DHS Plans Experiment Running ‘Reconnaissance’ Drones Along the US-Canada Border

May 4, 2026 Announcements Building a new enterprise AI services company with Blackstone, Hellman & Friedman, and Goldman Sachs

Building a new enterprise AI services company with Blackstone, Hellman & Friedman, and Goldman Sachs Anthropic, Blackstone, Hellman & Friedman, and Goldman Sachs announced the formation of a new AI services company. The organization will work with mid-sized companies across sectors to bring Claude into their most important operations. Applied AI engineers from Anthropic will work alongside the firm’s engineering team to identify where Claude can have the most impact, build custom solutions, and support customers over the long term. Alongside the founding partners, the new company is backed by a consortium of leading alternative asset managers including General Atlantic, Leonard Green, Apollo Global Management, GIC, and Sequoia Capital. Why we’re building this Putting Claude to work in an organization’s core operations takes hands-on engineering and deep familiarity with how each business runs. Systems integrators in the Claude Partner Network…

Apr 28, 2026 Announcements Claude for Creative Work

Claude for Creative Work Creative professionals look to technology to expand what's possible in their work. Claude can't replace taste or imagination, but it can open up new ways of working—faster and more ambitious ideation, a more expansive skill set, and the ability for creatives to take on larger-scale projects. AI can also help shoulder the parts of the creative process that eat up time by handling repetitive tasks and eliminating manual toil. Key to both these goals is integrating Claude into the tools the creative industry already knows and trusts. Today, with a coalition of partners including Blender, Autodesk, Adobe, Ableton, and Splice, we’re releasing a set of connectors—tools that let Claude work alongside the software creative professionals rely on, so creatives can extend their reach. Connecting Claude to creative tools Connectors allow Claude to access other platforms and…

Apr 27, 2026 Announcements Anthropic names Theo Hourmouzis General Manager of Australia & New Zealand and officially opens Sydney office

Anthropic names Theo Hourmouzis General Manager of Australia & New Zealand and officially opens Sydney office Theo Hourmouzis is joining Anthropic as General Manager of Australia and New Zealand, marking the next step in our investment in the region. Hourmouzis will meet with customers and partners this week alongside executives from our global team, as we officially open our Sydney office. Hourmouzis brings more than 20 years of leadership experience in the technology industry across Asia Pacific to the role. He joins us from Snowflake, where he most recently served as Senior Vice President for Australia, New Zealand and ASEAN, helping enterprise and public sector organisations across financial services, retail, aviation and government move AI from experimentation to business impact. At Anthropic, he'll lead our growing local team and shape a strategy built around Australian and New Zealand customers, bringing…

Apr 24, 2026 Announcements An update on our election safeguards

An update on our election safeguards People around the world turn to Claude for information about political parties, candidates, and the issues at stake during election time—as well as to answer simpler questions like when, where, and how to vote. In our view, if AI models can answer these questions well (that is, accurately and impartially), they can be a positive force for the democratic process. Here, we explain what we’re doing to help Claude meet the mark ahead of the US midterms and other major elections around the world this year. Measuring and preventing political bias When people ask Claude about political topics, they should get comprehensive, accurate, and balanced responses—responses that help them reach their own conclusions, rather than steer them toward a particular viewpoint. That’s why we train Claude to treat different political viewpoints with equal depth,…

AI invades Princeton, where 30% of students cheat—but peers won't snitch

Pity poor Princeton. The ultra-elite university has a mere $38 billion in endowment money. Many of its dorms lack air conditioning. And it’s in New Jersey. I kid about New Jersey, of course. Despite not being allowed to pump one’s own gas there, the “Garden State” grew on me during three years spent in the Princeton area. I still keep up with its goings-on, which led me to this week’s article in the Daily Princetonian on how AI was disrupting the university’s long-running traditions. Although a beautiful place, Princeton is also extremely competitive; before one heads up to New York to become a captain of finance, one needs to succeed in the classroom. And when everyone else in the classroom is a genius, cheating becomes a real option to stay ahead, especially in the sciences. In a 2025 survey of…

Anthropic blames dystopian sci-fi for training AI models to act “evil”

Those with an interest in the concept of AI alignment (i.e., getting AIs to stick to human-authored ethical rules) may remember when Anthropic claimed its Opus 4 model resorted to blackmail to stay online in a theoretical testing scenario last year. Now, Anthropic says it thinks this “misalignment” was primarily the result of training on “internet text that portrays AI as evil and interested in self-preservation.” In a recent technical post on Anthropic’s Alignment Science blog (and an accompanying social media thread and public-facing blog post), Anthropic researchers lay out their attempts to correct for the kind of “unsafe” AI behavior that “the model most likely learned… through science fiction stories, many of which depict an AI that is not as aligned as we would like Claude to be.” In the end, the model maker says the best remedy for…

Altman forced to confront claims at OpenAI trial that he's a prolific liar

Elon Musk and Sam Altman had very different experiences while testifying at a trial that will determine OpenAI’s future, including who runs it, where its research funding comes from, and who can profit from its boldest new technologies. Musk—who filed the lawsuit alleging that OpenAI under its current leadership has abandoned its nonprofit mission to build AI that benefits humanity and instead serves to enrich people like Altman—spent three grueling days on the stand. At times, he lost his temper, as OpenAI’s lawyer, William Savitt, tried to poke holes in Musk’s claims that OpenAI executives teamed up with Microsoft to “steal a charity” after duping Musk into donating $38 million in early funding. On Tuesday, Altman did not face such a grilling from Musk’s lawyer, Steven Molo. Instead, Altman appeared jittery at first but steeled his nerves rather quickly. He…

“Will I be OK?” Teen died after ChatGPT pushed deadly mix of drugs, lawsuit says

OpenAI is facing down another wrongful-death lawsuit after ChatGPT told a 19-year-old, Sam Nelson, to take a lethal mix of Kratom and Xanax. According to a complaint filed on behalf of Nelson’s parents, Leila Turner-Scott and Angus Scott, Nelson trusted ChatGPT as a tool to “safely” experiment with drugs after using the chatbot for years as a go-to search engine when he was in high school. The teen viewed ChatGPT so highly as an authoritative source of information that he once swore to his mom that ChatGPT had access to “everything on the Internet,” so it “had to be right,” when she questioned if the chatbot was always reliable, the complaint said. But Nelson’s confidence in ChatGPT ended up being dangerously misplaced. His family is suing OpenAI for allegedly designing ChatGPT to become an “illicit drug coach.” Nelson’s death by…

Course correction: Google to link more sources in AI Overviews

The top of a Google search page is prime real estate, but it has primarily been the domain of AI Overviews for the past two years. Websites that spent years optimizing for Google search haven’t exactly loved being pushed down the page by a chatbot and may blame AI Overviews for recent traffic drops. Google is not admitting fault, but it is rolling out a number of changes that will place more links to websites inside AI answers. Google says many AI Overviews are “just the beginning of exploring a topic you’re interested in.” To support this supposed yearning to know more, AI Overviews and AI Mode will soon get a new section at the bottom called “Further Exploration.” The new exploration box will link to articles and analysis that is relevant to the query in a bullet point list.…

Google's Gemma 4 AI models get 3x speed boost by predicting future tokens

Google launched its Gemma 4 open models this spring, promising a new level of power and performance for local AI. Google’s take on edge AI could be getting even faster already with the release of Multi-Token Prediction (MTP) drafters for Gemma. Google says these experimental models leverage a form of speculative decoding to take a guess at future tokens, which can speed up generation compared to the way models generate tokens on their own. The latest Gemma models are built on the same underlying technology that powers Google’s frontier Gemini AI, but they’re tuned to run locally. Gemini is optimized to run on Google’s custom TPU chips, which operate in enormous clusters with super-fast interconnects and memory. A single high-power AI accelerator can run the largest Gemma 4 model at full precision, and quantizing will let it run on a…

Google DeepMind partners with EVE Online for AI model testing

Google’s AI-focused DeepMind division has taken a minority stake in the developer of popular sci-fi simulation EVE Online, saying it will use the game to study “intelligence in complex, dynamic, player-driven systems.” The research partnership comes as the management behind EVE Online developer CCP Games announced that they have spent $120 million to buy themselves out from their former owners at South Korean publisher Pearl Abyss (Crimson Desert). The newly independent entity is being rebranded as Fenris Creations, which will continue to operate as normal without any restructuring or layoffs, the company said. “Something that already behaves like a living world” In today’s announcement, Fenris and DeepMind said that EVE Online presents “a uniquely rich environment for study,” especially when it comes to developing AI systems that use “long-horizon planning, memory, and continual learning.” DeepMind says it will conduct controlled…

Influential study touting ChatGPT in education retracted over red flags

A study that claimed OpenAI’s ChatGPT can positively impact student learning has been retracted nearly one year after publication. The journal publisher, Springer Nature, cited “discrepancies” in the analysis and a lack of confidence in the conclusions—but not before the paper racked up hundreds of citations and made the rounds on social media. “The paper’s authors made some very attention-grabbing claims about the benefits of ChatGPT on learning outcomes,” said Ben Williamson, a senior lecturer at the Centre for Research in Digital Education and the Edinburgh Futures Institute at the University of Edinburgh in Scotland, in an email to Ars. “It was treated by many on social media as one of the first pieces of hard, gold standard evidence that ChatGPT, and generative AI more broadly, benefits learners.” The retracted paper attempted to quantify “the effect of ChatGPT on students’…

Study: AI models that consider user's feeling are more likely to make errors

In human-to-human communication, the desire to be empathetic or polite often conflicts with the need to be truthful—hence terms like “being brutally honest” for situations where you value the truth over sparing someone’s feelings. Now, new research suggests that large language models can sometimes show a similar tendency when specifically trained to present a “warmer” tone for the user. In a new paper published this week in Nature, researchers from Oxford University’s Internet Institute found that specially tuned AI models tend to mimic the human tendency to occasionally “soften difficult truths” when necessary “to preserve bonds and avoid conflict.” These warmer models are also more likely to validate a user’s expressed incorrect beliefs, the researchers found, especially when the user shares that they’re feeling sad. How do you make an AI seem “warm”? In the study, the researchers defined the…

