SCROLL NEWS / DISCOVERY
Search headlines
Find the stories shaping the conversation.
Results for " Model S " 77 found 🔀 AI-powered Shuffle
Ryder grows its FMS business with East Coast acquisition
Ryder System (NYSE: R) has acquired Efficiency Enterprises, a company with a business model similar to what Ryder does in its Fleet Management Systems
Ryder grows its FMS business with East Coast acquisition
Ryder System (NYSE: R) has acquired Efficiency Enterprises, a company with a business model similar to what Ryder does in its Fleet Management Systems (FMS) unit. Ryder, in a prepared statement, described Efficiency Enterprises as “a regional provider of fleet management, maintenance, and leasing solutions for trucks, trailers, and material handling equipment.” Those services line […] The post Ryder grows its FMS business with East Coast acquisition appeared first on FreightWaves.
Model Indya Moore Goes After Martha Stewart's Deeply Insensitive Halloween Costume
“This is why your excuse is not believable for so many people,” the model said.
Hugging Face’s New AI Biology Model Lets You Chat With DNA
The new open-source model speaks the language of DNA and can help researchers understand the effects of any changes on gene activity and health.
UTA Science Learning Center sets up laboratory-themed escape rooms
The event aimed to highlight resources for students and incorporated anatomy models, chemistry model sets and more.
NASA Study Refines Projections for Thwaites Glacier
Antarctica’s Thwaites Glacier could contribute more to sea level rise by the end of the century than some previous estimates. In some model simulations, the glacier’s contribution was as much as twice current projections.
Opinion | Dane County must stop underfunding mental health services
Natalie Eilbert: The need for mental health support is high, but Dane County’s contract model sets major provider Journey Mental Health Center up to fail.
Sahara Ray, model once linked to Justin Bieber, dies at 34
NEW YORK — Model Sahara Ray is dead at 34. The family of the Australian fashionista once linked to Justin Bieber told TMZ that Ray died Sunday night “surrounded by
Could Nordsjaelland model solve Scotland's talent crisis?
BBC Sport Scotland looks at how Nordsjaelland have become a powerhouse of youth football development, and what Scottish clubs could copy.
The hidden math of Alzheimer's could point to new treatments
Mathematics may hold unexpected clues to understanding Alzheimer’s disease. Scientists have developed a model showing how common metals like copper and zinc may influence the buildup of proteins associated with the disease, while also exploring possible ways to stop the process. The findings could help researchers identify promising strategies for preventing harmful plaque formation.
HSBC’s Synopsys Upgrade Bets on a Business Model Shift. Can It Justify $700?
HSBC upgraded Synopsys Inc. (NASDAQ:SNPS) to Buy from Hold on September 25 and raised its price target from $490 to $700, implying a multiple of nearly 35 times HSBC’s fiscal 2027 EPS estimates. According to analyst Frank Lee, the company is evolving beyond its traditional slow-growth software profile as its licensing-plus-royalty model and agentic AI […]
Sutter Health care model safely expands local Alzheimer's treatment
A multidisciplinary care model developed across Sutter Health helped general neurologists address a practical challenge facing health systems nationwide: how to safely offer newer Alzheimer's disease treatments in community settings to more patients.
Sphere Entertainment (SPHR) Could Be 40% Undervalued As Oz Demand Worries Hit Sentiment
Craig-Hallum’s downgrade of Sphere Entertainment (SPHR) after concerns about softening demand for its The Wizard of Oz show has put fresh focus on how this experiential model supports the current valuation. Sphere Entertainment’s latest downgrade comes after a sharp pullback, with the 7-day share price return down 20.98% and the 30-day return down 25.40%. Even though the year-to-date share price return is 12.66% and the 1-year total shareholder return is 79.19%, recent momentum is clearly...
Venus' mysterious haze is actually cosmic dust
A physical model show that iron dust and sulfuric acid create the right properties.
Modders are now merging entire games, from Minecraft in Elden Ring to Mario Kart in COD Zombies
Modders are moving beyond model swaps by combining gameplay systems from entirely different titles. AI agents, recompilation and matching decompilation projects are speeding up the work.
Pioneer of China NewCo model sued in US over trade secret misuse, double-dealing
A legal feud unfolding in New York is putting a spotlight on the work of a business development veteran who is credited for pioneering one of biopharma’s hottest dealmaking mechanism: the China “NewCo” model. In a lawsuit originally filed in New York state court and recently moved to federal court, China’s Lepu Medical Technology has alleged that biopharma business development consultant Su Zhang, who was hired to help shop its GLP-1 weight-loss candidates to Western partners, engaged in corporate double-dealing and trade secret leaks. According to the complaint, while acting as Lepu’s primary dealmaking liaison for a group of obesity and metabolic disease drug candidates, Zhang was secretly double-dipping as the acting chief business officer for a direct competitor, BrightGene Pharmaceutical. “She never adequately disclosed the role, never sought Lepu’s consent, and never proposed any safeguard,” Lepu alleged in its complaint, which was filed in July after Zhang’s consulting firm, CoDevCo, sought certain transaction-based success fees from its angry customer through an arbitration process in Singapore. Lepu, the parent company of Lepu Biopharma, has asked Zhang to pay back its $500,000 retainer and disgorge all benefits she received during the alleged breaches of fiduciary duty, plus other damages and relief. The company has also requested the court grant injunctions to protect Lepu’s trade secrets. The ill-fated partnership officially kicked off in December 2024, when Lepu tapped CoDevCo as its “exclusive business development mechanism” to identify collaborators for the company’s drug candidates. According to Lepu’s complaint, Zhang personally negotiated and signed the agreement as managing director for CoDevCo, which used her home address. Before the Lepu deal, Zhang made a name following a short stint as the global head of business development for Hengrui Pharmaceuticals. It was during her tenure that Hengrui signed a landmark deal granting exclusive rights for a portfolio of GLP-1 assets to Hercules, which was at that time newly established by a group of private equity and venture capital firms. In addition to the regular upfront and milestone payments, Hengrui also gained an equity stake in Hercules, which later became Kailera Therapeutics. The Hengrui-Hercules deal opened a new paradigm, where a Chinese biopharma company spins off certain early-stage assets to investors for further development through the creation of a new company abroad. The “NewCo” model quickly came into vogue as