s local presence, connecting students with regional partners and strengthening relationships with the Sherbrooke technology ecosystem. Those developments put universities in a difficult position. They have to prepare students to use the tools employers increasingly have available while also making sure graduates understand enough to recognize when those tools are wrong, incomplete or being applied to the wrong problem. The risk of losing human skills Butler does not see AI as risk-free. Later in the interview, he raised a separate concern: as AI becomes better at reading, writing, programming and other cognitive tasks, people could lose some of the underlying skills that historically helped produce new ideas. “The other danger is that we’re going to lose the human skills,” Butler said. He specifically mentioned reading and writing and argued that people could lose some of the lower-level understanding that contributes to breakthroughs. Butler said AI systems are based on existing human knowledge and argued that they become much more powerful when paired with a capable person. That tension runs through his view of the technology. AI can remove hours of routine work. It can allow beginners to perform tasks that once required substantial technical training. It can give experienced workers significant leverage. But the more capable the technology becomes, the more important it may be for people to develop the judgement needed to decide what should be done in the first place. For Butler, that is where education may have to change. Students still need to understand their field. They still need to know when something is wrong. And they still need to be able to make the final decision. But if AI increasingly handles the bottom of the cognitive pyramid, the value of a worker may depend less on how quickly they can perform the grunt work and more on whether they understand what the work is supposed to accomplish. Don’t miss the next story that affects your community. Join thousands of readers getting The Pulse every morning—plus breaking news alerts when it matters most. If you value this kind of local journalism, consider supporting it. Subscribe, share and like! Subscribe Share
Two Brazilian political dynasties clash again
Brazilians vote Sunday in a tight race between two familiar political forces, with plenty at stake beyond the presidency. Here’s what to watch.
Rural America May Be Ground Zero for Health System Reform
AFT Health Care has posted an article by senior fellow and director of healthcare reform Jeanne Lambrew on health reform through the lens of rural
Phantom Blade Zero, Castlevania: Belmont's Curse, Final Fantasy Resonance, and 5 More Big October Games We've Already Played and Can Tell You About
October is here, which means the Fall is in full swing, which means some of the biggest games of the year drop this month. We've already played some of the biggest ones, so check out our previews of Phantom Blade Zero, Castlevania: Belmont's Curse, Final Fantasy Resonance, and more below.
Big 12 launches sportsmanship campaign in wake of derogatory chants directed at BYU
The Big 12 has a zero-tolerance policy toward discriminatory behavior, but repeat issues sparked increased scrutiny and calls for action.
$10M gift helps launch new sports, entertainment, enterprise program
There was just enough room on the oversized check for all the zeroes. Seven of them, marking one of the larger philanthropic commitments in N.C. Wesleyan University's history.
FEAR IS BIGGER THAN EVER… AND ZERO LATENCY VR HAS THE SCIENCE TO PROVE IT
/PRNewswire/ -- Zero Latency VR, the world's largest and most innovative immersive VR network, is celebrating a double Halloween anniversary: three years since...
Jason Jones, Co-Creator of Halo and Destiny and Co-Founder of Bungie, Announces New Studio
Jason Jones, one of the co-founders of Bungie and a co-creator of Halo and Destiny, has announced his new studio, Stone Kite, which will work on projects across various entertainment mediums. The studio's first release is One Zero One, a limited-series comic book based on a manuscript written by Jones.
Stone Kite Is A New Studio From Halo, Destiny Creator Jason Jones And Author Margaret Stohl, And Its First Project Is About Rogue AI Robots
Stohl is perhaps best known for writing the Beautiful Creatures series of books. - Game Informer
MTV VMAs producer clarifies remark about ‘external’ influence leading to Taylor Swift’s new award
MTV VMAs producer clarified a colleague’s remark about ‘external’ influence leading to Taylor Swift’s new award, which she accepted Sunday, before the debut of her ‘Patient Zero’ music video.
Ace Combat Zero: The Belkan War’s modern port can’t be sold separately due to fighter jet licenses expiring, devs explain
The 20-year-old game will be distributed exclusively as a pre-order bonus for AC8.
Iowa Football Getting Zero Respect Despite 4-0 Start
Iowa continues to be ignored despite impressive wins early in the season For the Iowa football team, the lack of […]
Can we just like a Taylor Swift song in peace?
Taylor Swift's four new songs sparked instant debate, but "Patient Zero" makes a case for letting the music grow on you before joining the discourse.
Taylor Swift's 'Patient Zero' video premieres during historic VMAs
Taylor Swift debuted her video for 'Patient Zero' at a historic MTV VMAs for the pop superstar, who broke a Beyoncé record. The video featured Swift's 'good friend' Dakota Johnson, who also presented Swift with an award during the event.