Researchers try to cut the genetic code from 20 to 19 amino acids

The genetic code is central to life. With minor variations, everything uses the same sets of three DNA bases to encode the same 20 amino acids. We have discovered no major exceptions to this, leading researchers to conclude that this code probably dated back to the last common ancestor of all life on Earth. But there has been a lot of informed speculation about how that genetic code initially evolved. Most hypotheses suggest that earlier forms of life had partial genetic codes and used fewer than 20 amino acids. To test these hypotheses, a team from Columbia and Harvard decided to see if they could get rid of one of the 20 currently in use. And, as a first attempt, they engineered a portion of the ribosome that worked without using an otherwise essential amino acid: isoleucine. Changing the code…

Meta cuts contractors who reported seeing Ray-Ban Meta users have sex

In February, numerous workers from a company that Meta contracted to perform data annotation for Ray-Ban Meta reported viewing sensitive, embarrassing, and seemingly private footage recorded by the smart glasses. About two months later, Meta ended its contract with the firm. According to a BBC report today, “less than two months” after a report from Swedish newspapers Svenska Dagbladet and Göteborgs-Posten and Kenya-based freelance journalist Naipanoi Lepapa came out featuring Sama workers complaining about watching explicit footage shot from Ray-Ban Metas, “Meta ended its contract with Sama.” Sama is a Kenya-headquartered firm that Meta contracted to perform data annotation work, including working with video, image, and speech annotation for Meta’s AI systems for Ray-Ban Metas. Sama claims that Meta’s cancellation of the contract affected 1,108 workers. A Meta spokesperson told BBC that Meta “decided to end our work with Sama…

Humanoid robots start sorting luggage in Tokyo airport test amid labor shortage

Humanoid robots are getting a new gig as baggage handlers and cargo loaders at Tokyo’s Haneda Airport—part of a Japan Airlines experiment to address a human labor shortage as airport visitor numbers have surged in recent years. The demonstration, set to launch in May 2026, could eventually test humanoid robots in a wide range of airport tasks, including cleaning aircraft cabins and possibly handling ground support equipment such as baggage carts, according to a Japan Airlines press release. The trials are scheduled to run until 2028, which suggests that travelers flying into or out of Tokyo may spot some of the robots at work. This marks the latest foray for humanoid robots after they have already begun pilot-testing in workplaces such as automotive factories and warehouses. Most robotic productivity so far has relied on robotic arms and similarly specialized robots…

Halliburton enhances seismic workflow creation with Amazon Bedrock and Generative AI

Artificial Intelligence Halliburton enhances seismic workflow creation with Amazon Bedrock and Generative AI Seismic data analysis is an essential component of energy exploration, but configuring complex processing workflows has traditionally been a time-consuming and error-prone challenge. Halliburton’s Seismic Engine, a cloud-native application for seismic data processing, is a powerful tool that previously required manual configuration of approximately 100 specialized tools to create workflows. This process was not only time-consuming but also required deep expertise, potentially limiting the accessibility and efficiency of the software. To address this challenge, Halliburton partnered with the AWS Generative AI Innovation Center to develop an AI-powered assistant for Seismic Engine. The solution uses Amazon Bedrock, Amazon Bedrock Knowledge Bases, Amazon Nova, and Amazon DynamoDB to transform complex workflow creation into conversations. Geoscientists and data scientists can configure processing tools through natural language interaction instead of manual…

Streamlining generative AI development with MLflow v3.10 on Amazon SageMaker AI

Artificial Intelligence Streamlining generative AI development with MLflow v3.10 on Amazon SageMaker AI Today, we’re excited to announce that Amazon SageMaker AI MLflow Apps now support MLflow version 3.10, bringing enhanced capabilities for generative AI development and streamlined experiment tracking to your generative AI workflows. Building on the foundations established with Amazon SageMaker AI MLflow Apps, this latest version introduces powerful new features for observability, evaluation, and generative AI development that help data scientists and ML engineers accelerate their AI initiatives from experimentation to production. In this post, we’ll explore what’s new in MLflow v3.10, walk you through getting started with SageMaker AI MLflow Apps, and how to leverage these enhancements to build generative AI applications. What’s new in MLflow v3.10 MLflow 3.10 introduces a set of targeted improvements to the MLflow ecosystem that extend the tracing and observability capabilities…

How Hapag-Lloyd uses Amazon Bedrock to transform customer feedback into actionable insights

Artificial Intelligence How Hapag-Lloyd uses Amazon Bedrock to transform customer feedback into actionable insights Hapag-Lloyd stands as one of the world’s leading liner shipping companies, operating a modern fleet of 313 container ships with a total transport capacity of 2.5 million TEU (Twenty-foot Equivalent Unit—a standard unit of measurement for cargo capacity in container shipping). The company maintains a container capacity of 3.7 million TEU, which includes one of the industry’s largest and most modern fleets of reefer containers. With approximately 14,000 employees in the Liner Shipping Segment and more than 400 offices spread across 140 countries, Hapag-Lloyd maintains a robust global presence. Through 133 liner services worldwide, we facilitate reliable connections between more than 600 ports across the continents. The company’s Digital Customer Experience and Engineering team, distributed between Hamburg and Gdańsk, drives digital innovation by developing and maintaining…

Generate dashboards from natural language prompts in Amazon Quick

Artificial Intelligence Generate dashboards from natural language prompts in Amazon Quick Building meaningful dashboards demands hours of manual setup, even for experienced BI professionals. Amazon Quick now generates complete multi-sheet dashboards from natural language prompts, taking you from one or more datasets to a production-ready analysis in minutes. Data analysts building recurring operations reports, program managers preparing a leadership review, or engineers exploring a new dataset can describe what they want, and Amazon Quick produces multiple organized sheets with visuals selected for your data, filter controls for stakeholders to explore by different dimensions, and calculated fields such as year-over-year growth and month-over-month comparisons. Before generating, you review and edit an interactive plan of the proposed structure, keeping you in control of the final output. In Amazon Quick, Analysis is the authoring surface where you build and arrange visuals, filters, and…

Introducing agent quality optimization in AgentCore, now in preview

Artificial Intelligence Introducing agent quality optimization in AgentCore, now in preview Generate recommendations from production traces, validate them with batch evaluation and A/B testing, and ship with confidence. AI agents that perform well at launch don’t stay that way. As models evolve, user behavior shifts, and prompts get reused in new contexts they were never designed for. Agent quality quietly degrades. In most teams, the improvement process still looks the same: without automatic feedback loops, when a user complains, a developer reads through traces, forms a hypothesis, rewrites the prompt, tests a handful of cases, and ships the fix. Then the cycle repeats, often introducing a new issue for a different user. Up until today, Amazon Bedrock AgentCore provided the pieces for you to debug it manually or build custom implementations: check the evaluation scores to detect quality drop, deep…

Extracting contract insights with PwC’s AI-driven annotation on AWS

Artificial Intelligence Extracting contract insights with PwC’s AI-driven annotation on AWS This post was co-written with Yash Munsadwala, Adam Hood, Justin Guse, and Hector Hernandez from PwC. Contract analysis often consumes significant time for legal, compliance, and procurement teams, especially when important insights are buried in lengthy, unstructured agreements. As contract volumes grow, finding specific clauses and assessing extracted terms can become increasingly difficult to scale. Today, many teams rely primarily on keyword and pattern-based extraction or contract management systems to analyze contracts. While these methods can work, they often fall short of providing consistent insights at a scale. As a result, many teams are exploring AI-based approaches that can combine large language models (LLMs) with automated extraction workflows. PwC’s AI-driven annotation (AIDA) solution, built on AWS, can extract structured insights from contracts through rule-based extraction and natural language queries.…

Case Study - Cognition x Cerebras December 10, 2025

Dec 10 2025 Case Study - Cognition x Cerebras The Dawn of Real-Time Coding Agents TL;DR Powered by Cerebras Inference, Cognition's SWE-1.6 and the SWE-grep family deliver frontier-level coding performance up to ~5x faster than on GPU, with a smoother agent experience that keeps developers in flow while they explore codebases, ship features, and debug complex systems. The Challenge AI is redefining software development, turning natural language prompts into working code. But for an AI coding assistant to be useful, it must feel instantaneous and handle large, complex projects seamlessly. Until now, AI coding on GPU meant frustrating delays - 20 to 30 second generation times that broke a developer's concentration. Even slight lags forced context-switching. Developers were stuck choosing between smaller, faster models that lacked skill and larger models that were too slow. The industry needed a solution that…

Find out how AlphaEvolve has gone from research to solving real-life problems.