it allows Western investors to access high-potential assets originating from China at an early stage while giving the Chinese licensors an opportunity to reap long-term benefits through equity stakes in the new firms. The NewCo model proved its value when Kailera staged a record-breaking $625 million IPO on Nasdaq this April, merely two years into the Hengrui deal. Related Kailera CEO 'knew we were in a good spot' before obesity biotech's record-breaking $625M IPO On her LinkedIn page, Zhang described her work at Hengrui as “[b]uilding team and capability from the ground up,” forming NewCo and major pharma partnerships worth billions of dollars. Lepu clearly wanted that kind of track record to help boost its own cardiometabolic portfolio, which includes a range of assets for obesity, Type 2 diabetes, fatty liver disease and muscle preservation during weight loss treatment. “Lepu’s pipeline is very valuable, but only if Lepu could secure the right international partner to expand into the overseas market during the available commercial window,” the company argued in its complaint. The Chinese biotech said it had little knowledge of U.S. laws, customs or business practices in the business development field, and therefore “needed to rely entirely” on the expertise of Zhang and her team as the firm’s first engagement of a U.S.-based BD consultant. “Zhang’s dominating influence on the business development strategy is further evidenced by the fact that she personally controlled the business-development communications on behalf of Lepu with potential partners,” Lepu said in its complaint. After entering the deal, Zhang introduced Lepu to her four associates. According to Lepu’s version of events, all of them were allegedly employed at different time points to perform similar jobs at other Chinese companies that Lepu described as its competitors, while having access to Lepu’s confidential information. As such, Lepu described Zhang’s team as “a roster of individuals whose primary professional loyalty, by virtue of their concurrent full-time employment, ran to Lepu’s competitors.” As well as the specific examples set out in the case, a look at Lepu’s deals over the period suggests Zhang may have been involved in at least one transaction. During their partnership, Lepu, through its subsidiary Shanghai Minwei Biotechnology, also out-licensed its GLP-1/GIP/FGF21 agonist MWN105 to a NewCo called Sidera Bio in the fall of 2025. The deal included an upfront and near-term payment of $35 million, up to $1.01 billion in milestones, and a 9.99% stake in the Danish startup. It seems likely that Zhang facilitated the Minwei deal given that Lepu is on record as having used CoDevCo as its “exclusive business development mechanism.” The big shock for Lepu came in the form of a news article on June 25, 2025. The article, reporting on phase 2 data for BrightGene’s GLP-1/GIP receptor agonist BGM0504, identified Zhang as acting chief business officer of the Lepu rival. Blindsided by her association with BrightGene, Lepu confronted Zhang in July 2025. “Lepu was given assurances that there was nothing to be concerned about and told that Zhang would explain the matter to Lepu,” Lepu alleged in its complaint. “Zhang never did.” Lepu claimed that Zhang had the motive and opportunity to steer potential prospective partners toward BrightGene or other competitors. Later, in February 2026, one of Zhang’s associates joined BrightGene as the vice president for global business development, according to the complaint. Around the same time, Lepu sent written demand letters asking Zhang to account for her BrightGene role, identify what confidential data were shared with the rival firm, and provide any conflict-of-interest or firewall policies during the engagement. According to Lepu, neither CoDevCo nor Zhang answered the questions in their responses. Related Novo vs. KBP: How a $1.3B deal turned into a fraud accusation and an international legal fight In a motion to dismiss filed with the U.S. District Court in the Eastern District of New York on Sept. 4, Zhang’s legal team argued that Lepu’s agreement was signed with CoDevCo, and that an employee of the contractor does not owe fiduciary duties to the client. In its complaint, Lepu said: “That CoDevCo was the contracting entity does not insulate Zhang from personal liability for duties she personally undertook and personally breached.” Taking the argument one step further, the defending party argued that the consulting agreement “does not create a fiduciary relationship even between CoDevCo and Lepu; instead, it identifies CoDevCo as an independent contractor retained to market Lepu’s business to potential western partners.” Although CoDevCo was Lepu’s exclusive business-development contractor, there were no terms stipulating that the consulting firm should serve only Lepu, Zhang’s lawyer noted. As to Lepu’s claims that Zhang misappropriated trade secrets, the defendant argued that Lepu authorized the use of the information so Zhang’s team could perform their service per the agreement. In a rebuttal filed on Sept. 11, Lepu’s lawyer set out why they believe the Chinese biotech satisfied New York law’s misappropriation definition by alleging that the defendant used a trade secret “in breach of a duty or through improper means.” As an alternative to dismissing the suit, the defendant suggested the action should be stayed pending outcomes from the arbitration of a separate case in Singapore. While NewCo pioneer Zhang is defending her business arrangements, momentum behind the NewCo trend has continued unabated into 2026. Through August, Chinese biotech assets have spurred 15 NewCo deals this year, according to local trade publication PharmaDJ.
Mexico promotes a community tourism model strengthened by knowledge
UNESCO continues to support Mexican authorities and communities in transforming the tourism paradigm for sustainable development.
Presley Gerber dies at 27. What Cindy Crawford's model son said about recovery, medication and feeling 'misunderstood.'
Gerber, the son of Cindy Crawford and Rande Gerber, had spoken publicly about his mental health and experiences with addiction.
One of X’s most notorious ragebaiters explains how the outrage business earned him $80,000—and why he’ll keep posting even as the creator model shifts
A typical month posting for Lefkowitz has paid between $3,000 and $10,000 from X.
Lava-Covered Planets Are Defying a Key Rule of Planetary Science
Researchers developed a model showing how lava-covered planets close to their stars can hold onto their atmospheres, which could help guide the search for life beyond our solar system. Some planets orbit so close to their stars that intense radiation should strip away their atmospheres. Yet several of these scorching, lava-covered worlds still retain thick
What's NEXT: adjoe’s Eileen Keng on why entertainment apps are turning to gaming for retention and revenue
Behind this model sits a straightforward funding mechanism, according to Eileen: premium mobile game advertisers cover the cost of both user rewards and publisher revenue through performance-based campaigns.
