Amazon has shut down KiroRank, the internal leaderboard that ranked employees by how many AI tokens they consumed through the company’s Kiro developer platform. The decision was announced internally by Amazon’s senior vice president of eCommerce services, who told staff: “Please don’t use AI just for the sake of using AI. Use it to solve customer problems, to help you solve business problems, to innovate.”
Sam Altman said that the rapid development of AI will not lead to a global "jobs apocalypse," and that the technology has not eliminated as many white-collar jobs as he had expected by this point. "I'm delighted to be wrong about this," Altman told CBA Chief Executive Matt Comyn. "I thought there would have been more impact on entry-level white-collar jobs being eliminated by now than has actually happened."
Anthropic says Opus 4.8 is "around four times less likely than its predecessor to allow flaws in code it has written to pass unremarked", meaning the model is better at flagging its own uncertainty rather than confidently shipping broken work. The alignment assessment, included in the system card, says the model shows "substantially lower" rates of misaligned behavior than Opus 4.7.
Uber's president and chief operating officer Andrew Macdonald said in an interview released over the weekend that the company can't draw a clear line between its rising AI spending and the features it's actually shipping to customers. The comments are the most direct public admission yet from a senior executive at a major enterprise AI customer that the math isn't working.
Pope Leo XIV released his first encyclical today, Magnifica Humanitas (“Magnificent Humanity”), focused on “the protection of the human person in the age of artificial intelligence.”
Sam Altman told the current Y Combinator class that OpenAI will invest $2 million in tokens in every startup in the cohort. The startups give up equity in return. Y Combinator partner Tyler Bosmeny called it a “mic drop moment.”
ClickUp, the $4 billion productivity software company, laid off 22 percent of its workforce. Founder and CEO Zeb Evans announced the cut on X, framing it not as cost-cutting but as a structural bet on AI. The company is also introducing salary bands that reach $1 million per year in cash for employees Evans calls "100x contributors".
Steve Wozniak, the Apple co-founder, became one of the few commencement speakers this year to get applause rather than boos for talking about AI. Speaking at Grand Valley State University earlier in May, Wozniak told graduates: "You have AI — actual intelligence." The line drew laughs and cheers.
Microsoft began canceling internal Claude Code licenses, giving thousands of developers in its Windows, Microsoft 365, Teams, and Surface organizations until June 30 to switch to GitHub Copilot CLI.
Peter Steinberger, the founder of the open-source coding agent OpenClaw and now an engineer at OpenAI, published a screenshot of his team's 30-day OpenAI API dashboard, showing $1,305,088.81 in usage charges across 603 billion tokens and 7.6 million requests.
Security research firm Calif disclosed the first publicly documented macOS kernel memory corruption exploit on Apple's M5 silicon , the first to bypass Memory Integrity Enforcement, the hardware-assisted memory safety system Apple introduced as the marquee security feature of its M5 and A19 chips.
OpenAI announced that an internal general-purpose reasoning model had autonomously disproved a central conjecture in discrete geometry, producing the first instance of an AI system independently solving a prominent open problem at the frontier of an active mathematical field.
OpenAI co-founder Andrej Karpathy has joined Anthropic's pre-training team, where he will start a new team focused on using Claude to accelerate pre-training research itself.
Dr. Axel Rauschmayer, author of the long-running JavaScript reference blog 2ality.com and a series of widely cited language books, took both his blog and his free online books offline.
Vercel Labs released an experimental low-level systems programming language in the same design space as C, Rust, and Zig, built from the ground up so AI coding agents are the intended consumers of compiler output, diagnostics, and repair guidance.
Amazon shuts down the internal leaderboard at the center of the tokenmaxxing problem
Amazon has shut down KiroRank, the internal leaderboard that ranked employees by how many AI tokens they consumed through the company’s Kiro developer platform. The decision was announced internally by Amazon’s senior vice president of eCommerce services, who told staff: “Please don’t use AI just for the sake of using AI. Use it to solve customer problems, to help you solve business problems, to innovate.”