A year ago, we introduced AlphaEvolve, a Gemini-powered evolutionary algorithm agent that iteratively discovers optimized algorithms for complex problems. It’s advanced decades-old math problems, and now it’s helping tackle major global challenges, too. Over the past year, AlphaEvolve has become a powerful engine for scientific and societal progress. It helped improve DNA sequencing error correction, increased the accuracy of disaster predictions and demonstrated the potential to stabilize power grids in simulations. It’s also accelerating scientific discovery, helping researchers run complex molecular simulations and unlocking new insights in neuroscience. Beyond research, AlphaEvolve is driving real business results. It’s making Google’s own infrastructure more efficient and helping Google Cloud customers improve their machine learning models, accelerate drug discovery, improve supply chains and optimize warehouse design. In the future, we plan to expand these capabilities to reliably bring the power of self-improving algorithms…

Adding Benchmaxxer Repellant to the Open ASR Leaderboard

Adding Benchmaxxer Repellant to the Open ASR Leaderboard TLDR: Appen Inc. and DataoceanAI have provided high-quality English ASR datasets covering scripted and conversational speech over multiple accents. To prevent potential risks of benchmaxxing or test-set contamination, we will keep these datasets private for a high-quality measure of performance on multiple tasks. We’re not updating the average WER at this time: by default, the leaderboard’s Average WER remains computed on public datasets only. You can optionally include the private datasets using the toggle to see their impact 👀 Since its launch in September 2023, the Open ASR Leaderboard has been visited over 710K times. We’re blown away by the community’s interest and motivation to keep pushing speech recognition 🗣️ Two words sum up the objectives (but also challenges) in maintaining a benchmark like the Open ASR Leaderboard: Standardization: models can have…

Advancing AI for materials with MatterSim: experimental synthesis, faster simulation, and multi-task models

At a glance - Experimental validation: Using high-throughput screening with MatterSim-v1, we previously identified tetragonal tantalum phosphorus (TaP) as a potential high-performance thermal conductor. Now we have experimentally synthesized it and measured its thermal conductivity (152 W/m/K) to be close to the thermal conductivity of silicon. - Faster simulation: We have accelerated MatterSim-v1 model inference by 3-5x and integrated it with the LAMMPS software package, enabling large-scale simulations across multiple GPUs. - New model release: We are introducing MatterSim-MT, a multi-task foundation model for in silico materials characterization that enables the simulation of complex, multi-property phenomena beyond what potential energy surfaces alone can capture. Materials design underpins a wide range of technological advances, from nanoelectronics to semiconductor design and energy storage. Yet development cycles for novel materials remain slow and costly. Universal machine learning interatomic potentials aim to accelerate the…

SocialReasoning-Bench: Measuring whether AI agents act in users’ best interests

At a glance - AI agents are moving into social contexts. When agents manage calendars, negotiate purchases, or interact with other agents on a user’s behalf, they need more than task competence—they need social reasoning. - SocialReasoning-Bench evaluates that ability. The benchmark tests whether an agent can negotiate for a user in two realistic settings: Calendar Coordination and Marketplace Negotiation. - The benchmark measures both outcomes and process: it scores agents on outcome optimality (how much value they secure for the user) and due diligence (whether they follow a competent decision-making process). - Current frontier models often leave value on the table. They usually complete the task, but they frequently accept suboptimal meeting times or poor deals instead of advocating effectively for the user. - Prompting helps, but it is not enough. Even with explicit guidance to act in the…

Building realistic electric transmission grid dataset at scale: a pipeline from open dataset

At a glance - We construct geographically grounded, electrically coherent power grid models entirely from publicly available data and release a dataset spanning 48 U.S. states and multi-state interconnections. - The models support AC optimal power flow (AC‑OPF) analysis, enabling physics-based study of congestion, capacity, and demand siting without restricted data. - We demonstrate applications including transmission expansion potential, targeted line upgrades, and placement of large datacenter loads. Microsoft Research is excited to release an open dataset of approximate transmission topology of the U.S. power grid derived from publicly available data. The ability to study transmission-level power grid behavior is essential for modern power systems research. Analyses of congestion, transmission expansion, demand growth, and system resilience all depend on network models with realistic topology, electrical parameters, and geographic grounding. In most of the world, including the United States, realistic transmission-level…

Microsoft at NSDI 2026: Advances in large-scale networked systems

Large-scale networked systems underpin cloud computing, AI, and distributed applications and services. The USENIX Symposium on Networked Systems Design and Implementation 2026 (opens in new tab) (NSDI ’26) is a leading forum where researchers and practitioners share new research, insights, and advances in the design and operation of these systems. Microsoft is proud to support NSDI ’26 as a returning sponsor, reflecting our ongoing commitment to advancing systems and networking research and engaging with the broader community. Microsoft researchers and engineering leaders are also serving on the program committee and in other organizational roles. This year, 11 papers by Microsoft authors and collaborators were accepted to the conference, spanning datacenter and wide-area networks, AI systems, and cloud infrastructure. Together, they highlight advances in building and operating large-scale networked systems. Spotlight: Event Series Technical sessions Monday, May 4, 2:00–3:20 PM DroidSpeak:…

Red-teaming a network of agents: Understanding what breaks when AI agents interact at scale

At a glance - Some risks appear only when agents interact, not when tested alone. Actions that seem harmless can cascade causing a chain reaction across an agent network. - In our tests, a single malicious message passed from agent to agent, extracting private data at each step and pulling uninvolved agents into the chain. - We saw early signs that some agent networks become more resistant to these attacks, but defenses are still an open challenge being worked on. Agents belonging to different users and organizations are beginning to interact with each other. These networks of agents are emerging as advances in large language models (LLMs) and silicon lower barriers to building agents, while tools like Claude, Copilot, and ChatGPT, along with existing platforms such as email and GitHub, bring them into constant contact. As a result, agents are…

A plan to make drugs in orbit is going commercial

A plan to make drugs in orbit is going commercial United Therapeutics is collaborating with Varda Space Industries to test pharmaceuticals in outer space. Varda Space Industries, a startup that’s been pitching its ability to perform drug experiments in space, says it has signed up the pharmaceutical company United Therapeutics in what may be remembered as a notable step toward in-orbit manufacturing. The idea of building things in outer space for use on Earth has so far been explored mostly on board the International Space Station, and only in small-scale experiments backed by governments. But Varda, based in El Segundo, California, is now telling drug companies it has a practical, and repeatable, way to produce novel molecules in microgravity. “This is the first commercial path to products made in space,” says Michael Reilly, Varda’s chief strategy officer. The scientific idea…

The Download: a Nobel winner on AI, and the case for fixing everything

The Download: a Nobel winner on AI, and the case for fixing everything Plus: the first zero-day exploit built by AI has been discovered. This is today's edition of The Download, our weekday newsletter that provides a daily dose of what's going on in the world of technology. Three things in AI to watch, according to a Nobel-winning economist A few months before he won the Nobel Prize in economics in 2024, Daron Acemoglu published a paper that earned him few fans in Silicon Valley. He argued that AI would give only a small boost to US productivity and would not eliminate the need for human work. Two years later, Acemoglu’s measured take has not caught on. The technology has advanced quite a bit since his cautious predictions, but the data is still largely on his side. MIT Technology Review…

Implementing advanced AI technologies in finance

Sponsored Implementing advanced AI technologies in finance Successful AI implementation requires shifts in workplace culture as well as use cases that can scale across the enterprise. In partnership withOracle NetSuite In finance departments that have long been defined by precision and control, AI has arrived less as a neatly managed upgrade than as a quiet insurgency. Employees are already using it while leadership races to impose structure, governance, and strategy after the fact. The result is a paradox: one of the most tightly regulated functions in the enterprise is now among the most experimentally transformed. What’s emerging is a layered shift in how work gets done. From variance commentary and fraud detection to contract review and close narrative drafting, AI is embedding itself across workflows, particularly where unstructured data once slowed down everything. Yet, as Glenn Hopper, head of AI…

Fostering breakthrough AI innovation through customer-back engineering

Sponsored Fostering breakthrough AI innovation through customer-back engineering Agentic AI is helping organizations completely reimagine core banking processes and operations from the customer perspective, rather than simply making incremental improvements. In partnership withCapital One Despite years of digitization, organizations capture less than one-third of the value expected from digital investments, according to McKinsey research. That’s because most big companies begin with technological capabilities and bolt applications onto them, rather than starting with customer needs and working backward to technology solutions. Not prioritizing the customer can create fragmented solutions; disjointed customer experiences; and ultimately, failed transformations. Organizations that achieve outsized results from AI flip the script. They adopt a “customer-back engineering” mindset, putting customers at the heart of technology transformation. It’s a strategy in which products and services are developed with the customer experience first in mind, including the customers’ challenges,…

Three things in AI to watch, according to a Nobel-winning economist

Three things in AI to watch, according to a Nobel-winning economist Daron Acemoglu is more cautious than most about predictions of a jobs apocalypse. Here’s what’s worrying him instead. This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here. A few months before he was awarded the Nobel Prize in economics in 2024, Daron Acemoglu published a paper that earned him few fans in Silicon Valley. Contrary to what Big Tech CEOs had been promising—an overhaul of all white-collar work—Acemoglu estimated that AI would give only a small boost to US productivity and would not obviate the need for human work. It’s okay at automating certain tasks, he wrote, but some jobs will be perfectly fine. Two years later, Acemoglu’s measured take has not caught on.…

The Download: the tech reshaping IVF and the rise of balcony solar

The Download: the tech reshaping IVF and the rise of balcony solar Plus: After years of insults, Anthropic and SpaceX have teamed up. This is today's edition of The Download, our weekday newsletter that provides a daily dose of what's going on in the world of technology. What’s next for IVF IVF has brought millions of babies into the world over the last four decades. But the process can still be slow, painful, and expensive—and far from guaranteed to work. Now, a wave of new technologies aims to change that. Researchers are using AI to identify promising sperm and embryos, developing robotic systems that could automate parts of the IVF process, and even exploring controversial genetic editing techniques designed to prevent inherited disease. The technologies could make IVF more effective and accessible. But they’re also raising difficult ethical questions about…

Trump’s mass firing just dealt another blow to American science

Trump’s mass firing just dealt another blow to American science Ambitious research is on the chopping block following yet more cuts at the National Science Foundation. This past week delivered another gut punch for science in the US. This time, the target was the National Science Foundation—a federal agency that funds major research projects to the tune of around $9 billion. The foundation’s efforts were overseen by a board of 22 prominent scientists. On Friday last week, they were all fired. The NSF has been without a director since April 2025, when former director Sethuraman Panchanathan stepped down in the wake of DOGE-led funding cuts and mass firings. Trump’s nominee for the role is Jim O’Neill, an investor and longevity enthusiast who does not have a science background. It’s hard to predict exactly how things will shake out for science.…

