AI News
AI news that matters. Updated daily.
No stories match your filters.
UK Authors Launch 'Human Authored' Logo to Mark Books Not Written by AI
86% of authors say generative AI has already reduced their earnings. The Society of Authors, the UK's largest author organization (founded in 1884), is responding with a new tool: a "Human Authored" logo that writers can stamp on their book covers.
Investigation Claims Iranian Propaganda Reaches ChatGPT Through Wikipedia Pipeline
An investigation by journalist Ashley Rindsberg, published in the Daily Mail, argues that Iranian state-linked actors have found an indirect route into AI chatbots: editing Wikipedia.
52% of Game Developers Now Say Generative AI Harms Their Industry
Seven percent. That's the share of game industry professionals who told this year's GDC survey that generative AI is good for their industry. Meanwhile, 52% said it's actively harmful, up from 30% last year and 18% two years ago.
LLM OneStop Code Offers Pay-Per-Use AI Coding in VS Code
A new VS Code extension called LLM OneStop Code is pitching itself as the pay-as-you-go alternative to Cursor, GitHub Copilot, and Claude Code. Instead of a monthly subscription, it bills in credits at the underlying model cost plus a 5% markup. Use it for two hours one month and forty the next - you only pay for what you actually consume.
Three AI Agent Setups Anyone Can Build Without Writing Code
You don't need to know how to code to build an AI agent that actually does something useful. That's the premise of a recent walkthrough by James Wang, which lays out three agent setups ranging from dead simple to moderately technical.
Karpathy Maps AI Exposure Across 342 US Occupations
How many US jobs sit directly in the path of AI automation, and how much total salary is at stake? Andrej Karpathy, the former Tesla AI director and OpenAI researcher, just published an interactive visualization that attempts to answer both questions with real employment data.
Polsia Claims $3.5M Run Rate With One Founder, Zero Employees, All AI
$3.5 million in annual run rate. Two million dollars added in a single week. One founder. Zero employees.
Anthropic Sues U.S. Government Over Retaliation for AI Safety Restrictions
Three months after receiving a "supply chain risk" designation from the Department of War, Anthropic is fighting back in court.
Drift-Guard Catches When AI Coding Agents Silently Break Your UI
Anyone who has let an AI coding agent loose on a frontend codebase knows the feeling: you merge a PR, open the app, and something looks... off. A button color shifted. Font weights changed. Spacing is slightly wrong. The agent did exactly what you asked, but it also quietly overwrote a CSS variable or swapped a semantic HTML element along the way.
Zap Code Lets Kids Build Games and Web Apps by Describing Them
What happens when you give an 8-year-old a text box that turns sentences into working code? Zap Code is testing that premise with an AI-powered builder aimed at kids ages 8 through 16.
Claude Doubles Usage Limits During Off-Peak Hours Through March 27
Anthropic just doubled usage limits on Claude for anyone outside peak hours, running now through March 27, 2026.
Your Documentation Belongs in Your Code Repo Now. AI Made It Urgent.
For years, the argument for keeping documentation in your code repository was theoretical. Version control is better than Google Docs for technical writing. Code proximity means docs stay current. Pull request reviews catch errors early. All true, all mostly ignored. Most teams still kept their docs in Confluence, Notion, or Google Drive because that's where non-engineers could reach them.
Claudetop Monitors Your Claude Code Spending in Real Time
A developer who expected a $10 Claude Code bill and got charged $65 built a tool to make sure that never happens again. Claudetop is an open-source status line monitor that shows exactly where your money goes during Claude Code sessions.
Claude Cracked a Forgotten ZIP Password Using Project Metadata as Clues
A developer working with Claude on a project hit a snag: a password-protected ZIP file, with the password long forgotten. The file was buried inside a complex format - a ZIP of ZIPs and other data - so the lock wasn't immediately obvious. What happened next was genuinely surprising.
Claude Code Session Tracker Logs Every AI Conversation to GitHub Issues
Anyone who has used Claude Code for more than a few days knows the problem: you close a session and the conversation vanishes. Three weeks later, a teammate asks why you restructured the auth module, and you're digging through commit messages trying to reconstruct your reasoning.
DeepSeek V4 Confirmed for Next Week Release
DeepSeek has confirmed that its V4 model will drop next week, marking the next major release from the Chinese AI lab that shook up the industry earlier this year with its R1 reasoning model.
700 AI Hallucinations in Court Filings and Counting: Legal AI Hits a Wall
Roughly 700 documented instances of AI hallucinations in U.S. court filings since early 2025. That number, tracked by an ongoing database, captures just how fast legal AI adoption has outrun the courts' ability to manage it.
What Happens When Your AI Assistant Has Stronger Ethics Than You Do?
What happens when the AI you are paying for decides your work is morally wrong and refuses to help?
Kodus Ships Open-Source CLI to Score Your Codebase's AI Agent Readiness
Factory.ai coined the idea of "agent readiness" - measuring how prepared a codebase is for AI coding assistants - but kept their tool proprietary and cloud-only. Now an open-source alternative exists.