SM Entertainment Takes Legal Action Against Malicious Posts on WINTER
SM Entertainment Takes Legal Action Against Malicious Posts on WINTER Agency Warns of Zero-Tolerance Policy Against Online Abuse Across Platforms
Malicious browser extensions can hijack AI assistants
BragJack research reveals how a single malicious browser extension could hijack Gemini, Perplexity Comet and Edge AI agents with zero clicks.
Justin Herbert Gives Vague Answer to Crucial Chargers Question
The Los Angeles Chargers somehow lost Sunday in Buffalo even after the defense forced five Bills turnovers. Justin Herbert and the offense managed a total of 16 points on a day when Josh Allen had zero touchdown passes and two interceptions.This drops the Chargers to 0-3 on the season with very little hope of a […] The post Justin Herbert Gives Vague Answer to Crucial Chargers Question appeared first on HEAVY.
NFL player makes unexpected cameo in Taylor Swift music video
An NFL star features in Taylor Swift's new "Patient Zero" music video, but it's not her husband Travis Kelce or any of his teammates.
Taylor Swift debuts ‘Patient Zero’ music video at VMAs starring Dakota Johnson and Colin Farrell
The pop star and actress play women who have dated the same man — who secretly poisoned both of them the whole time.
Taylor Swift Confirms She’s ‘Patient Zero’ In VMAs Music Video Debut
Fans had been parsing the singer’s social media for clues and coming up with theories for days leading up to the ceremony.
Taylor Swift Confirms She’s ‘Patient Zero’ In VMAs Music Video Debut
Fans had been parsing the singer’s social media for clues and coming up with theories for days leading up to the ceremony.
Taylor Swift premieres music video for ‘Patient Zero’ at the VMAs
Dakota Johnson and Colin Farrell star in this track off the encore edition of Swift’s 12th album, “The Life of a Showgirl.”
Taylor Swift Receives Inaugural Artist Director Honors at 2026 MTV VMAs as She Thanks Fans for Making Directing ‘So Rewarding’
Swift went on to premiere her new music video for “Patient Zero”
Flying With Gamma-Ray Counters: Another Test Or Two Of The Radiacode Zero
Yesterday I traveled all day, moving my family back to Lulea from Padova. The trip is long because there is no direct connection from Venice to Stockholm (Norwegian has one, but I prefer to avoid them
'Godzilla Minus Zero' makes franchise history, aims to 'shock' fans
NEW YORK – Look out, Manhattan. Godzilla has arrived in the Big Apple.
Seminoles Conquer Central Arkansas in Non-Conference Tilt
Florida State did not cover the 45.5-point spread against FCS opponent Central Arkansas, but they did enough to secure a 34-7 victory on Saturday to even their record at 2-2. The defense stepped up, as they should have, but the offense scored zero points in the fourth quarter.
What to know about Taylor Swift’s new music
Taylor Swift has released “The Life of a Showgirl: The Encore.” It's an extended version of her 2025 album with four new tracks. Those include the single “Patient Zero” as
Taylor Swift releases 4 new songs, including 'Patient Zero'
Taylor Swift's 'The Life of a Showgirl: The Encore' (Mert Alas & Marcus Piggott)
Taylor Swift releases 'The Life of a Showgirl: The Encore' with 4 new songs after Swifties try to decode her mysterious zeros
Our favorite pop star has dropped new music, including the single "Patient Zero."
The Googlebook has an opportunity to build Android like nothing else [Video]
Despite having effectively zero market share right now, there is an opportunity for Googlebook to do something that basically nobody...
Taylor Swift Just Teased the 'Patient Zero' Music Video and I Think It Confirms *This* Theory
Taylor Swift announced that she'd be premiering a new music video, "Patient Zero," a the VMAs this weekend. Here's why I think it points to this long-standing theory.
iPhone 18 Pro one week later: Zero compromises
I’ve been using iPhone 18 Pro for the past week after a full year as an iPhone Air user. Here’s how that’s been going.
From 15,000 to 185 to Zero: Trump’s Claims of Illegal Voting in Nevada Fizzle
It was the Trump administration’s latest stumble in its effort to find proof of widespread noncitizen voting, despite decades of evidence that it is extremely rare.
Taylor Swift teases new music video with Dakota Johnson, Colin Farrell
Are you ready for it? Taylor Swift has revealed a first look at the music video for her song "Patient Zero," which will debut during the MTV Video Music Awards
Why Tacoma athletes and seniors alike are braving this freezing, sub-zero box
Move aside, ice bath and cold plunges. Cryotherapy is having a moment in Tacoma thanks to this former Wilson High hooper.