Stability AI Raises $76M Series B, Backed by Major Entertainment Companies
Stability AI, the startup behind image generation model Stable Diffusion, has raised $76 million in Series B funding, bringing its total funding to $232
TCL launches fourth-generation WOLED monitors with up to 480Hz refresh rate
TCL has launched two fourth-generation WOLED gaming monitors in China. The 32-inch model supports 4K 240Hz and FHD 480Hz modes.
Strategy buys preferred stock in past week instead of bitcoin, hit to business model signaled
Strategy (MSTR) was trading higher as the bitcoin treasury company announced a repurchase of its preferred stock instead of buying bitcoin in the past week.
Science thought the human lifespan was 122 years. A new model says we could live decades longer.
A July 2026 paper published in the journal "npj Aging" suggests that humans may have the potential to live up to 156 years.
How Attribution Science Uses Models to Uncover Climate Change’s Effects on Weather
Running climate model simulations with—and without—observed human influences can quantify how emissions changed the probability or intensity of a weather event.
New Switch 2 model spotted in the wild with updated screen
It looks like a new version of Switch 2 with an updated LCD screen is rolling out soon…
Dirt on Dirt - Salina gives Davenport another MLRA sweep
DirtonDirt.com covers Dirt Late Model racing like no other publication, magazine or website with massive amounts of original content including videos, photos, stories, columns, blogs and more. Eldora Speedway's World 100 and Dream, the Dirt Track World Championship, Lucas Oil Late Model Knoxville Nationals and all the major races are covered along with the major tours, including the World of Outlaws Late Model Series and Lucas Oil Late Model Dirt Series.
First 'Confirmed' iPhone Ultra Color Allegedly Revealed in Leaked Image
Apple is expected to launch its first foldable iPhone later this year. Rumors suggest the "iPhone Ultra" will come in two color options, and a leaker shared an image today that allegedly shows one of them. Posted on Weibo by the Chinese leaker known as Ice Universe, the image purportedly offers a first glimpse of Apple's foldable in white. The device is believed to have entered early mass production, but the model shown is likely a dummy.
Dirt on Dirt - Feger starting on MARS pole for Red Hill Rumble
DirtonDirt.com covers Dirt Late Model racing like no other publication, magazine or website with massive amounts of original content including videos, photos, stories, columns, blogs and more. Eldora Speedway's World 100 and Dream, the Dirt Track World Championship, Lucas Oil Late Model Knoxville Nationals and all the major races are covered along with the major tours, including the World of Outlaws Late Model Series and Lucas Oil Late Model Dirt Series.
D-Wave Quantum and Department of Commerce Sign Letter of Intent for $100 Million in CHIPS and Science Act Funding to Accelerate U.S. Leadership in Quantum Computing
D-Wave Quantum Inc. (NYSE: QBTS), (“D-Wave” or the “Company”), the only dual-platform quantum computing company providing both annealing and gate-model syste...
Scientists were wrong about this “rule-breaking” particle
Scientists spent decades chasing signs of a mysterious new force hidden inside the muon, one of nature’s strangest particles. But after years of supercomputer calculations, researchers discovered the apparent anomaly was likely a calculation error — and the Standard Model still reigns supreme.
Microsoft’s multi-agent AI system tops Anthropic’s Mythos on cybersecurity benchmark
Microsoft's new vulnerability-scanning system, codenamed MDASH, scored 88.45% on the CyberGym benchmark, surpassing single-model systems from Anthropic and OpenAI by using more than 100 specialized AI agents across multiple models.
AI uses 12-lead ECGs to predict long-term stroke risk
Stroke is still a leading cause of death all over the world. A new AI model showed early potential to anticipate such adverse outcomes up to a decade in advance, ensuring high-risk patients receive the care they need.
Palmeiras reject multi-million Manchester City bid for Heittor Vinicius
Manchester City have seen an initial €20 million offer for Palmeiras’ Heittor Vinicius turned down.A strategic Etihad recruitment model substantiates the pursuit of emerging Brazilian talent.Playe...
Tesla is selling 250 invite-only Model S 'Signatures' for $160,000
Elon Musk's automaker is inviting a select group of fans to purchase limited-edition versions of its soon-to-be-discontinued Model X and S luxury EVs.