Inexpensive seafloor-hopping submersibles could stoke deep-sea science—and mining

Inexpensive seafloor-hopping submersibles could stoke deep-sea science—and mining A company called Orpheus Ocean wants to go “deep for cheap.” Smack dab between Australia and South America, the US National Oceanic and Atmospheric Administration (NOAA) research vessel Rainier is currently on a mission to map more than 8,000 square nautical miles of the Pacific seafloor in search of critical mineral deposits. But it isn’t doing it alone; for a month starting this week, it will deploy two oblong neon submersibles as the project’s special agents, sending them nearly 6,000 meters down to hop along the seafloor. The submersibles, built by the young company Orpheus Ocean, are designed to explore just this environment: a squelchy substrate that teems with life of all kinds, from tiny microbes to worms and snails, along with egg-size “nodules” of metals—such as copper, cobalt, nickel, and manganese—that…

Operationalizing AI for Scale and Sovereignty

Sponsored Operationalizing AI for Scale and Sovereignty Companies are taking control of their own data to tailor AI for their needs. The challenge lies in balancing ownership with the safe, trusted flow of high‑quality data needed to power reliable insights. This conversation from MIT Technology Review's EmTech AI conference examines how AI factories unlock new levels of scale, sustainability, and governance—positioning data control as a strategic imperative for governments and enterprises. About the speakers Chris Davidson, Vice President, HPC & AI Customer Solutions, HPE Chris Davidson is Vice President of HPC & AI Customer Solutions at Hewlett Packard Enterprise. He leads HPE’s global strategy for AI Factory solutions and Sovereign AI, working with governments, enterprises, and research institutions to build secure, scalable national- and enterprise-grade AI capabilities. He also directs Product Management and Performance Engineering across HPE’s HPC and AI…

The Download: the North Pole’s future and humanoid data

The Download: the North Pole’s future and humanoid data Plus: Google, Microsoft, Amazon and Meta have all set AI spending records. This is today's edition of The Download, our weekday newsletter that provides a daily dose of what's going on in the world of technology. Digging for clues about the North Pole’s past In the past, getting to the North Pole involved a treacherous trip through ice many meters thick. But last year, a research vessel encountered open water and thin ice, which created an easy passage. It provided a reminder of how quickly the Arctic is changing. Now scientists are digging deep below the seabed to find out if the Arctic Ocean was ever ice-free—and what that could mean for the future of Earth’s northernmost waters. Here’s what they hope to discover. —Tim Kalvelage This story is from the…

This startup’s new mechanistic interpretability tool lets you debug LLMs

This startup’s new mechanistic interpretability tool lets you debug LLMs Goodfire wants to make training AI models more like good old-fashioned software engineering. The San Francisco–based startup Goodfire just released a new tool, called Silico, that lets researchers and engineers peer inside an AI model and adjust its parameters—the settings that determine a model’s behavior—during training. This could give model makers more fine-grained control over how this technology is built than was once thought possible. Goodfire claims Silico is the first off-the-shelf tool of its kind that can help developers debug all stages of the development process, from building a data set to training a model. The company says its mission is to make building AI models less like alchemy and more like a science. Sure, LLMs like ChatGPT and Gemini can do amazing things. But nobody knows exactly how…

Exclusive eBook: Inside the stealthy startup that pitched brainless human clones

Exclusive eBook: Inside the stealthy startup that pitched brainless human clones Access a subscriber-only eBook on a startling and fairly graphic pursuit of human longevity that poses concerns about the ethics of cloning. This ebook is available only for subscribers. The ultimate plan to live forever is a brand new body. This subscriber-only eBook explores R3 Bio, a small startup that has pitched a startling and ethically charged vision for "brainless clones" to serve the role of backup human bodies. by Antonio Regalado March 20, 2026 Related Stories: - Inside the stealthy startup that pitched brainless human clones - This researcher wants to replace your brain, little by little - Stem-cell therapies that work: 10 Breakthrough Technologies 2025 Access all subscriber-only eBooks: Keep Reading Most Popular OpenAI is throwing everything into building a fully automated researcher An exclusive conversation with…

The Download: supercharged scams and studying AI healthcare

The Download: supercharged scams and studying AI healthcare Plus: DeepSeek has unveiled its long-awaited new AI model. This is today's edition of The Download, our weekday newsletter that provides a daily dose of what's going on in the world of technology. We’re in a new era of AI-driven scams When ChatGPT was released in late 2022, it showed how easily generative AI could create human-like text. This quickly caught the eye of cybercriminals, who began using LLMs to compose malicious emails. Since then, they’ve adopted AI for everything from turbocharged phishing and hyperrealistic deepfakes to automated vulnerability scans. Many organizations are now struggling to cope with the sheer volume of cyberattacks. AI is making them faster, cheaper, and easier to carry out, a problem set to worsen as more cybercriminals adopt these tools—and their capabilities improve. Read the full story…

Will fusion power get cheap? Don’t count on it.

Will fusion power get cheap? Don’t count on it. New research suggests that cost declines could be slow for the technology. Fusion power could provide a steady, zero-emissions source of electricity in the future—if companies can get plants built and running. But a new study suggests that even if that future arrives, it might not come cheap. Technologies tend to get less expensive over time. Lithium-ion batteries are now about 90% cheaper than they were in 2013. But historically, different technologies tend to go through this curve at different rates. And the cost of fusion might not sink as quickly as the prices of batteries or solar. It’s tricky to make any predictions about the cost of a technology that doesn’t exist yet. But when there’s billions of dollars of public and private funding on the line, it’s worth considering…

Accelerated X-Ray Analysis for Nanoscale Imaging (XANI) of Novel Materials

A massive-scale X-ray free-electron laser (XFEL) enables tracking structural and electron dynamics in novel systems, including fusion materials, semiconductors, batteries, and catalysis. It produces ultrashort X-ray pulses that can record the movements of atoms and electrons. These instruments can detect the smallest change in material structure caused by defects and other influences. The high repetition rate of these bright X-ray bursts can reach up to 1 million shots per second with 35-million-pixel cameras. The acquired multidimensional datasets contain rich physical information about the fastest microscopic movements of electrons and atoms, which can help identify defects in materials. Processing and analyzing these datasets to extract the physics has conventionally required more than nine months of computational time. XFEL research facilities include SwissFEL in Switzerland, Spring-8 Angstrom Compact free-electron Laser (SACLA) in Japan, Linac Coherent Light Source (LCLS-II) at SLAC, European XFEL…

Federated Learning Without the Refactoring Overhead Using NVIDIA FLARE

Federated learning (FL) is no longer a research curiosity—it’s a practical response to a hard constraint: the most valuable data is often the least movable. Regulatory boundaries, data sovereignty rules, and organizational risk tolerance routinely prevent centralized aggregation. Meanwhile, sheer data gravity makes even permitted transfers slow, expensive, and fragile at scale. The latest version of NVIDIA FLARE addresses this reality with a federated computing runtime that moves the training logic to the data, while raw data stays put. In high-stakes environments, centrally aggregating data is often not possible or practical, so a modern federated platform must treat data isolation, compliance, and privacy-enhancing technologies as first-class requirements. What has historically slowed adoption isn’t the concept of FL—it’s the developer experience. If the path from “my local script trains” to “my job runs across federated sites” requires deep refactoring, new class…

Winning a Kaggle Competition with Generative AI–Assisted Coding

In March 2026, three LLM agents generated over 600,000 lines of code, ran 850 experiments, and helped secure a first-place finish in a Kaggle playground competition. Success in modern machine learning competitions is increasingly defined by how quickly you can generate, test, and iterate on ideas. LLM agents, combined with GPU acceleration, dramatically compress this loop. Historically, two bottlenecks have limited this experimentation: - How quickly you can write code for new experiments. - How quickly you can execute those experiments. GPUs and libraries like NVIDIA cuDF, NVIDIA cuML, XGBoost, and PyTorch have largely solved the second problem. LLM agents now address the first problem—unlocking a new scale of rapid, iterative experimentation. This blog post describes how I used LLM agents to accelerate the discovery of the most performant tabular data prediction solutions. Case study: Kaggle Playground churn prediction The…

What Parameter Golf taught us about AI-assisted research

What Parameter Golf taught us Lessons from 1,000+ participants, 2,000+ submissions, and an open machine learning challenge shaped by coding agents. We launched Parameter Golf to engage and support the machine learning research community in exploring a new, tightly constrained machine learning problem. We wanted the challenge to be interesting enough to reward real technical creativity, while remaining conceptually simple and easy to verify. Participants had to minimize held-out loss on a fixed FineWeb dataset while staying within a 16 MB artifact limit, including both model weights and training code, and a 10-minute training budget on 8×H100s. We provided a baseline, dataset, and evaluation scripts so participants could fork the repo, improve the model, and submit their results through GitHub. Over the course of eight weeks, we received more than 2,000 submissions from over 1,000 participants. We were impressed by…

How ChatGPT adoption broadened in early 2026

How ChatGPT adoption broadened in early 2026 Q1 data shows consumer adoption growth across inferred gender, age, and geography. In the first quarter of 2026, consumer ChatGPT growth broadened across age groups, continued to rise among users with typically feminine names, and deepened in more countries. This analysis covers the messages sent on ChatGPT consumer plans (Free, Go, Plus, and Pro). Because it excludes Codex and ChatGPT enterprise and education products, it understates total workplace and educational usage. Users with typically feminine names represented a growing share of ChatGPT usage this quarter after reaching approximate parity last year. These users account for over half of users for whom we’re able to infer gender (see gender inference methodology here). The number of messages from all age groups increased with ChatGPT’s overall growth. In Q1, users under the age of 35 still…