Altman now says AI hasn't taken the jobs he thought it would
Sam Altman said that the rapid development of AI will not lead to a global "jobs apocalypse," and that the technology has not eliminated as many white-collar jobs as he had expected by this point. "I'm delighted to be wrong about this," Altman told CBA Chief Executive Matt Comyn. "I thought there would have been more impact on entry-level white-collar jobs being eliminated by now than has actually happened."
Anthropic launches Opus 4.8, betting on bigger workflows at the moment customers are pushing back on cost
Anthropic says Opus 4.8 is "around four times less likely than its predecessor to allow flaws in code it has written to pass unremarked", meaning the model is better at flagging its own uncertainty rather than confidently shipping broken work. The alignment assessment, included in the system card, says the model shows "substantially lower" rates of misaligned behavior than Opus 4.7.
Uber's president says AI spending is getting "harder to justify"
Uber's president and chief operating officer Andrew Macdonald said in an interview released over the weekend that the company can't draw a clear line between its rising AI spending and the features it's actually shipping to customers. The comments are the most direct public admission yet from a senior executive at a major enterprise AI customer that the math isn't working.
The Pope’s first encyclical is about AI, and Anthropic is in the room
Pope Leo XIV released his first encyclical today, Magnifica Humanitas (“Magnificent Humanity”), focused on “the protection of the human person in the age of artificial intelligence.”
Sam Altman offers $2 million in tokens to every Y Combinator startup, in exchange for equity
Sam Altman told the current Y Combinator class that OpenAI will invest $2 million in tokens in every startup in the cohort. The startups give up equity in return. Y Combinator partner Tyler Bosmeny called it a “mic drop moment.”
ClickUp lays off 22% of staff, says the survivors can earn $1 million
ClickUp, the $4 billion productivity software company, laid off 22 percent of its workforce. Founder and CEO Zeb Evans announced the cut on X, framing it not as cost-cutting but as a structural bet on AI. The company is also introducing salary bands that reach $1 million per year in cash for employees Evans calls "100x contributors".
Wozniak found a way to talk about AI at a graduation without getting booed
Steve Wozniak, the Apple co-founder, became one of the few commencement speakers this year to get applause rather than boos for talking about AI. Speaking at Grand Valley State University earlier in May, Wozniak told graduates: "You have AI — actual intelligence." The line drew laughs and cheers.
Microsoft canceled its internal Claude Code licenses after token-based billing made the cost untenable
Microsoft began canceling internal Claude Code licenses, giving thousands of developers in its Windows, Microsoft 365, Teams, and Surface organizations until June 30 to switch to GitHub Copilot CLI.
OpenClaw founder spends $1.3 million in a month running 100 AI coding agents
Peter Steinberger, the founder of the open-source coding agent OpenClaw and now an engineer at OpenAI, published a screenshot of his team's 30-day OpenAI API dashboard, showing $1,305,088.81 in usage charges across 603 billion tokens and 7.6 million requests.
Security firm bypasses Apple's MIE protections on M5 in five days using Anthropic's Mythos assistance
Security research firm Calif disclosed the first publicly documented macOS kernel memory corruption exploit on Apple's M5 silicon , the first to bypass Memory Integrity Enforcement, the hardware-assisted memory safety system Apple introduced as the marquee security feature of its M5 and A19 chips.
OpenAI says internal model has disproved central Erdős conjecture, with proof verified by external mathematicians
OpenAI announced that an internal general-purpose reasoning model had autonomously disproved a central conjecture in discrete geometry, producing the first instance of an AI system independently solving a prominent open problem at the frontier of an active mathematical field.
Karpathy joins Anthropic to lead AI-accelerating-AI research
OpenAI co-founder Andrej Karpathy has joined Anthropic's pre-training team, where he will start a new team focused on using Claude to accelerate pre-training research itself.
One of JavaScript's most cited reference blogs, taken offline after AI crawlers overwhelm independent author
Dr. Axel Rauschmayer, author of the long-running JavaScript reference blog 2ality.com and a series of widely cited language books, took both his blog and his free online books offline.
Vercel releases Zero, a systems language whose compiler speaks JSON to AI agents
Vercel Labs released an experimental low-level systems programming language in the same design space as C, Rust, and Zig, built from the ground up so AI coding agents are the intended consumers of compiler output, diagnostics, and repair guidance.