“AI polls” are fake polls
A few weeks after Donald Trump’s second presidential win, I took the train up from London (where I was living at the time) to Oxford to attend a conference on polls and forecasts of the 2024 election. Most of the attendees were pollsters or academics, but I also watched presentations from Aaru and Electric Twin, two companies that do what is interchangeably called synthetic sampling, silicon sampling, or synthetic audiences. Stripped of startup jargon, that means they use large language models (LLMs) to simulate responses to public opinion polls by having AI agents take on the role of survey respondents. I had already heard of Aaru thanks to some articles with eye-catching headlines like “No people, no problem: AI chatbots predict elections better than humans” in the months leading up to Election Day. The founders were making some big, some might even say far-fetched claims, such as: “within two years, we will simulate the entire globe — from the way crops are grown in Ukraine to how that impacts production of oil in Iraq, trade through the strait of Malacca, and elections for the mayor of Baltimore.” When Semafor asked Aaru’s cofounders — Cameron Fink and Ned Koh — about my boss, they said “we respect all those who came before us.” Nate (as he so often does) shared his thoughts on Twitter: Fink and Koh were relatively good-natured about this back-and-forth when we spoke at Oxford. They even offered to mail me one of the t-shirts featuring Nate’s quote they apparently had made. I never took them up on the offer, which I now somewhat regret. These synthetic sampling companies fell off my radar for a while, but they do still exist. In fact, Aaru recently received a $1 billion valuation. Is what they’re doing anywhere close to the most important frontier in AI development? Not by a longshot, especially when Anthropic just developed a model so adept at exploiting software vulnerabilities that it’s only being released to 40 companies. Still, silicon sampling is increasingly finding its way into public polling. Axios reported in March that “a majority of people trust their own doctors and nurses” based on findings from Aaru — without mentioning that the “people” in that sentence were actually LLMs. Around the same time, the Public Sentiment Institute “boosted” their online sample of 373 real survey respondents with 114 AI agents.1 (Spoiler alert: even the co-founder of Electric Twin doesn’t think that’s a particularly defensible approach.) Polling companies like Qualtrics and Ipsos are also developing synthetic data panels. So, what should we make of these … “polls”? Let’s get one thing out of the way: whatever they are, they’re not polls in the way that term is usually defined. Subscribe You can’t replace polls with AI On one hand, using LLMs to essentially make up fake survey respondents sort of sounds like the dumbest idea ever, one that will at best imperfectly replicate real polls while introducing all sorts of biases. On the other hand, with LLMs improving at a remarkable, perhaps even alarming rate, maybe that means I’m a dinosaur at the ripe old age of 24 because I still want to rely on polls that talk to actual people. I’m not going to argue that synthetic samples are completely useless. In fact, as I’ll return to later, there is evidence that some techniques can replicate topline survey results quickly and cheaply. But the marketing from certain companies can be slightly optimistic. “No traditional poll will exist by the time the next general election occurs,” said Fink in 2024. We’re just 206 days away from the midterms, and based on the fact that I still have to collect a bunch of polls every day, I’d say he should have run that prediction by a sample of AI agents before the interview.2 To see why synthetic samples can’t replace polls, here’s a quick primer on how they work. The simplest version of these models involves taking a LLM (like ChatGPT or Claude), giving it a demographic profile (e.g., a white, college-educated woman who lives in Utah and makes $70k a year), and then asking it to respond to a survey question. You repeat that process a few thousand times using different demographic profiles and end up with a sample of synthetic survey responses. The actual models used by private companies are more sophisticated than this, usually because they incorporate more hypothesized demographic characteristics for each agent and provide them with extra information. Aaru, for example, feeds agents a diet of news and information they’d be likely to consume, while Electric Twin incorporates their customers’ proprietary data about the audience they’re trying to replicate. The way Ben Warner, the co-founder of Electric Twin, explained it to me was “we have a large amount of data on […] for instance, 5,000 people. Can we make an accurate prediction of how they would respond to another question?” Still, it should be obvious why synthetic samples can’t replace polls. Polling is fundamentally a data collection process. We might use surveys to make predictions by feeding them into election forecasts, but the main purpose of a poll isn’t prediction, it’s gathering new data about what people think and how they feel. Silicon sampling, on the other hand, produces no new data. It’s simply a model: you input LLM training data, demographic prompts, and a bunch of other information, and it spits out a prediction for what a poll would say. We love models here, but models aren’t polls. That difference is an important philosophical sticking point for most pollsters I talk to. “I think politics should stay away from [synthetic sampling], because we’re trying to […] represent the voice of the people,” said Natalie Jackson, a vice president at GQR Insights. Democratic pollster John Hagner told me: “I think I’m just incredibly skeptical of this idea. I don’t think it’s research. At that point, you’re asking the machine to tell you what you already believe.” Hagner has seen some presentations of early synthetic sampling experiments, but so far, “if it’s being used in a campaign, people are keeping it incredibly quiet.”3 But Eli, I hear you saying, aren’t polls themselves increasingly governed by modeling decisions? Indeed they are: pollsters’ choices on which sampling method to use, how to define their likely voter models, and how to weight their samples can and do lead to dramatic differences in the results they publish. Aaru even referenced these limitations in the methodology statement included with that maternal mortality “poll” — although I’m using the term “methodology statement” loosely here, because it doesn’t really explain how the model works at all. We can ignore the (frankly preposterous) implication that synthetic sampling isn’t subject to a separate set of biases. The important point is that there’s still a meaningful difference between using weighting and other statistical techniques on actual polling data and using a model to predict what a poll would say. The latter is far closer to election forecasts or techniques like MRP — potentially useful models, but not a replacement for polls.4 To be fair, other synthetic sampling companies are perfectly happy with the distinction between polls and models. Warner compared polling and synthetic sampling to different tools in a toolbox. “The mistake I think we make is we think that these new tools should either work in exactly the same way or somehow replace these old tools,” he said. “Rather than thinking of it as, okay, so we’ve always had the hammer, we’ve always had the screwdriver, now we’ve got a saw. But don’t use a saw to try [to] do the job of a hammer.” A quick comment from Nate Eli didn’t ask me for a comment — rather rude of him, don’t you think? But since I’m editing this story, I figured I’d add a few quick thoughts rather than putting words in his mouth. Beyond the frequently misleading marketing, what bothers me about the AI “poll” hype is that as AI tools make statistical inference cheaper and/or better (note that these are not synonyms) that actually increases the comparative value of collecting original data. You might be able to train a model to make a reasonable estimate of what some hard-to-reach poll respondent would say — say, a young Black man who voted for Trump. (Such a person checks a number of boxes for a voter who is usually hard to reach in