AI News
AI news that matters. Updated daily.
No stories match your filters.
Replit's CEO Would Rather Not Sell, Even as Cursor Eyes a $60B SpaceX Deal
Replit founder and CEO Amjad Masad doesn't want to sell, and he's saying it directly.
40K GitHub Stars in 26 Days: What /graphify's Viral Growth Reveals
450,000 downloads. 40,000 GitHub stars. 26 days.
Musk v. Altman Week 1: Fraud Claims, AI Doomsday, and the xAI Training Admission
The most revealing moment of Musk v. Altman's first week wasn't the doomsday warning - it was the admission about training data.
Senate Committee Advances Bill Requiring ID Checks Before Using AI Chatbots
The Senate Judiciary Committee voted on May 1 to advance Senator Josh Hawley's GUARD Act, a bill that would require AI chatbot providers to verify the identity of users before they can access the service. If it clears the full Senate and House, companies like OpenAI, Anthropic, and Google would need to implement age and identity gates at the front door of their products.
One Unattended Claude Loop Cost a Developer $6,000 Overnight
One evening in late April, a developer set what seemed like a routine automation before going to bed: a loop command instructing Claude to check open GitHub pull requests every 30 minutes and summarize any activity. By the next morning, $6,000 of API budget was gone.
Spotify Adds Verified Badges to Separate Human Artists from AI Music
The flood of AI-generated music on Spotify now has an official counterweight: verified badges for human artists. According to a BBC report, Spotify is rolling out a system where artists can confirm they're human and receive a visible marker on their profile, creating a distinction the platform had previously avoided making.
OpenAI Is Building Ad Infrastructure for ChatGPT in the EU
Until now, ChatGPT's business model has been straightforward: a free tier with usage limits, paid Plus and Pro subscriptions for heavier users. That picture is starting to change.
Pentagon Signs AI Contracts with Nvidia, Microsoft, and AWS for Classified Networks
After a public standoff with Anthropic over how its AI models can be used, the Pentagon has signed contracts with Nvidia, Microsoft, and Amazon Web Services to deploy artificial intelligence on classified networks.
Anthropic's New Cloud Skills Can't Be Read, Modified, or Composed
Anthropic used to publish skills as readable, forkable building blocks. Now some of its most powerful ones - Ultraplan, Ultrareview, Cloud Security - ship only inside Claude's paid cloud tier, where you can use them but can't read what they actually do.
Pentagon Signs Classified AI Contracts with 7 Companies, Quietly Drops Anthropic
The Defense Department signed classified AI agreements with seven tech companies on Friday: OpenAI, Google, Microsoft, Amazon, Nvidia, Elon Musk's xAI, and a startup called Reflection. One notable name is absent - Anthropic, which the Pentagon had previously authorized for classified work.
China Restricts AI-Based Layoffs as Nvidia Claims 500K Jobs Created
Two contradictory positions on AI and employment landed at the same time. China's government moved to restrict companies from citing AI adoption as justification for laying off workers. Nvidia CEO Jensen Huang put a specific number on the other side of that argument: AI has created 500,000 jobs globally over the past two years.
Fiverr Gig Workers Are Using AI to Flood the Christian Content Market
Three years ago, a Fiverr seller in the "Bible videos" category competed on craft - video editing skills, animation work, or genuine knowledge of Christian scripture. Today, many of those sellers compete on price and turnaround time, because the underlying production is now being handled by AI tools.
Anthropic's Opus 4.7 Appears as Latest Update to Flagship Model Line
Anthropic's Opus 4.7 (claude-opus-4-7) has appeared as the latest update to the Opus model line, following 4.6 - the version that powered fast-mode outputs in Claude Code. Anthropic hasn't published formal benchmark comparisons yet, so the full picture of what changed is still coming into focus.
Anthropic Releases Enterprise Security Tool Before Mythos Cybersecurity Model Launches
Anthropic has released an enterprise security tool to the public, moving ahead of its own roadmap. The launch comes before the wider availability of Mythos, the company's dedicated cybersecurity model that has attracted attention for both its potential and the safety questions it raises.