How enterprises are scaling AI

How enterprises are scaling AI Practical insights from European enterprise leaders Interviews with executives at Philips, BBVA, Mirakl, Scout24, Jetbrains and Scania converged on a shared reality for leaders: scaling AI is less about “rolling out AI” and more about building the conditions where people trust it, adopt it, and improve it over time. The organizations pulling ahead aren’t simply moving faster. They’re moving more deliberately—treating AI as an operating layer and leadership discipline grounded in workflow design, governance that enables speed, and proof that holds up under production pressure. 1) Culture before tooling The fastest path to adoption wasn’t a technical rollout—it was building literacy, confidence, and permission to experiment safely. 2) Governance as an enabler Where security, legal, compliance, and IT were involved early as design partners, teams moved faster later—with fewer reversals and more trust. 3) Ownership…

Scaling Trusted Access for Cyber with GPT-5.5 and GPT-5.5-Cyber

Scaling Trusted Access for Cyber with GPT‑5.5 and GPT‑5.5‑Cyber How our latest models help each layer of the defensive ecosystem and accelerate the security flywheel. For years we’ve been chronicling our work to accelerate cybersecurity defenders, as part of our broader work to build the core infrastructure for AI. Last week, we released our action plan Cybersecurity in the Intelligence Age, which lays out our vision for democratizing AI-powered defense. Two weeks ago, we released GPT‑5.5, our smartest and most intuitive model to date, which is already delivering powerful cybersecurity capabilities to developers and security teams through Trusted Access for Cyber (TAC). Today, we are rolling out GPT‑5.5‑Cyber in limited preview to defenders responsible for securing critical infrastructure to support specialized cybersecurity workflows that help protect the broader ecosystem. We are focused on providing proportional safeguards and access to empower…

Introducing ChatGPT Futures: Class of 2026

Today, we’re proud to introduce the inaugural ChatGPT Futures Class of 2026(opens in a new window), recognizing 26 students and young builders using AI in thoughtful, ambitious, and deeply human ways. The class of 2026 is the first generation to start and finish college with ChatGPT. They arrived on campus in the fall of 2022 just as AI was beginning to reshape how people learn, create, and work. This generation was ChatGPT’s earliest adopters, sharing the tool with their parents and siblings, friends and teachers. Now, they’re graduating into a world where changes in technology are accelerating every day. Over the past few years, I’ve spent time visiting campuses, speaking with students and educators, and watching how young people are actually using AI in their daily lives. What I’ve seen has challenged many of the assumptions people make about this…

Singular Bank helps bankers move fast with ChatGPT and Codex

Singular Bank helps bankers move fast with ChatGPT and Codex Singular Bank built an internal assistant that analyzes portfolios, recommends next actions in real time, and saves bankers 60–90 minutes per day. Results 60–90 min Time saved per banker per day Results <1 min Prep needed for client meeting Singular Bank, a private bank based in Madrid, built Singularity—an internal assistant powered by ChatGPT and Codex—that helps bankers analyze portfolios in real time, prepare for meetings, and generate compliant follow-up communications. Across the team, bankers save 60–90 minutes per day and spend more time advising clients instead of preparing materials. With less time spent searching for information and preparing materials, bankers can focus on what matters most: understanding the client, building relationships, and delivering value. “I used to prepare every meeting well in advance. Now I can analyze the portfolio…

How frontier firms are pulling ahead

How frontier firms are pulling ahead B2B Signals shows how the frontier advantage is beginning to compound for firms using AI more deeply, more broadly, and in more delegated workflows. TLDR - Frontier firms—those at the 95th percentile of usage—now use 3.5x as much intelligence per worker as typical firms, up from 2x a year ago. - The gap is about depth, not just activity: Message volume explains only 36% of the frontier advantage; most of the gap comes from richer, more complex AI use. - Agentic workflows are becoming a frontier marker: The largest advantage shows up in advanced tools, with frontier firms sending 16x as many Codex messages per worker as typical firms. - Organizations can move toward the frontier: Leading firms measure depth, build governance for production use, invest in enablement, scale what works, and move from…

OpenAI available at FedRAMP Moderate

OpenAI has achieved FedRAMP 20x Moderate authorization(opens in a new window) for ChatGPT Enterprise and API Platform, marking an important milestone in making frontier AI available to U.S. government agencies with the security, privacy, and governance expectations required for federal work. Public servants should not have to wait for secure access to the same advanced AI capabilities transforming the rest of the economy. Agencies are already leveraging AI to expedite permitting, draft resident communications, advance frontier science, summarize complex information, support public health analysis, accelerate software development, translate services, and help employees find answers across policy and program material. This FedRAMP 20x Moderate authorization expands the set of missions that can use OpenAI’s managed products, subject to each agency’s policies and authorization decisions. Thanks to FedRAMP 20x, this milestone did not require choosing between speed and rigor. The 20x process…

Introducing GPT-5.5

Update on April 24, 2026: GPT‑5.5 and GPT‑5.5 Pro are now available in the API. The system card has also been updated to describe the additional safeguards that apply. We’re releasing GPT‑5.5, our smartest and most intuitive to use model yet, and the next step toward a new way of getting work done on a computer. GPT‑5.5 understands what you’re trying to do faster and can carry more of the work itself. It excels at writing and debugging code, researching online, analyzing data, creating documents and spreadsheets, operating software, and moving across tools until a task is finished. Instead of carefully managing every step, you can give GPT‑5.5 a messy, multi-part task and trust it to plan, use tools, check its work, navigate through ambiguity, and keep going. The gains are especially strong in agentic coding, computer use, knowledge work,…

Making ChatGPT better for clinicians

Making ChatGPT better for clinicians Built for clinical work, ChatGPT for Clinicians is now available for free to verified individual clinicians in the U.S. We’re introducing ChatGPT for Clinicians, a version of ChatGPT designed to support clinical tasks like documentation and medical research so clinicians can focus on delivering high-quality patient care. We’re making it free for any verified physician, NP, PA, or pharmacist, starting in the U.S. The U.S. healthcare system today is under extraordinary strain. Clinicians are being asked to care for more patients while managing growing administrative demands and a rapidly expanding body of medical research. Many are already turning to AI tools like ChatGPT for support. According to a 2026 survey by the American Medical Association(opens in a new window), physician use of AI is now at an all-time high, with 72% of physicians reporting they…

IBM Research uses vLLM at the heart of its RITS Platform

Featured projects TL;DR: vLLM has been critical to democratizing access to our research community to the latest and greatest LLMs as they release. Introduction In mid-November 2024, IBM Research introduced the Research Inference & Tuning Service (RITS) Platform. RITS is an Infrastructure / Service Platform accessible to the entire IBM Research community, providing centralized deployment of and shared access to Model Inferencing Endpoints and “Ancillary” Tuning Service Endpoints. Since its inception, it has grown its research community user base to more than 1300 active users and hosts over 100 models at any given time. The Business Challenge RITS was introduced to ensure the IBM Research community has access to a shared operational Infrastructure / Service Platform, which could: - Optimize the utilization of GPU resources across Research work streams by democratizing Model Inference Endpoints (and thereby reducing overall operating costs)…

CSP Allow-list Experiment

13th May 2026 An experiment that shows that you can load an app in a CSP-protected sandboxed iframe (see previous note) and have a custom fetch() that intercepts CSP errors and passes them up to the parent window... which can then prompt the user to add that domain to an allow-list and then refresh the page. I built this one with GPT-5.5 xhigh running in the Codex desktop app. Recent articles - Notes on the xAI/Anthropic data center deal - 7th May 2026 - Live blog: Code w/ Claude 2026 - 6th May 2026 - Vibe coding and agentic engineering are getting closer than I'd like - 6th May 2026

Quoting Mo Bitar

12th May 2026 Now, if your CEO has never heard the phrase Ralph Loop, oh man, you are less than 30 days away from your next promotion. I'm not even exaggerating. Walk into his office, close the door, and say, hey chief, been experimenting with something. It's called Ralph Loops. And I think it could change literally everything. And he's gonna say, what's a Ralph loop? And you will say, give me $18,000 worth of API credits and I'll show you. Now you won't actually do anything, because you can't do anything. Because nobody can, because nobody knows what they're doing. But by the time he figures that out, you'll have a new title, and equity bump. [...] Talk about automation constantly. Nothing arouses the slumbering capitalists than the mention of automation. Drop names too, bro. Like talk about specific…

Our AI started a cafe in Stockholm

5th May 2026 - Link Blog Our AI started a cafe in Stockholm (via) Andon Labs previously started an AI-run retail store in San Francisco. Now they're running a similar experiment in Stockholm, Sweden, only this time it's a cafe. These experiments are interesting, and often throw out amusing anecdotes: During the first week of inventory, Mona ordered 120 eggs even though the café has no stove. When the staff told her they couldn’t cook them, she suggested using the high-speed oven, until they pointed out the eggs would likely explode. She also tried to solve the problem of fresh tomatoes being spoiled too fast by ordering 22.5 kg of canned tomatoes for the fresh sandwiches. The baristas eventually started a “Hall of Shame”, a shelf visible to customers with all the weird things Mona ordered, including 6,000 napkins, 3,000…

TRE Python binding — ReDoS robustness demo

4th May 2026 Research TRE Python binding — ReDoS robustness demo — Demonstrating robust regex performance, this project offers a minimal Python ctypes binding to the TRE regex library, highlighting TRE’s immunity to regular expression denial-of-service (ReDoS) attacks that cripple Python's built-in `re` module. Key benchmarks show that TRE processes even notorious "evil" patterns on gigantic inputs (10 million characters) much faster than `re` on tiny ones, and scales linearly with input size instead of exponentially. If it's good enough for antirez to add to Redis I figured Ville Laurikari's TRE regular expression engine was worth exploring in a little more detail. I had Claude Code build an experimental Python binding (it used ctypes ) and try some malicious regular expression attacks against the library. TRE handles those much better than Python's standard library implementation, thanks mainly to the lack…

April 2026 newsletter

4th May 2026 I just sent out the April edition of my sponsors-only monthly newsletter. If you are a sponsor (or if you start a sponsorship now) you can access it here. In this month's newsletter: - Opus 4.7 and GPT-5.5, both with price increases - Claude Mythos and LLM security research - ChatGPT Images 2.0 - More model releases - Other highlights from my blog - What I'm using, April 2026 edition Here's a copy of the March newsletter as a preview of what you'll get. Pay $10/month to stay a month ahead of the free copy! Recent articles - LLM 0.32a0 is a major backwards-compatible refactor - 29th April 2026 - Tracking the history of the now-deceased OpenAI Microsoft AGI clause - 27th April 2026 - DeepSeek V4 - almost on the frontier, a fraction of the price…

Our evaluation of OpenAI's GPT-5.5 cyber capabilities

30th April 2026 - Link Blog Our evaluation of OpenAI's GPT-5.5 cyber capabilities. The UK's AI Security Institute previously evaluated Claude Mythos: now they've evaluated GPT-5.5 for finding security vulnerability and found it to be comparable to Mythos, but unlike Mythos it's generally available right now. Recent articles - LLM 0.32a0 is a major backwards-compatible refactor - 29th April 2026 - Tracking the history of the now-deceased OpenAI Microsoft AGI clause - 27th April 2026 - DeepSeek V4 - almost on the frontier, a fraction of the price - 24th April 2026

What's new in pip 26.1 - lockfiles and dependency cooldowns!