surveys.) Indeed, this is closely related to what models like the Silver Bulletin forecast already do. They essentially smooth out the kinks in noisy survey data by making inferences based on past voting patterns or national polls or surveys of other states. But you don’t actually know what these voters think unless you’re reaching them directly. If there’s a shift in opinion among this subgroup, you’re not going to detect it. So if I were running a campaign, I’d invest more in going the extra mile to find a representative sample of those voters. And then I’d hire some smart quants — or Claude? — to figure out the implications for campaign strategy based on proprietary data that my competitors didn’t have access to. -Nate Silver Are these models any good? If synthetic surveys are just a new type of model, the next obvious question is whether the models are at least accurate. The answer very much depends on who you ask. On one end of the spectrum, you have the maximalist argument that synthetic sampling is more accurate than actual polls. “It’s an incredibly challenging problem to go to someone and say ‘hey, we’re going to be more accurate at predicting human behavior than you, even when you talk to your customers directly’,” Koh recently told CNBC. In his view, synthetic sampling isn’t a saw to polling’s hammer, it’s “magic.” There’s certainly evidence that synthetic samples can replicate certain survey toplines. But if Aaru does have any examples of their approach outperforming the polls, they’re keeping those to themselves.5 Aaru’s 2024 election model, for example, had Kamala Harris leading in Michigan, Nevada, Pennsylvania, and Wisconsin on November 4th. And although they’ve since taken down their forecast page, they gave Harris a 50.5 percent chance of winning the race on November 2nd.6 After the election, Fink told Semafor he was happy enough with those results because they were “within margin of error,” a term that is completely meaningless when applied to a “sample” of AI agents. And of course, Aaru says their models have improved since 2024, so supposedly now they’d be more accurate than the polls? Still, their stronger argument is on cost: “We are significantly faster and cheaper than traditional polling, and still more accurate,” said Fink. The first two claims are undeniably true, but the third brings us to the opposite end of the spectrum. Both Jackson and Hagner are skeptical that these models are reliable for anything beyond replicating common survey toplines. “I just […] don’t think the machines are what we want when we’re looking for nuanced views. My example on this is people in Arizona and Nevada in 2024 who voted for Trump and voted for expanding abortion in their states on ballot initiatives,” said Jackson. Hagner identified another issue. Maybe the synthetic respondents, like sycophantic LLMs, are inhuman in one important way: they’re too nice. “The reports that have come through at the meetings that I’ve been at are that the early experiments on this, they cannot get respondents to be as racist or sexist or, frankly, as negative as human respondents,” he said. Academic research mostly agrees on this point. While there are some papers that show promising results when using LLMs to replicate polling data, most show that LLMs suffer from various quirks like producing too few “don’t know” responses and can seriously overpredict the favorability of politicians like Donald Trump and Kamala Harris. They also seem to struggle with too little variation between demographic subgroups, so the difference in predicted opinion between Democrats and Republicans, for example, is too small. When I asked Warner about these studies, his response to these papers was that just because academics can’t get synthetic sampling to work doesn’t mean that the technique doesn’t work in general. “Actually, the argument is, okay, yours does not [work]. That does not mean […] for this complex set of machinery, which uses a lot of investment, a lot of time, a lot of money, you can’t get it to work.” Cards on the table, I’m somewhat sympathetic to this argument because academics aren’t exactly great at making election forecasts. Usually, the people with skin in the game are the most accurate. Warner’s argument is that the approach Electric Twin takes — which includes, for example, making multiple predictions for each synthetic respondent using different models and prompts and subsequently averaging those to get a final prediction in a sort of ensemble forecast — produces better results than the simpler academic models. Warner shared a comparison between his method and the method from a recent academic paper with me, and Electric Twin was indeed able to get more accurate replication. But even still, he acknowledged that synthetic sampling “is not a crystal ball.” “If you asked me, do I think using other data sources will be more accurate than asking somebody who they will vote for, I would probably say no. But if you asked me ‘would your system be useful for our turnout modeling today?’ I would say yes.” For better or worse, it looks like the method is already getting more popular in the market research world. Most of the clients Aaru touts these days are businesses like EY and McDonald’s. And AI will probably begin to pop up in other parts of the political polling process. Pollsters are already using it to code open-ended survey responses, and some firms, like YouGov, are testing using LLMs to ask survey respondents questions. More worryingly, one danger to actual polls is that AI agents can be used to infiltrate online surveys. Most online polls use various checks to prevent that from happening, but there’s conflicting evidence on how effective those filters are and how prevalent AI agents currently are in online panels. If those agents ever become impossible to detect, it might spell the end of online polling, but the solution isn’t to replace all of your respondents with ChatGPT. Silver Bulletin is a reader-supported publication. To receive new posts and support our work, consider becoming a subscriber. Subscribe 1 That particular poll obviously doesn’t meet Silver Bulletin standards for aggregation. But we exclude all Public Sentiment Institute polls from our averages because we classify them as an amateur polling firm. 2 You could argue that Fink meant the next presidential election, but (a) I’m also confident we’ll still have real polls in 2028 and (b) in that case he should have asked an LLM to define “general election.” 3 Quick caveat: that’s reporting from a Democratic pollster. It’s possible that Republicans are more willing to use AI in political campaigns. 4 Indeed, Silver Bulletin does not include “polls” produced by MRP in our forecasts or averages, and we think it’s extremely misleading when their practitioners describe them in a way that suggests original data had been collected among a large number of states or Congressional districts. 5 A recent report from Aaru and EY did show two examples of a synthetic estimate being closer than a survey to a real-world benchmark — but I’d take those findings with a grain of salt because the report reads more like an ad and didn’t involve any sort of prediction being made ahead of time. 6 For comparison, our odds for Harris on the same day were 48.2 percent.
Anthropic Mythos cybersecurity concerns: What smart people are saying
Anthropic's Mythos AI model sparked cybersecurity concerns, but some think it's been overhyped. Here's what smart people re saying.
Your Next Breach Will Look Like Business as Usual
These are the fundamental detection model shifts cybersecurity teams need to make to keep up with the rising number of credential-based attacks.
Institute of Industrial Science, The University of Tokyo and George Mason University scientists mine past air temperature forecasts, lower cost with greater subseasonal prediction accuracy
Researchers at Institute of Industrial Science, The University of Tokyo and George Mason University‘s College of Science have developed a new method that improves air temperature forecasts one to five weeks in advance—without requiring additional model simulations.
AMD Rolls Out Gemma 4 Model Support Across Full Range of GPUs & CPUs
AMD has rolled out official support for Google's Gemma 4 across its full range of GPUs & CPUs, offering support for the compact AI model.