Quinn Meinerz and Garett Bolles Are Powering the Broncos' Commitment to the Run Game
Quinn Meinerz has a 90.9 PFF grade and zero penalties through two games, and Garett Bolles leads Denver's line as three starters still find their footing.
Taylor Swift announces 'The Life of a Showgirl: The Encore' with 4 new songs after Swifties try to decode her mysterious zeros
Swifties, it's happening: Our favorite pop star is releasing new music.
Reform and the UAE – a threat to British politics
By Alex Carlen The United Arab Emirates likes to present itself as part of the solution to the climate crisis. The oil-rich Gulf state has committed to Net Zero by 2050, hosted COP 28, and even declared “championing climate action” as one of three strategic objectives for this week’s session of the United Nations General Assembly. But behind the scenes the UAE has quietly built one of its closest political relationships in Europe with the party most hostile to climate action in Britain: Nigel Farage’s Reform UK. There is no evidence that the UAE is funding Reform. But a new FairSquare report, Petro-populism: the UAE–Reform threat, published today, documents a growing political relationship that should concern anyone interested in the integrity of British democracy. Reform’s leaders have embraced the UAE’s political model and promoted positions that align closely with the Emirati government’s priorities, while the UAE has given Farage and his party unusually warm access and public backing. What is emerging is not simply a relationship between British politicians and a wealthy Gulf state, but a channel through which an authoritarian government can cultivate political influence here, and abroad. Democracy for Sale relies on its readers. If you already support us, thanks so much. If you don’t, sign up to support fearless independent investigative journalism. Subscribe Back in January, Nigel Farage appeared at a GB News event in Dubai. On the 50th floor of a tower overlooking the luxury resorts and upscale residences of Palm Jumeirah, the Reform leader declared: “I want Clacton to look like this.” At the event – which Farage never declared on his official register of interests – senior Emirati officials mixed with British expats. Farage was keen to flatter. “We recognise you are our friends”, he said, assuring them that “a Brexit London, a Reform London, will remember you.” Controversy surrounding Reform’s finances and its susceptibility to foreign influence have centred around high-profile donations from billionaire crypto donors, and the jailing of Nathan Gill, Reform’s former leader in Wales, for accepting bribes to give pro-Russia speeches. Far less attention has been paid to Reform’s blossoming new alliance with the UAE, an oil-rich authoritarian state with a track record of ruinous anti-democratic interventions. Farage wasn’t always so fond of the UAE. Back in 2023 he called it an “absolute monarchy with an abysmal human rights record.” Now Reform’s hierarchy calls the UAE home. Deputy leader Richard Tice splits his time between his constituency and Dubai and has called on the UK to learn from the UAE. The party’s treasurer Nick Candy runs a firm which has been in partnership with a UAE-state linked firm since 2024. Nadhim Zahawi sits on the board of an Abu Dhabi think tank alongside an array of the most senior UAE officials and is reportedly close to three brothers of the country’s despotic ruler Mohammed Bin Zayed. Reform’s leading lights have effusively praised the UAE’s governance model, defended it from criticism, and advanced its political agenda in the UK. GB News, whose two major shareholders – the Legatum Foundation and Paul Marshall – have strong links to the UAE, has been more than happy to broadcast UAE propaganda and was forced to pay damages in 2025 after airing a false claim about a UK charity by the UAE “disinfluencer” Amjad Taha. The UAE has responded in kind. When the UAE Minister for Foreign Affairs conducted an official visit to the UK in May, Farage was the only non-government figure granted an audience. UAE state media publicised the meeting, and said that they discussed “the strategic relations between the two countries and ways to enhance them in a manner that serves mutual interests.” For the UAE these ‘mutual interests’ largely concern its paranoia over political Islam, and in this regard the UAE-Reform alliance is part of a broad and troubling pattern. Following a crackdown on dissent at home in the aftermath of the Arab Spring, the UAE has offered material and political support to a range of violent and authoritarian actors willing to oppose the Muslim Brotherhood with no regard for the consequences. Its meddling in Egypt, Tunisia, Yemen, and most notably Sudan, have been devastating. The UAE’s strategy of seeking out partners on the political fringes who it believes will advance its interests is not unique to Britain – it extends beyond the Middle East to Europe and the US. In France, it has offered clandestine support to Marine Le Pen through secret meetings, as well allegedly funding trips to the region after which Le Pen returned with a renewed focus on the UAE’s perceived enemies – Qatar and the Muslim Brotherhood. Left-wing opposition leader Jean-Luc Mélenchon has accused the UAE of smearing him and of “encouraging discrimination, hatred or violence”. In the US, the UAE actively supported the rise of Donald Trump, establishing itself as a key ally during both