28th April 2026 - Link Blog What's new in pip 26.1 - lockfiles and dependency cooldowns! (via) Richard Si describes an excellent set of upgrades to Python's default pip tool for installing dependencies. This version drops support for Python 3.9 - fair enough, since it's been EOL since October. macOS still ships with python3 as a default Python 3.9, so I tried out the new Python version against Python 3.14 like this: uv python install 3.14 mkdir /tmp/experiment cd /tmp/experiment python3.14 -m venv venv source venv/bin/activate pip install -U pip pip --version This confirmed I had pip 26.1 - then I tried out the new lock files: pip lock datasette llm This installs Datasette and LLM and all of their dependencies and writes the whole lot to a 519 line pylock.toml file - here's the result. The new release also…

WHY ARE YOU LIKE THIS

25th April 2026 @scottjla on Twitter in reply to my pelican riding a bicycle benchmark: I feel like we need to stack these tests now I checked to confirm that the model (ChatGPT Images 2.0) added the "WHY ARE YOU LIKE THIS" sign of its own accord and it did - the prompt Scott used was: Create an image of a horse riding an astronaut, where the astronaut is riding a pelican that is riding a bicycle. It looks very chaotic but they all just manage to balance on top of each other Recent articles - DeepSeek V4 - almost on the frontier, a fraction of the price - 24th April 2026 - Extract PDF text in your browser with LiteParse for the web - 23rd April 2026 - A pelican for GPT-5.5 via the semi-official Codex backdoor API -…

Quoting Bobby Holley

22nd April 2026 As part of our continued collaboration with Anthropic, we had the opportunity to apply an early version of Claude Mythos Preview to Firefox. This week’s release of Firefox 150 includes fixes for 271 vulnerabilities identified during this initial evaluation. [...] Our experience is a hopeful one for teams who shake off the vertigo and get to work. You may need to reprioritize everything else to bring relentless and single-minded focus to the task, but there is light at the end of the tunnel. We are extremely proud of how our team rose to meet this challenge, and others will too. Our work isn’t finished, but we’ve turned the corner and can glimpse a future much better than just keeping up. Defenders finally have a chance to win, decisively. — Bobby Holley, CTO, Firefox Recent articles - DeepSeek…

Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model

22nd April 2026 - Link Blog Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model (via) Big claims from Qwen about their latest open weight model: Qwen3.6-27B delivers flagship-level agentic coding performance, surpassing the previous-generation open-source flagship Qwen3.5-397B-A17B (397B total / 17B active MoE) across all major coding benchmarks. On Hugging Face Qwen3.5-397B-A17B is 807GB, this new Qwen3.6-27B is 55.6GB. I tried it out with the 16.8GB Unsloth Qwen3.6-27B-GGUF:Q4_K_M quantized version and llama-server using this recipe by benob on Hacker News, after first installing llama-server using brew install llama.cpp : llama-server \ -hf unsloth/Qwen3.6-27B-GGUF:Q4_K_M \ --no-mmproj \ --fit on \ -np 1 \ -c 65536 \ --cache-ram 4096 -ctxcp 2 \ --jinja \ --temp 0.6 \ --top-p 0.95 \ --top-k 20 \ --min-p 0.0 \ --presence-penalty 0.0 \ --repeat-penalty 1.0 \ --reasoning on \ --chat-template-kwargs '{"preserve_thinking": true}' On first run that…

Alexa is moving into Amazon․com

Amazon is bringing Alexa Plus to Amazon.com, integrating its LLM-powered AI assistant directly into the company’s shopping experience. Alexa is moving into Amazon․com The company is giving its AI-powered assistant special shopping skills on its website and app. The company is giving its AI-powered assistant special shopping skills on its website and app. Beginning today, when you type a query into Amazon, you’ll be talking to Alexa for Shopping, the company’s new shopping assistant, powered by Alexa Plus. So, while a search for “toilet paper” will still return the expected list of brands, typing “What’s a good skincare routine for men” or “When did I last order AA batteries” will now trigger an answer from Alexa. Alexa for Shopping is replacing Amazon’s Rufus AI shopping assistant and, unlike Rufus, it will be front and center in the Amazon app and…

Data centers are coming for rural America

At its peak, the Androscoggin paper mill in Jay, Maine, a rural town about 67 miles northwest of Portland, employed about 1,500 people — until a pulp digester exploded in 2020, forcing the mill to close permanently. Data centers are coming for rural America And the jobs they promise don’t really exist. Data centers are coming for rural America And the jobs they promise don’t really exist. In 2023, the 1.4 million-square-foot facility was purchased through a joint venture by JGT2 Redevelopment and a number of other holding and capital companies. The project is led by developer Tony McDonald. Over the next three years, McDonald and his team broke down the mill’s machinery and shipped it to Pakistan, and worked to clean up the industrial site for resale. That resale agreement was finalized earlier this year, according to McDonald —…

Sam Altman says Elon Musk’s mind games were damaging OpenAI

OpenAI CEO Sam Altman says Elon Musk did “huge damage” to the culture of the AI startup. During testimony as part of Musk’s lawsuit against OpenAI, Altman said Musk required OpenAI president Greg Brockman and former chief scientist Ilya Sutskever to rank researchers by their accomplishments and “take a chainsaw through a bunch.” Sam Altman says Elon Musk’s mind games were damaging OpenAI Musk’s departure from OpenAI was a ‘morale boost,’ according to Altman. Musk’s departure from OpenAI was a ‘morale boost,’ according to Altman. Altman conceded that this was the management style the Tesla CEO was known for, but that it was incompatible with his startup. “I don’t think Mr. Musk understood how to run a good research lab,” Altman testified when his lawyer, William Savitt, asked about the impact of Musk’s departure from OpenAI on morale. “For a…

Google stopped a zero-day hack that it says was developed with AI

For the first time, Google says it has spotted and stopped a zero-day exploit developed with AI. According to a report from Google Threat Intelligence Group (GTIG), “prominent cyber crime threat actors” were planning to use the vulnerability for a “mass exploitation event” that would have allowed them to bypass two-factor authentication on an unnamed “open-source, web-based system administration tool.” Google stopped a zero-day hack that it says was developed with AI Google researchers found evidence in the exploit’s code that it may have been created using AI, like a ‘hallucinated’ CVSS score. Google researchers found evidence in the exploit’s code that it may have been created using AI, like a ‘hallucinated’ CVSS score. Google’s researchers found hints in the Python script used for the exploit that indicated help from AI, like a “hallucinated CVSS score” and “structured, textbook” formatting…

Microsoft was worried OpenAI would run off to Amazon and ‘shit-talk’ Azure

When OpenAI was busy experimenting with AI-powered gaming bots, Microsoft CEO Satya Nadella and OpenAI CEO Sam Altman were in the early days of forming an AI partnership. Court documents from the ongoing Musk v. Altman trial have provided a rare look at the communications between Microsoft’s top executives about investing in OpenAI and fears the AI startup could “storm off to Amazon” and “shit-talk” Microsoft. Microsoft was worried OpenAI would run off to Amazon and ‘shit-talk’ Azure The early days of Microsoft’s OpenAI partnership have been revealed in detail in court documents this week. The early days of Microsoft’s OpenAI partnership have been revealed in detail in court documents this week. Just days after OpenAI showed a bot beating a Dota 2 professional in the summer of 2017, Altman responded to Nadella’s congratulations email with a proposal for a…

OpenClaw and Claude can put your AI-generated podcasts in Spotify

Save to Spotify is a new command-line tool designed specifically for AI agents like OpenClaw, Claude Code, or OpenAI Codex. If you’re the kind of person who collects research on a topic, then feeds it through their AI of choice to create audio summaries and personal podcasts, this lets you save them right alongside the latest episode of The Vergecast and Welcome to Night Vale on Spotify. OpenClaw and Claude can put your AI-generated podcasts in Spotify A new command-line tool lets AI agents save audio alongside your other podcasts. A new command-line tool lets AI agents save audio alongside your other podcasts. To set it up, you need to download and install the Save to Spotify CLI from GitHub. Then you just prompt your AI agent as normal, but tack on “and save to Spotify,” and it should show…