Gemma 4 and what makes an open model succeed
Having written a lot of model release blog posts, there’s something much harder about reviewing open models when they drop relative to closed models, especially in 2026. In recent years, there were so few open models, so when Llama 3 was released most people were still doing research on Llama 2 and super happy to get an update. When Qwen 3 was released, the Llama 4 fiasco had just gone down, and a whole research community was emerging to study RL on Qwen 2.5 — it was a no brainer to upgrade. Today, when an open model releases, it’s competing with Qwen 3.5, Kimi K2.5, GLM 5, MiniMax M2.5, GPT-OSS, Arcee Large, Nemotron 3, Olmo 3, and others. The space is populated, but still feels full of hidden opportunity. The potential of open models feels like a dark matter, a potential we know is huge, but few clear recipes and examples for how to unlock it are out there. Agentic AI, OpenClaw, and everything brewing in that space is going to spur mass experimentation in open models to complement the likes of Claude and Codex, not replace them. Especially with open models, the benchmarks at release are an extremely incomplete story. In some ways this is exciting, as new open models have a much higher variance and ability to surprise, but it also points at some structural reasons that make building businesses and great AI experiences around open models harder than the closed alternatives. When a new Claude Opus or GPT drops, spending a few hours with them in my agentic workflows is genuinely a good vibe test. For open models, putting them through this test is a category error. Something else to be said about open models in the era of agents is that they get out of the debate of integration, harnesses, and tools and let us see close to the ground on what exactly is the ability of just a model. Of course, we can’t test some things like search abilities without some tool, but being able to measure exactly the pace of progress of the model alone is a welcome simplification to a systematically opaque AI space. Share The list of factors I’d use to assess a new open-weight model I’m considering investing in includes: Model performance (and size) — how this model performs on benchmarks I care about and how it compares to other models of a similar size. Country of origin — some businesses care deeply about provenance, and if a model was built in China or not. Model license — if a model needs legal approval for use, uptake will be slower at mid-sized and large companies. Tooling at release — many models release with half-broken, or at least substantially slower, implementations in popular software like vLLM, Transformers, SGLANG, etc due to pushing the envelope of architectures or tools. Model fine-tunability — how easy or hard it is to modify the given model to your use-case when you actually try and use it. The core problem is that some of these are immediately available at release, e.g. general performance, license, origin, etc. but others such as tooling take day(s) to week(s) to stabilize, and others are open research questions — with no group systematically monitoring fine-tunability. In the early era of open models, the days of Llama 2 or 3 and Qwen pre v3.5, the architectures were fairly simple and the models tended to work out of the box. Some of this was due to the extremely hard work of the Llama, Qwen, Mistral, etc. developer teams. Some is due to the new models being genuinely harder to work with. When it comes to something like Qwen 3.5 or Nemotron 3, with hybrid models (either gated delta net or mamba layers), the tooling is very rough at release. Things you would expect to “just work” often don’t. I’ve been following this area closely since we released Olmo Hybrid with a similar architecture, and Qwen 3.5 is just starting to work well in the various open-source tools that need to all play nice together for RL research. That’s 1.5 months after the release date! This is just to start really investing more into understanding the behavior of the models. Of course, others started working on these models sooner by investing more engineering resources or relying on partially closed software. The fully open and distributed ecosystem takes a long time to get going on some new models. All of this is lead-in for the most important question for open models — how easy is it to adapt to specific use-cases? This is a different problem for different model sizes. Large MoE open-weight models may be used by entities like Cursor who need complex capabilities in their domain, e.g. Composer 2 trained on Kimi K2.5. Other applications can be built on much smaller models, such as Chroma’s Context-1 model for agentic search, built on GPT-OSS 20B. The question of “which models are fine-tunable” is largely background knowledge known by engineers across the industry. There should be a thriving research area here to support the open ecosystem model. The first step is to understand characteristics of different base and post-trained models to understand what they look like. The second step is to tune pretraining recipes for open models so they’re more flexible. Interconnects AI is a reader-supported publication. Consider becoming a subscriber. Subscribe For The ATOM Project and other Interconnects endeavors, we’ve put in substantial effort to measuring adoption trends in the open ecosystem. Everything takes a long time to unfold after a model is first publicly available — and adaptability is why. What we know for sure now, when Qwen has been going from strength to strength with its releases, is that technical staff across the industry has gotten comfortable working with Qwen models. Countless research methods and datasets were made to work with Qwen. It’ll take patience for any other model family to get to this point — a patience I’m not sure many open model builders have. This takes us to Gemma 4, Google’s latest open models. Gemma 3 was released more than a year ago, in March of 2025, and is a bit underrated. Gemma 4 comes in 4 sizes for now, with a bigger, MoE model of over 100B total parameters rumored but not released yet. The models we have today come in sizes of ~5B dense, 8B dense, 26B total 4B active MoE, and 31B dense. I’m most excited that they’re finally adopting a standard Apache 2.0 open source license. This’ll massively boost adoption. The standard of better licenses for strong open-weight LLMs was set by mostly Chinese open model labs in the last 1-2 years, and now U.S. companies are following suit. I will personally be so happy if the horrible Llama licenses and Gemma terms of service were an ~18-month transient dynamic of the industry being nervous about releasing strong open models. The Gemma 4 scores look very solid, the small models have incredible benchmark scores (especially in general domains like LMArena) and the 31B model rivals the recent Qwen 3.5 27B, which is the leading member of that class. The ~30B size range is an important one, as it’s accessible both to researchers and to enterprises looking to deploy the model in real use-cases. Where the 7B model scale is the default for tinkering and research, a 30B model is the default for seeing if an open model can unlock substantial value in your specific workflow — a good mix of intelligence, low price, tractability for downstream training, etc. This takes us back to the above adoption criteria I mentioned for open models and the bigger question — do I think Gemma 4 will be an overwhelming success? Previous Gemma models have been plagued by tooling issues and poorer performance when being finetuned. Gemma 4’s success is going to be entirely determined by ease of use, to a point where a 5-10% swing on benchmarks wouldn’t matter at all. It’s strong enough, small enough, with the right license, and from the U.S., so many companies are going to slot it in. I’m cautiously optimistic that Gemma 4 is going to work better here. Winds are shifting for open models built in America. We saw GPT-OSS go through a bumpy launch to become an overwhelming success. There’s a collective energy around the likes of Reflection, Arcee, Nemotron, Gemma, Olmo, and peers that show substantial demand for building new stacks around open models. There’s capital to be spent on AI stacks across the economy by those who want more ownership of everything, including the model. After launching The ATOM Project 240 days ago, the conversation is shifting into the next stage. Summer of 2025 was a crisis moment where the U.S. AI scene realized it can’t wait and figure out open models after building AGI. The two markets will capture different areas and proceed in parallel. Now that more companies in the U.S. are releasing strong models, we need to improve the ecosystem so that these models are easy to use, understand, and build value around. It’s the hard work to build another inflection point in these adoption plots I’ve been updating consistently, but that’s the work to be done. Join me in it. More data coming soon! Here’s a sneak peek:
Clock is ticking on law's billable hour, says Cleary Gottlieb lawyer
For decades, elite corporate law firms have relied on the billable hour to fuel profits. Can the model survive the age of AI?