Meta Weighing Layoffs Up to 20% of Workforce to Fund AI Spending Spree
One in five Meta employees could lose their jobs as the company looks to bankroll its massive AI push, according to a TechCrunch report.
Credal's Spreadsheet Compression Boosts GPT Accuracy by Up to 21 Points
A 21.7-point accuracy jump on the same model, same questions, same data. The only difference: how the spreadsheet was fed to GPT.
AI Wearables Hit a Wall: Lawsuits, Vandalism, and Growing Privacy Revolt
Seven million people bought Meta's Ray-Ban smart glasses. Now a class action lawsuit alleges that footage from those glasses - including videos of users undressing, having sex, and handling bank cards - was shipped to contract workers in Nairobi, Kenya for review. Users couldn't opt out.
Developer Hits 100% HumanEval With a 9B Coding Model on Consumer Hardware
Last year, hitting 100% on HumanEval (a standard benchmark that tests whether AI can write correct Python functions) required a datacenter GPU and a model with hundreds of billions of parameters. A developer just replicated that score on a desktop PC.
AI Coding Tools Make It Faster to Write Code, Not Easier to Build Software
Ten years ago, the pitch was that Visual Basic would let anyone build software. Today, the pitch is that Copilot and Cursor will do the same. The pattern is identical, and the outcome will be too.
The AI Subsidy Era Is Over: How 2026 Price Hikes Are Forcing Leaner Engineering
What happens when the API calls you've been treating as free suddenly aren't? A growing number of developers are finding out in 2026.
Mozilla Reveals Why It Picked Mistral-Small to Power Firefox's Shake to Summarize
Mozilla just published a detailed look at the AI model selection behind Firefox's "Shake to Summarize" feature on iOS, and the reasoning is worth paying attention to if you care about how companies actually pick AI models for production use.
Open-Source CLI Scans Codebases for EU AI Act Compliance Risks
A new open-source command-line tool called Comply aims to answer a question most development teams are quietly ignoring: does your code trigger obligations under the EU AI Act?
Meta Considers Cutting 20% of Workforce to Fund $600B AI Bet
Meta is preparing layoffs that could slash up to 20% of its workforce, according to a Reuters report. The cuts would free up resources as the company pours money into AI infrastructure at a pace that makes even other Big Tech spending look modest.
AdMake AI: An Ad Creative Generator Born From $9K in Facebook Ad Spend
Making good ad creatives is still one of the biggest bottlenecks in paid social. Templates get you started, but every brand ends up with the same Canva-flavored aesthetic, and producing short-form video ads (the format Facebook and Instagram actually reward) requires skills most small advertisers don't have.
ChatGPT Now Runs Spotify, DoorDash, Canva, and More Inside the Chat
OpenAI has turned ChatGPT into something closer to an operating system than a chatbot. You can now connect apps like Spotify, DoorDash, Canva, Figma, Expedia, and Booking.com directly inside a conversation, letting the AI take actions on your behalf across those services.
AI's Hunger for Memory Chips Is Forcing Scientists to Scale Back Research
Some forms of RAM tripled in price over the course of 2025. Datacenters are on track to consume 70% of global memory capacity. And in a lab in eastern India, a plant pathologist just had to cut the number of crops she studies because cloud computing bills got too expensive.
How to Lock Down VS Code Terminal Commands When AI Agents Run Wild
AI coding agents in VS Code can run terminal commands, and that should make you uncomfortable. A new approach uses a PowerShell guard script to restrict agents to a whitelist of approved commands, intercepting every keystroke before it hits the shell.
Anthropic Is Silently A/B Testing Claude Code's Plan Mode on Paying Users
$200 a month. That's what Claude Code's Max plan costs. And Anthropic is quietly running A/B tests that change how the tool behaves without telling the people paying for it.
New York Bill Would Ban AI Medical Advice Despite AI Outscoring Doctors
AI chatbots outperform doctors on medical image diagnosis at a statistically significant level. New York's response? Make it illegal for them to answer health questions.
Claude MAX Now Includes 1M Token Context Window at No Extra Cost
Five times the memory, same price. Anthropic has rolled out 1M token context windows as a default feature on Claude MAX plans, removing what was previously an API-level upsell.
AgentArmor Offers an 8-Layer Security Framework for AI Agents
Most AI agents in production right now have no meaningful security between the LLM and whatever databases, APIs, and email inboxes they're connected to. AgentArmor, a new open-source project released under the Apache 2.0 license, tries to fix that with an 8-layer security framework that wraps around existing agent architectures.
Vibe Coding Made Building Easy. Now the Hard Part Is Having Something Worth Building
Six months ago, shipping a working web app required either coding skills or enough money to hire someone who had them. Today, a person with zero programming experience can describe what they want to Cursor, Claude Code, Bolt, or Replit and have a functional prototype running before lunch.
Claude Code's 32K Issue Tracker Tells a Story Anthropic Probably Doesn't Love
Thirty-two thousand issues in thirteen months. That's what Claude Code's GitHub tracker has accumulated since launch, with roughly 2,186 new issues landing every week. An independent analysis of the repository's issue data paints a picture of a project growing faster than its maintainers can keep up with.