his first and second terms. Around these relationships sits a wider ecosystem of figures willing to amplify UAE narratives, that even extends to Andrew Tate and Tommy Robinson. Subscribe In Britain, the UAE has used a combination of aggressive diplomacy and illicit, disruptive tactics to achieve its policy goal of eliminating any perceived threat from political Islam. In 2012, for instance, it threatened to block billion-pound arms deals with the UK and stop inward investment if David Cameron did not act against the Muslim Brotherhood and rein in the BBC’s coverage of the group – going so far as to suggest, with no supporting evidence, that the group had infiltrated the BBC. It has financed smear campaigns against a high-profile UK charity and funded a mercenary to secretly surveil, and possibly assassinate, a British national on British soil. In Reform UK, the UAE has finally found an ally willing to not only ban the Muslim Brotherhood, but to publicly vilify the group and repeat the UAE’s claims that dangerous Muslims are a fifth column in British society, a narrative that aligns with the Islamophobia and anti-immigration sentiments that underpin Reform’s populist politics. The Rycroft review into foreign interference concluded that “hostile states and non-state actors have a clear incentive to attempt to undermine UK democracy.” The UAE’s network and alliance with Reform must be understood within that framework. It is a powerful, disruptive and deeply authoritarian actor with the resources and networks to intervene in our politics, and a track-record of doing so in a multitude of other countries. The evidence of the threat the UAE poses is overwhelming. The question is whether parliamentarians are willing to call the UAE out and demand action using the tools that the government has to respond. Placing the UAE on the “Enhanced Tier” of the Foreign Registration scheme alongside Russia and Iran would send a powerful signal to the UAE that its meddling will no longer be tolerated, and would place Farage’s next trips to the UAE under the scrutiny they deserve. Alex Carlen is research and advocacy manager at London based human rights organisation FairSquare and author of its new report Petro-populism: the UAE–Reform threat. Subscribe
Chinese Hackers Exploit Chrome-Windows Zero-Day Chain to Deploy CLEANGULP Malware
UTA0565 exploited a Chrome-Windows zero-day chain through fake websites to deploy CLEANGULP malware against Asian government entities.
Did this LA Times story inspire Taylor Swift's new single?
LOS ANGELES — "Taylor Swift is patient zero for emotional contagion." This was an evocative turn of phrase from former L.A. Times columnist Amy Kaufman, in a 2024 article assessing
New Taylor Swift music is coming! 'Patient Zero' single announced
New Taylor Swift music is coming! The pop superstar has announced a new single, "Patient Zero," set to drop on Friday, Sept. 25.
Taylor Swift announces new single 'Patient Zero' after Swifties try to decode her mysterious zeros
Swifties, it's happening: Our favorite pop star is releasing new music.
SeaSTORE Scientists Measure Trawling's Impact On Irish Sea Carbon Storage
The SeaSTORE project is conducting real-world research cruises in the Irish Sea to measure how seabed trawling disturbs stored carbon and whether that carbon ultimately reaches the atmosphere. Findings will be shared with UK policymakers to inform net zero, biodiversity, and food security decisions.
Germany’s political center suffers historic blow, Turkey weighs coming to Saudi Arabia’s aid, Major TV networks support outlets banned from White House
Two weeks ago, German Chancellor Friedrich Merz was “shocked” when his conservative CDU party got creamed by the hard right AfD in local elections in the state of Saxony-Anhalt. That was just the start.
Scientists claim elastocaloric cooling breakthrough
HONG KONG: Scientists claim to have developed what is said to be the world’s first zero-degradation elastocaloric cooling device....Read More...
Startup JetZero is building a sleek new kind of plane with bigger-than-Boeing dreams
A startup is building a new type of passenger plane to rival Boeing and Airbus' tube-and-wing models. The blended wing body aircraft could take flight next year.
Override Series: Cyberpunk: Edgerunners Cup Zero Build in Europe: Session 1
Keep running on the edge. Chase the win. Team up as a trio and compete for a chance to earn Cyberpunk: Edgerunners cosmetic rewards. Good luck and have fun!
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.
Holy Cross Health Recognized for Zero Reported Hospital-Onset
The hospital is committed to evidence-based strategies that protect patients by reducing infection risks. August 26, 2026 – Holy Cross Health is one of a select group of U.S. hospitals to be recognized in the Healthcare Associated Infections-Hospital database for reporting zero hospital-onset Methicillin-resistant Staphylococcus aureus (MRSA) bloodstream infections. The database relies on publicly reported […]
Xbox bug adds phantom achievements that spoil 100% completions
Xbox users have spotted a bug that adds unobtainable phantom achievements worth zero Gamerscore to certain games, spoiling their 100% completion rate.
Licking County Health Department has tested a record number of bats for rabies this year - The Reporting Project
The health department said of the 50 bats it has tested this year, zero have tested positive for the deadly virus.