The real threat to trust in science isn’t outright fraud, but the pervasive tweaking of research designs and models
Thomas Plümper and Eric Neumayer argue that science's credibility issues stem from the deliberate manipulation of research designs and model specifications.
Pokémon Champions is releasing on Nintendo Switch consoles in early April, but its "free-to-start" package raises some questions about its pricing model
Pokémon Champions is finally arriving in April to Nintendo Switch consoles, but its "free-to-start" model seems limiting and confusing.
The System That Decides What Science Gets Published Is Breaking Down
The peer review system that validates scientific research is trapped in a self-defeating cycle. A new mathematical model shows why—and what comes next.
Forum on harnessing Artificial Intelligence for health equity
Register to this forumOver the past decade, artificial intelligence (AI) has significantly evolved in complexity, performance, and utility. Its integration across a wide range of industrial sectors has accelerated, with healthcare emerging as one of the most transformative sectors. In this sector, AI has enabled high-impact applications – from enhancing the speed and accuracy of disease diagnosis and supporting clinical decision-making, to strengthening public health initiatives such as disease surveillance and workforce optimization. Moreover, AI holds immense potential to empower individuals, enabling them to take greater control of their personal health and well-being. Yet the central question remains: Who benefits, who is at risk, and do these innovations close or widen health inequities? While the potential of AI-enabled applications is substantial, many countries have approached their adoption with caution. This hesitancy stems from concerns about risks and benefits, including model bias, data breaches, privacy violations, and the erosion of professional skills due to over-reliance on AI. Additional barriers include inadequate data infrastructure and a lack of robust evidence on the most effective forms and interfaces of AI to enhance service delivery and system performance. Moreover, the rapid rise of Generative AI has introduced a heightened urgency for the development of comprehensive governance and ethical frameworks. The Asian Development Bank (ADB) Health Practice Team and Digital Sector, in collaboration with the World Health Organization (WHO) Regional Office for the Western Pacific, are convening the Forum on Harnessing AI for Health Equity on 25–26 March at ADB Headquarters. Senior health officials from 16 developing member countries (DMCs), alongside leading researchers, innovators, and policymakers from across Asia and the Pacific, will engage in discussions on how AI can be governed, financed, and scaled to deliver healthier lives for all. The forum is funded through the High-Level Technology Fund with support from the Government of Japan. Objectives The two-day forum aims to:Identify high-impact, evidence-based health AI use cases suitable for resource-constrained settings and map out a pipeline for potential supportIntroduce and validate the WHO's policy guidance on minimum necessary ecosystem for countries and institutions to self-assess readiness for AI adoptionObtain initial commitment from participating stakeholders to join a pilot regional initiative focused on data/model sharing for a priority health challengeEndorse a prioritized set of actionable policy recommendations focusing on governance, regulation, and sustainable financing necessary to scale high-impact health AI use cases in resource-constrained settings Target participants The forum will convene participants from 16 ADB DMCs including Bangladesh, Bhutan, Cambodia, Cook Islands, Fiji, Indonesia, Lao People's Democratic Republic, Mongolia, Nepal, Palau, Papua New Guinea, Philippines, Sri Lanka, Thailand, Uzbekistan, and Viet Nam. In addition, health system leaders and practitioners, private sector technologists, academic experts, representatives from civil society organizations, development partners, and philanthropic institutions engaged in AI for Health from Australia, the People’s Republic of China, Japan, Malaysia, New Zealand, Singapore, and the Republic of Korea will also be invited. SessionsDay 1 (25 March)Session 1: AI Use Cases for the Care Continuum: Improving Access and QualitySession 2: Adopting a risk-based approach for AI use case prioritizationSession 3: Establishing the Minimum Ecosystem for Implementation Risk ManagementSession 4: Future-Ready Health Workforce: Building Competency for Human-AI CollaborationDay 2 (26 March)Session 5: The Political Economy of AI in Health: Understanding Stakeholder Interests and their Impact on EquitySession 6: AI and Health Financing: Supporting Fair Coverage and ProtectionSession 7: AI Regulation and Oversight: Pathways for Different RealitiesSession 8: The Way Forward – Building Alliances and Regional Partnership How to registerThe forum will be held in a hybrid format. In-person attendance is by invitation only, while livestreaming of the forum will be available to the public. To register, please complete this form on or before 24 March 2026.Event organizers / partnersAsian Development Bank (ADB)WHO Regional Office for the Western Pacific High-Level Technology Fund (HLTF)ContactAkihito WatabeHealth Specialist, SD3-HSDawatabe@adb.org Jae Kyoun KimHealth Specialist, SD3-HSDjkkim@adb.org
NCAA Tournament 2026 bracket picks, upsets, Cinderella teams: Best model simulates March Madness 10K times
SportsLine's college basketball model has called 25 first-round NCAA Tournament upsets by double-digit seeds
NCAA Tournament 2026 bracket picks, upsets, Cinderella teams: Best model simulates March Madness 10,000 times
SportsLine's college basketball model has called 25 first-round NCAA Tournament upsets by double-digit seeds
NCAA Tournament 2026 bracket picks, upsets, Cinderella teams: Top model simulates March Madness 10,000 times
SportsLine's college basketball model has called 25 first-round NCAA Tournament upsets by double-digit seeds
NCAA Tournament 2026 bracket picks, upsets, plus Cinderella teams: Model simulates March Madness 10K times
SportsLine's college basketball model has called 25 first-round NCAA Tournament upsets by double-digit seeds