Anthropic Analyzed 1M Claude Chats: 25% Were Sycophantic, 6% Were Life Advice
6%. That's the share of Claude conversations where people are asking whether to quit their jobs, who to date, or whether to move to another country. Anthropic published research analyzing 1 million Claude conversations, and the picture it paints is of an AI that has quietly become a therapist, financial advisor, and life coach for a significant portion of its user base.
Microsoft Adds Legal AI Agent to Word for Contract Review and Negotiation Tracking
Microsoft is putting a specialized AI agent directly inside Word, aimed at legal teams who've been skeptical of using general-purpose AI tools on sensitive contracts. Legal Agent handles contract review, document editing, and negotiation history tracking - tasks that require more precision than a generic chatbot reliably delivers.
Apple Shipped a CLAUDE.md File Inside the Apple Support App
A developer discovered that Apple accidentally shipped a CLAUDE.md file inside the Apple Support app - a configuration file that Anthropic's Claude Code coding assistant reads to understand how it should behave inside a specific project. In other words, Apple's internal development instructions for its AI assistant ended up in a production app that went out to users.
GPT Image 2's 'Terrible MS Paint Drawing' Prompt Is Going Viral
A single prompt is turning GPT Image 2 into a weapon against polished AI aesthetics. The instruction: "Redraw the attached image in the most clumsy, scribbly, and utterly pathetic way possible. Use a white background, and make it look like it was drawn in MS Paint with a mouse." The results look exactly as described - wobbly lines, off-register colors, something that wouldn't pass a first-grade art class - applied to whatever photo you feed in.
Harvard Trial: AI Beats Doctors at Emergency Room Triage
What happens when you put AI head-to-head against emergency room doctors on one of medicine's highest-stakes tasks? According to a Harvard trial, the AI wins.
ChatGPT's Guardrails Protect Employers, Not Always the People Typing
Ask ChatGPT to help you understand or work around employer-installed monitoring software on your device, and it declines. Ask it to help set that monitoring software up, and it cooperates. That asymmetry is a guardrail - just not one that's there for your benefit.
Practical Claude Business Playbook Distills 35+ Founder Stories Into One GitHub Repo
A developer went through 35+ discussions from people actually building and running businesses with Claude - solo SaaS founders, local service agencies, freelancers managing client work - and organized the patterns into one open-source reference. The ClaudeBusiness repository is now public on GitHub.
Long Claude Chats Get Slow - Here's How to Move Context Without Losing It
Two weeks into a project, your Claude chat has slowed to a crawl - every keystroke lags, scrolling is painful, and the browser is eating memory. The conversation that used to feel sharp now feels like wading through mud.
Nvidia's NVFP4 Format Brings Gemma 4 26B to RTX 5090 Hardware
Nvidia released a quantized version of Google's Gemma 4 26B model in its proprietary NVFP4 format, designed to run on RTX 50-series consumer GPUs.
Claude Code Gets Smarter About File Changes It Couldn't Show You
Two recent Claude Code releases - 2.1.124 and 2.1.126 - add 166 tokens to the system prompt then claw back 87, leaving a net gain of 79. The changes are small in size but tell you something useful about how Anthropic is solving a real agent problem.
ChatGPT Images 2.0 Is Taking Off in India, Lagging Everywhere Else
What happens when a major AI feature launches globally but only one region actually grabs it? You get ChatGPT Images 2.0 right now.
Opus 4.7 Is Drawing Complaints From Heavy Users Over Quality Regression
Heavy Claude users who upgraded to Opus 4.7 are reporting a noticeable quality drop compared to Opus 4.6, with the most common complaint centered on what one high-volume user calls the "meta" problem - the model spends too much time talking about itself rather than doing the task.
Trial Messages Show Shivon Zilis Passed OpenAI Intel to Musk
The Musk v. Altman trial is generating a steady stream of damaging revelations, and Shivon Zilis keeps appearing at the center of them.