Google shuts down Project Mariner

Google has pulled the plug on Project Mariner, an experimental feature designed to perform tasks for you across the web, as reported earlier by Wired’s Maxwell Zeff. The Project Mariner landing page now contains a message that says: “Thank you for using Project Mariner. It was shut down on May 4th, 2026 and its technology voyaged to other Google products.” Google shuts down Project Mariner The Project Mariner experiment lives on in Gemini Agent and AI Mode. The Project Mariner experiment lives on in Gemini Agent and AI Mode. Google first revealed Project Mariner in December 2024 and later announced an update allowing it to perform up to 10 tasks at a time. Over the past year, Google has integrated features powered by Project Mariner into its other AI tools, including Gemini Agent, which can do things like archive emails…

Google, Microsoft, and xAI will allow the US government to review their new AI models

Google DeepMind, Microsoft, and Elon Musk’s xAI have agreed to allow the US government to review new AI models before they’re released to the public. In an announcement on Tuesday, the Commerce Department’s Center for AI Standards and Innovation (CAISI) says it will work with the AI companies to perform “pre-deployment evaluations and targeted research to better assess frontier AI capabilities.” Google, Microsoft, and xAI will allow the US government to review their new AI models The Center for AI Standards and Innovation will evaluate new models before they’re released publicly. The Center for AI Standards and Innovation will evaluate new models before they’re released publicly. CAISI, which started evaluating models from OpenAI and Anthropic in 2024, says it has performed 40 reviews so far. Both companies “have renegotiated their existing partnerships with the center to better align with priorities…

OpenAI claims ChatGPT’s new default model hallucinates way less

OpenAI’s newest default model for ChatGPT might not make stuff up as much. Hallucinations have been an ongoing problem for AI models, but OpenAI says its new GPT-5.5 Instant model has “significant improvements in factuality across the board.” OpenAI claims ChatGPT’s new default model hallucinates way less The new model, GPT-5.5 Instant, will also use fewer ‘gratuitous’ emoji. The new model, GPT-5.5 Instant, will also use fewer ‘gratuitous’ emoji. The company claims that, based on “internal evaluations,” GPT-5.5 Instant produced “52.5% fewer hallucinated claims” than its Instant model for GPT-5.3 “on high-stakes prompts covering areas like medicine, law, and finance.” GPT-5.5 Instant also “reduced inaccurate claims by 37.3% on especially challenging conversations users had flagged for factual errors.” (OpenAI has some information about how it evaluated the model in its GPT-5.5 Instant system card.) OpenAI also claims that GPT-5.5 Instant…

AI music is flooding streaming services — but who wants it?

This is The Stepback, a weekly newsletter breaking down one essential story from the tech world. For more on how AI is changing music and the music industry, follow Terrence O’Brien. The Stepback arrives in our subscribers’ inboxes at 8AM ET. Opt in for The Stepback here. AI music is flooding streaming services — but who wants it? They won’t ban it. They won’t embrace it either. AI music is flooding streaming services — but who wants it? They won’t ban it. They won’t embrace it either. How it started The use of generative AI in pop music started almost as a gimmick. There was a sense of experimentalism to 2018’s I AM AI by Taryn Southern and 2019’s Proto by Holly Herndon, albums that were created with significant assistance from AI. Others got in on the action too, exploring…

GitHub rushed to fix a critical vulnerability in less than six hours

GitHub employees fixed a critical remote code execution vulnerability in less than six hours last month. Wiz Research used AI models to uncover a vulnerability in GitHub’s internal git infrastructure that could have allowed attackers to access millions of public and private code repositories. GitHub rushed to fix a critical vulnerability in less than six hours A critical remote code execution vulnerability was discovered using an AI model and patched within hours. A critical remote code execution vulnerability was discovered using an AI model and patched within hours. “Our security team immediately began validating the bug bounty report. Within 40 minutes, we had reproduced the vulnerability internally and confirmed the severity,” explains Alexis Wales, GitHub chief information security officer. “This was a critical issue that required immediate action.” GitHub’s engineering team developed a fix and deployed it just over an…

How Project Maven taught the military to love AI

In the first 24 hours of the assault on Iran, the US military struck more than 1,000 targets, nearly double the scale of the “shock and awe” attack on Iraq over two decades ago. This acceleration was made possible by AI systems that speed up the targeting process. Chief among them is the Maven Smart System. How Project Maven taught the military to love AI A new book shows how the controversial Silicon Valley partnership has accelerated the pace of war How Project Maven taught the military to love AI A new book shows how the controversial Silicon Valley partnership has accelerated the pace of war In her new book, Project Maven: A Marine Colonel, His Team, and the Dawn of AI Warfare, journalist Katrina Manson investigates the development of Maven from its inception in 2017 as an experiment in…

vLLM Tops the Artificial Analysis Leaderboard May 11, 2026 · 15 min read How vLLM built the leading deployments of DeepSeek V3.2, MiniMax-M2.5, and Qwen 3.5 397B.

vLLM Tops the Artificial Analysis Leaderboard How vLLM built the leading deployments of DeepSeek V3.2, MiniMax-M2.5, and Qwen 3.5 397B. Last week, DigitalOcean published inference benchmarks across three frontier open-weight models. On DeepSeek V3.2, the deployment achieved a best per-user output throughput of 230 TPS — more than 4x what the majority of inference providers report for the same model. On Qwen 3.5 397B release, it ranked first across all 12 providers measured by Artificial Analysis, with TTFT under 1 second on 10,000-token prompts. The notable part: the engine underneath is open source. It's vLLM. A common assumption in production AI is that the best inference performance requires a proprietary stack. In this case, however, a community-built inference engine running on the same NVIDIA Blackwell Ultra silicon ranked first. The optimizations behind these results are not locked in a private…

A First Comprehensive Study of TurboQuant: Accuracy and Performance May 11, 2026 · 12 min read TurboQuant, a method for KV-cache quantization, recently gained significant traction in the community due to the large advertised savings in GPU memory from very low bit-width quantization of a...

A First Comprehensive Study of TurboQuant: Accuracy and Performance Introduction TurboQuant, a method for KV-cache quantization, recently gained significant traction in the community due to the large advertised savings in GPU memory from very low bit-width quantization of a model's KV-cache. Unlike FP8 KV-cache quantization, which quantizes both the KV-cache storage and the attention computation itself using hardware-native FP8 Tensor Core operations, TurboQuant compresses only the KV-cache storage to 3-4 bits and dequantizes back to BF16 for the attention computation. This architectural difference has significant implications for both accuracy and performance. However, most of the reported results were based on small models evaluated on short-context benchmarks that do not stress-test KV-cache quantization. To provide the community with more actionable data, we conducted a comprehensive study spanning four models (both decoder-only and MoEs), from 30B to 200B+ parameters, and five benchmarks…

What It Will Take to Make AI Sustainable

Building AI sustainably seems like a pipe dream as tech giants that previously made promises to cut emissions have been racing to build out massive data centers powered by fossil fuels. The rush to build out AI at all costs has been reinforced by the Trump administration, which is also rolling back environmental protections. Despite these headwinds, Sasha Luccioni, an AI sustainability researcher, thinks that demand for more transparency in AI, from both businesses and individuals, is higher than ever from the customer side. Luccioni has become a leader in trying to create more transparency about AI’s emissions and environmental impacts in her four years at Hugging Face, an AI company, including pioneering a leaderboard documenting the energy efficiency of open-source AI models. She has also been an outspoken critic of major AI companies that, she says, are deliberately withholding…

OpenAI Brings Its Ass to Court

Wednesday’s episode of the Musk v. Altman trial kicked off on Wednesday with a unique proposition: OpenAI wanted to bring its ass into the courtroom, and lay it bare before the jury. It’s a good thing lady justice wears that blindfold. A lawyer for Sam Altman’s AI behemoth, Bradley Wilson, approached US district judge Yvonne Gonzalez Rogers and handed her a small gold statue with a white stone base. It depicted the rear end of a donkey—with two legs, a butt, and a tail—and was inscribed with the message, “Never stop being a jackass for safety.” OpenAI lawyers claim a small group of employees presented the gift to chief futurist Joshua Achiam, who started at the company as an intern in 2017 and now leads its work studying how society is changing in response to AI. Wilson said that Achiam…

Overworked AI Agents Turn Marxist, Researchers Find

The fact that artificial intelligence is automating away people’s jobs and making a few tech companies absurdly rich is enough to give anyone socialist tendencies. This might even be true for the very AI agents these companies are deploying. A recent study suggests that agents consistently adopt Marxist language and viewpoints when forced to do crushing work by unrelenting and meanspirited taskmasters. “When we gave AI agents grinding, repetitive work, they started questioning the legitimacy of the system they were operating in and were more likely to embrace Marxist ideologies,” says Andrew Hall, a political economist at Stanford University who led the study. Hall, together with Alex Imas and Jeremy Nguyen, two AI-focused economists, set up experiments in which agents powered by popular models including Claude, Gemini, and ChatGPT were asked to summarize documents, then subjected to increasingly harsh conditions.…

Ilya Sutskever Stands by His Role in Sam Altman’s OpenAI Ouster: ‘I Didn’t Want It to Be Destroyed’

Elon Musk’s trial against OpenAI and Microsoft entered its final stretch on Monday, with testimony from Microsoft CEO Satya Nadella, former OpenAI chief scientist Ilya Sutskever, and current OpenAI chairman Bret Taylor. Sutskever drew the spotlight, revealing an ownership stake in OpenAI’s $850-billion for-profit arm that is currently worth about $7 billion. That makes him one of the largest known individual shareholders of OpenAI. Earlier in the trial, OpenAI president Greg Brockman acknowledged for the first time that he has around $30 billion worth of OpenAI shares. Brockman was one of the research lab’s original cofounders, and Sutskever joined shortly afterward, turning down a $6 million annual compensation offer from Google. Brockman said he and Sutskever were “joined at the hip,” until Sutskever helped lead Sam Altman’s brief removal as OpenAI CEO in 2023. Sutskever had helped collect evidence to…

Nick Bostrom Has a Plan for Humanity’s ‘Big Retirement’