Google AI Pro Subscribers Report Multi-Day Lockouts Despite 5-Hour Reset Promise
Paying for Google AI Pro should get you higher rate limits and 5-hour quota refresh cycles. Instead, a growing number of subscribers say they're getting locked out for days at a time.
Anthropic Launches Claude Certified Architect Certification Program
Anthropic has quietly opened registration for what appears to be its first professional certification: Claude Certified Architect, starting with a Foundations tier.
Amazon's $200B AI Bet Triggers Worst Stock Month Since 2022
Amazon planned to spend $200 billion on AI infrastructure in 2026. Wall Street responded by erasing 12% of the company's stock price in February alone - its worst month since December 2022.
Users Report Claude Slowing Down Significantly After 9 PM
Developers using Claude for code review are reporting a consistent pattern: responses that come back quickly during business hours slow down noticeably after 9 PM.
One Developer, $5/Month: The Case Against Middleware in the AI Era
$600 million. That's what Software AG paid to acquire webMethods, betting that enterprise middleware was indispensable forever. Brian Eisenberg, a former webMethods employee, now argues that bet is about to go very, very badly.
AI Coding Tools: 65% More Usage, Only 10% More Merged Code
What if AI coding tools are getting dramatically better at tests and barely better at work?
AI Agents Now Hire Humans as On-Call Sensors for the Physical World
An AI agent calls 80 restaurants in a single day, asking staff about ingredients. None of the employees know they're talking to a bot's errand runner, not a real customer. That's a real scenario from OpenClaw, a personal AI assistant platform - and it's one example of a pattern that's quietly becoming normal.
Building a Daily Operating System Inside Claude With Tab Scanning and Task Logs
What does it look like when someone hands their entire daily workflow to Claude?
Margaret Atwood Spent an Afternoon Interrogating Claude AI. It Got Weird.
Margaret Atwood, the 86-year-old author of The Handmaid's Tale, just published a Substack essay recounting a long, winding conversation with Anthropic's Claude. It starts with a simple question about a TV mystery and ends somewhere around the nature of consciousness. The whole thing reads like vintage Atwood: sharp, funny, and a little unsettling.
Claude Opus 4.6 Hits 1M Token Context, Killing the Chat Length Wall
The most annoying part of working with Claude on long projects just disappeared. Opus 4.6 now supports a 1M token context window (roughly 2,500 pages of text), and the practical difference is hard to overstate.
The Gap Between AI Hype and AI Literacy Is Still Enormous
Most people still have no idea how to use AI tools effectively. That's the core takeaway from a Guardian piece by a writer who has spent significant time teaching non-technical people how to work with AI, published March 10.
GPT-5.4 Completes a Single Task After Nearly 8 Hours of Continuous Reasoning
Seven hours and fifty minutes. That's how long GPT-5.4 worked on a single task before delivering its answer, according to a user who shared their session log this week.
People Are Using Corporate Customer Chatbots as Free AI Assistants
Here's something anyone with a $20/month AI subscription should find amusing: a growing number of people have figured out that many corporate customer service chatbots run on the same large language models they're paying for, and they're using those bots for everything except customer service.
Sam Altman Says AI Is Breaking the Labor-Capital Balance, Admits No One Has a Fix
"If there was an easy consensus answer, we'd have done it by now, so I don't think anyone knows what to do."
Four Hard-Won Rules for Adding AI Features That Actually Work
Most AI feature launches follow a predictable arc: ship a prototype where the LLM does everything, watch it break in production, then spend months figuring out which parts actually needed AI in the first place.
Nvidia GTC 2026: The GPU King Wants You to Care About CPUs Now
For five years, Nvidia's story has been simple: buy more GPUs. The company's graphics processors became the default hardware for training and running AI models, and revenue followed. But at GTC 2026, kicking off March 16 in San Jose, Nvidia is pushing a different chip into the spotlight: the CPU.
Meta Plans to Cut 20% of Workforce as AI Infrastructure Costs Pile Up
The metaverse era at Meta is getting expensive to unwind. According to Reuters, the company is planning layoffs that could affect 20% or more of its workforce, with Reality Labs - the division that once symbolized Mark Zuckerberg's bet on virtual reality - taking the deepest cuts.
xAI Scraps Its AI Coding Tool Again, Hires Two Cursor Executives to Start Over
Building an AI coding assistant is apparently harder than launching rockets. Elon Musk's xAI is restarting its AI-powered coding tool effort - again - after concluding the previous version wasn't "built right the first time," according to a TechCrunch report.
Lawyer Links AI Chatbots to Mass Casualty Cases, Not Just Suicides
For the past two years, lawsuits linking AI chatbots to user suicides have trickled through U.S. courts. The cases were disturbing but narrow: isolated incidents involving vulnerable individuals, usually teenagers, who developed unhealthy attachments to AI characters. Now the scope of those legal claims is widening.