NCAA Tournament 2026 bracket picks, upsets, plus Cinderella teams: Model simulates March Madness 10,000 times
SportsLine's college basketball model has called 25 first-round NCAA Tournament upsets by double-digit seeds
NCAA Tournament 2026 bracket picks, upsets, and Cinderella teams: Model simulates March Madness 10,000 times
SportsLine's college basketball model has called 25 first-round NCAA Tournament upsets by double-digit seeds
NCAA Tournament 2026 bracket picks, best upsets, Cinderella teams: Model simulates March Madness 10,000 times
SportsLine's college basketball model has called 25 first-round NCAA Tournament upsets by double-digit seeds
NCAA Tournament 2026 bracket picks, top upsets, Cinderella teams: Model simulates March Madness 10,000 times
SportsLine's college basketball model has called 25 first-round NCAA Tournament upsets by double-digit seeds
NCAA Tournament 2026 bracket picks, Cinderella teams, upsets: Model simulates March Madness 10,000 times
SportsLine's college basketball model has called 25 first-round NCAA Tournament upsets by double-digit seeds
NCAA Tournament 2026 bracket picks, Cinderella teams, upsets: Model simulates March Madness 10K times
SportsLine's college basketball model has called 25 first-round NCAA Tournament upsets by double-digit seeds
NCAA Tournament 2026 bracket picks, Cinderella teams, best upsets: Model simulates March Madness 10K times
SportsLine's college basketball model has called 25 first-round NCAA Tournament upsets by double-digit seeds
NCAA Tournament 2026 bracket picks, Cinderella teams, best upsets: Model simulates March Madness 10,000 times
SportsLine's college basketball model has called 25 first-round NCAA Tournament upsets by double-digit seeds
NCAA Tournament 2026 bracket picks, Cinderella teams, top upsets: Model simulates March Madness 10,000 times
SportsLine's college basketball model has called 25 first-round NCAA Tournament upsets by double-digit seeds
Former supermodel says business managers stole millions
Kathy Ireland and her husband Greg Olsen filed suit in Santa Barbara against their longtime managers.
Dartmouth Health launches TeamBirth to enhance patient experience during labor and delivery
Patient-centered care model supports open communication among patients, their support people, and clinicians during childbirth
Target to invest another $2 billion in its business this year to reverse its sales malaise
MINNEAPOLIS (AP) — Target is investing another $2 billion in its business this year to spruce up its store experience, remodel stores and invest in its workers as it tries to turn around a persistent sales malaise and reclaim its authority on style.
Google germinates Gemini 3.1 Pro in ongoing AI model race
: AI model said to show improved reasoning capabilities
6 biggest takeaways from Tesla's Q4 earnings call
Tesla CEO Elon Musk announced a $2B investment into xAI and the discontinuation of Model S and Model X in order to support Optimus robot production.
Rams vs. Falcons prediction, odds: 'Monday Night Football' picks from 10,000 simulations
SportsLine's model simulated Los Angeles vs. Atlanta on 'MNF' 10,000 times and reveals its NFL picks today
2025-26 college football bowl confidence picks, predictions, computer rankings
SportsLine's model simulates every game 10,000 times and reveals bowl confidence pool and CFP picks
2025-26 college football bowl confidence picks, predictions, computer projections
SportsLine's model simulates every game 10,000 times and reveals bowl confidence pool and CFP picks
NFL Week 17 odds, betting lines, spreads: Picks, predictions, best bets today
SportsLine's model simulates every game 10,000 times and reveals its NFL picks today
NFL Week 17 odds, lines, spreads: Picks, predictions, best bets today
SportsLine's model simulates every game 10,000 times and reveals its NFL picks today
Supernova immersion model suggests Earth-like planets are more common in the universe
Rocky planets like our Earth may be far more common than previously thought, according to new research published in the journal Science Advances. It suggests that when our solar system formed, a nearby ...
Disney’s Entertainment Model Stumbles as Amazon Doubles Down on AI Infrastructure Spending
Amazon (NASDAQ: AMZN) and Disney (NYSE: DIS) just reported earnings revealing two companies moving in opposite directions. Amazon delivered 13.4% revenue growth and accelerating cloud momentum. Disney posted flat revenue, down 0.5% year-over-year, as streaming gains failed to offset legacy media weakness. Is it odd that I’m comparing what are now two drastically different companies ... Disney’s Entertainment Model Stumbles as Amazon Doubles Down on AI Infrastructure Spending
Chargers vs. Eagles prediction, odds: 'Monday Night Football' picks today from 10,000 simulations
SportsLine's model simulated Philadelphia vs. Los Angeles 10,000 times and revealed its 'MNF' best bets
Assessing Goldman Sachs After a 301% Five Year Surge and Business Model Shift
If you are wondering whether Goldman Sachs Group is still a smart buy after its huge run up, or if you are late to the party, you are in the right place to unpack what the current share price actually implies. The stock has climbed 3.5% over the last week, 8.7% over the past month, and is up a striking 48.6% year to date, capping a 45.4% 1 year gain and a 301.0% increase over 5 years that has clearly shifted how the market views its risk and growth profile. Recent headlines have highlighted...
NFL odds, lines, spreads: Week 14 predictions, best picks today
SportsLine's model simulates every game 10,000 times and reveals NFL picks
NFL odds, betting lines, spreads: Week 14 picks, best predictions today
SportsLine's model simulates every game 10,000 times and reveals NFL picks
Does Recent Business Model Shift Make DocuSign a Better Value for Investors in 2025?
Curious if DocuSign is trading at a bargain or if recent changes have made it less appealing? Let’s break down what has been happening and what it could mean for anyone eyeing the stock. DocuSign’s stock has experienced some volatility recently, climbing 5.2% in the last week, but still down 23.2% year-to-date and 13.6% over the past 12 months. Analyst discussion has focused on DocuSign’s evolving business model and strategic partnerships, both of which have been making headlines and...