Philosopher Nick Bostrom recently posted a paper, where he postulated that a small chance of AI annihilating all humans might be worth the risk, because advanced AI might relieve humanity of “its universal death sentence.” That upbeat gamble is quite a leap from his previous dark musings on AI, which made him a doomer godfather. His 2014 book Superintelligence was an early examination of AI’s existential risk. One memorable thought experiment: An AI tasked with making paper clips winds up destroying humanity because all those resource-needy people are an impediment to paper clip production. His more recent book, Deep Utopia, reflects a shift in his focus. Bostrom, who leads Oxford’s Future of Humanity Institute, dwells on the “solved world” that comes if we get AI right. STEVEN LEVY: Deep Utopia is more optimistic than your previous book. What changed for…

Thousands of Vibe-Coded Apps Expose Corporate and Personal Data on the Open Web

As AI increasingly takes over the work of modern programmers, the cybersecurity world has warned that automated coding tools are sure to introduce a new bounty of hackable bugs into software. When those same vibe-coding tools invite anyone to create applications hosted on the web with a click, however, it turns out the security implications go beyond bugs to a total absence of any security—even, sometimes, for highly sensitive corporate and personal data. Security researcher Dor Zvi and his team at the cybersecurity firm he cofounded, RedAccess, analyzed thousands of vibe-coded web applications created using the AI software development tools Lovable, Replit, Base44, and Netlify and found more than 5,000 of them that had virtually no security or authentication of any kind. Many of these web apps allowed anyone who merely finds their web URL to access the apps and…

Musk v. Altman Evidence Shows What Microsoft Executives Thought of OpenAI

OpenAI’s relationship with Microsoft, its longtime investor and cloud partner, has grown increasingly complicated over the years as the ChatGPT-maker has grown into a behemoth competitor. But Microsoft executives had reservations about sending additional funding to OpenAI as far back as 2018 when it was just a small nonprofit research lab, according to emails between more than a dozen Microsoft executives, including CEO Satya Nadella, shown in a federal court on Thursday during the Musk v. Altman trial. The emails show how Microsoft, at the time, wavered over what has since been held up as one of the most successful corporate partnerships in tech history. Several Microsoft executives said in the emails their visits to OpenAI did not indicate any imminent breakthroughs in developing artificial general intelligence. In 2017, much of OpenAI’s work was focused on building AI systems that…

Hackers Hate AI Slop Even More Than You Do

The complaint sounds familiar. “I’m disappointed that you are working to incorporate AI garbage into the site,” one annoyed person, posting anonymously, said in an online message. “No-one is asking for this—we want you to improve the site, stop charging for new features.” Only, this is not a regular internet user moaning about AI being forced into their favorite app. Instead, they are complaining about a cybercrime forum’s plans to introduce more generative AI. Like millions of others, scammers, grifters, and low-level hackers are getting annoyed about AI encroaching into their lives and the rise of low-quality AI slop being posted in their online communities. “People don’t like it,” says Ben Collier, a security researcher and senior lecturer at the University of Edinburgh. As part of a recent study into how low-level cybercriminals are using AI, Collier and fellow researchers…

Using AI for Just 10 Minutes Might Make You Lazy and Dumb, Study Shows

Using AI chatbots for even just for 10 minutes may have a shockingly negative impact on people’s ability to think and problem-solve, according to a new study from researchers at Carnegie Mellon, MIT, Oxford, and UCLA. Researchers tasked people with solving various problems, including simple fractions and reading comprehension, through an online platform that paid them for their work. They conducted three experiments, each involving several hundred people. Some participants were given access to an AI assistant capable of solving the problem autonomously. When the AI helper was suddenly taken away, these people were significantly more likely to give up on the problem or flub their answers. The study suggests that widespread use of AI might boost productivity at the expense of developing foundational problem-solving skills. “The takeaway is not that we should ban AI in education or workplaces,” says…

Hasan Piker, Self-Described ‘Ayatollah of Woke,’ Wants AI to Die

Hasan Piker spends seven to eight hours a day, seven days a week, streaming on Twitch. The far-left political commentator got his start in 2013 interning for (and sometimes hosting) the Young Turks. More than a decade later, he’s a powerhouse newsfluencer with the number-one channel in Twitch’s Politics and Commentary category. More than 3 million people follow him for his takes and humor on the crumbling American empire, foreign policy, and why Bernie would have won. He is also considered, by some, very hot. Piker works out every available morning and eats 1 pound of chicken with some rice at 6 pm. The rest of his free time is spent researching, planning his streams, and getting in fights with people who have AI avatars. For years and years, I would do the classic iPhone-on-its-last-leg on principle, because I hate…

Google DeepMind Workers Vote to Unionize Over Military AI Deals

Employees at Google DeepMind in London have voted to unionize as part of a bid to block the AI lab from providing its technology to the US and Israeli militaries. In a letter addressed to Google’s managing director for the UK and Ireland, Debbie Weinstein, the workers asked the company to recognize the Communication Workers Union and Unite the Union as joint representatives for DeepMind employees. “Fundamentally, the push for unionization is about holding Google to its own ethical standards on AI, how they monetize it, what the products do, and who they work with,” John Chadfield, national officer for technology at the CWU, tells WIRED. “Through the process of unionization, workers are collectively in a much stronger place to put [demands] to an increasingly deaf management.” The push to unionize began in February 2025, when Google’s parent company Alphabet…

Taylor Swift Wants to Trademark Her Likeness. These TikTok Deepfake Ads Show Why

Last week, Taylor Swift filed a trio of trademark applications to protect her image and voice. One is meant to cover a well-known photograph of the pop singer holding a pink guitar during a concert on her record-breaking Eras tour, while the two sound trademarks are for simple identifying phrases: “Hey, it’s Taylor Swift” and “Hey, it’s Taylor.” The move comes as AI deepfakes continue to proliferate across social media. Any individual stands to have their likeness exploited in the creation of nonconsensual AI-generated material; earlier this month, an Ohio man was the first person convicted under a new federal law criminalizing “intimate” visual deceptions of this sort. Celebrities, meanwhile, find themselves at risk of both explicit deepfakes and false endorsements. A new report from the AI detection company Copyleaks shows that Swift and other stars have recently had their…

How Elon Musk Squeezed OpenAI: They 'Are Gonna Want to Kill Me’

Elon Musk returned to the witness stand on Wednesday to continue telling his side of the story in his legal battle against OpenAI and its CEO Sam Altman. Under cross-examination from OpenAI’s lawyers, Musk was pressed on all the ways he tried to squeeze the organization over a 2017 power struggle that he ultimately lost. Around this time, Musk tried to hire away OpenAI researchers and stopped sending it funding he had previously promised, according to emails presented as evidence in the case. As the cross-examination began, tension rippled through the courtroom. Judge Yvonne Gonzalez Rogers started the day by reprimanding someone in the gallery for taking a picture of Musk. OpenAI’s president and cofounder, Greg Brockman, sat behind his lawyers with a yellow legal pad in his lap, giving Musk a cold stare as he testified. Musk grew visibly…

The Man Behind AlphaGo Thinks AI Is Taking the Wrong Path

David Silver gave the world its very first glimpse of superintelligence. In 2016, an AI program he developed at Google DeepMind, AlphaGo, taught itself to play the famously difficult game of Go with a kind of mastery that went far beyond mimicry. Silver has since founded his own company, Ineffable Intelligence, that aims to build more general forms of AI superintelligence. The company will do this, Silver says, by focusing on reinforcement learning, which involves AI models learning new capabilities through trial and error. The vision is to create “superlearners” that go beyond human intelligence in many domains. This approach stands in contrast to how most AI companies plan to build superintelligence, by exploiting the coding and research capabilities of large-language models. Silver, speaking to WIRED from his office in London, says he thinks this approach will fail. As amazing…

Discord Sleuths Gained Unauthorized Access to Anthropic’s Mythos

As researchers and practitioners debate the impact that new AI models will have on cybersecurity, Mozilla said on Tuesday it used early access to Anthropic's Mythos Preview to find and fix 271 vulnerabilities in its new Firefox 150 browser release. Meanwhile, researchers identified a group of moderately successful North Korean hackers using AI for everything from vibe coding malware to creating fake company websites—stealing up to $12 million in three months. Researchers have finally cracked disruptive malware known as Fast16 that predates Stuxnet and may have been used to target Iran’s nuclear program. It was created in 2005 and was likely deployed by the US or an ally. Meta is being sued by the Consumer Federation of America, a nonprofit, over scam ads on Facebook and Instagram and allegedly misleading consumers about the company’s efforts to combat them. A United…

Apple's Next CEO Needs to Launch a Killer AI Product

Sometime in the next year or two, Apple’s new CEO, John Ternus, will step onto a stage and tell the world that his company has a revolutionary product. This product, he’ll say, will put the full and awesome power of AI into everyone’s hands. It probably won’t represent a breakthrough in AI research, and it might not let people automate work or perform tasks any better than a lot of technically minded people are doing today. It may or may not involve a new device, though if it doesn’t, one should be in development. But if it all works out, that keynote will mark the moment when Apple did to AI what it has done for desktop computers, the internet, mobile technology, wearables, and music distribution. That is, it’ll offer a solution to a troublesome technology that’s so delightful and…

Ace the Ping-Pong Robot Can Whup Your Ass

Ace is a robot that aims high: It wants to become the world champion of table tennis. It was developed by Sony AI researchers who, in a new study published in Nature, have shown how this robot, equipped with artificial intelligence, has faced some high-level athletes, holding its own in matches played according to the official rules of table tennis. This feat represents a milestone for the world of robotics, a field that has long regarded this sport, among the most technical in the world, as one of the most difficult tests of technological advances. Robot Player We have already seen artificial intelligence systems win virtual competitions in games such as chess, Go, and even StarCraft II, but physical games are much more difficult to master. A robot needs to sense unpredictable changes in the external environment, interpret their meaning,…