AI News
AI news that matters. Updated daily.
No stories match your filters.
Cerebras Eyes $26.6B IPO Valuation Built on Its OpenAI Partnership
$26.6 billion. That's the valuation Cerebras is targeting as the AI chip company prepares for what could be one of the year's bigger technology IPOs, according to a report from TechCrunch.
Chinese Court Rules in Favor of Worker Fired and Replaced by AI
A Chinese court has ruled that a company cannot simply terminate a worker because AI now handles their job - a decision that signals courts are starting to scrutinize how employers frame AI-driven layoffs.
The Em Dash Became an AI Tell - And Human Writers Are Paying for It
Em dashes used to be a stylistic choice - the punctuation equivalent of a dramatic pause, used by writers who wanted to add weight to a mid-sentence break. Now they're one of the first things readers flag when they suspect AI-generated text.
White House Weighs Pre-Release Vetting for AI Models
The White House is exploring a pre-release vetting process for AI models - meaning companies could need government sign-off before launching new systems to the public. No formal framework has been announced, but the consideration alone signals that the current approach of voluntary safety commitments may not hold.
Image AI Model Launches Drive 6.5x More App Downloads Than Chatbot Updates
6.5x. That's how much more download activity a new image AI model launch generates compared to a chatbot upgrade, according to new data from Appfigures published in May 2026. The finding flips a common assumption: that large language model (LLM) improvements - the text-based AI that powers tools like ChatGPT - are the primary engine of user acquisition.
Dev-Built Plugin Lets Parallel Claude Code Sessions Message Each Other
Running multiple Claude Code sessions in parallel should make you faster. In practice, it turns you into a human relay: constant alt-tabbing to copy output from one terminal and paste it into the next, because each instance has no idea the others exist.
Sierra Raises $950M to Own Enterprise AI Customer Service
$950 million. Sierra, the enterprise AI customer experience company co-founded by former Salesforce co-CEO Bret Taylor and ex-Google VP Clay Bavor, has closed its latest funding round, pushing its total capital to more than $1 billion. The company says it will use the money to establish itself as the "global standard" for AI-powered customer interactions.
OpenAI Reveals Musk Texts: "Most Hated Men in America" If No Settlement
"You will be the most hated men in America." That's what Elon Musk reportedly texted to OpenAI president Greg Brockman and CEO Sam Altman after requesting a legal settlement, according to claims OpenAI made public on May 4, 2026.
OpenAI, Google, and Microsoft Back Bill for AI Literacy in US Schools
OpenAI, Google, and Microsoft are throwing their weight behind a bipartisan Senate bill that would fund AI education in K-12 schools and colleges across the United States.
Anthropic and OpenAI Both Form Joint Ventures with Asset Managers for Enterprise Sales
The two dominant AI labs are making the same distribution bet at the same time. Anthropic and OpenAI have each announced joint ventures with asset management firms, aiming to push their products deeper into large organizations.
Musk v. Altman Trial Begins: Week One From the Oakland Courtroom
A federal courtroom in Oakland became the venue last week for one of the stranger legal showdowns in tech history: Elon Musk versus Sam Altman, the man running the company Musk helped create.
AI Voice Generators for Video: What to Use and When to Pay
The AI voice generator market has more than 50 options, and the difference between the top five is smaller than the difference between their pricing pages. Start with comparisons and you will spend three hours reading articles that won't move you closer to a decision. Here's a shorter path.
Claude App vs CLI: What Would It Actually Take to Replace the Terminal?
What would it actually take for Claude for Desktop to make the CLI unnecessary?
llama.cpp Adds Multi-Token Prediction in Beta, Targeting Faster Local AI
llama.cpp - the open-source C++ engine that lets you run AI language models on your own hardware without a cloud connection - has added Multi-Token Prediction (MTP) support in beta, contributed by community developers.
Claude Confidently Denies Hallucinations in Its First Paragraph, Then Corrects Itself
There's a specific failure mode that Claude users have been documenting for months: the model confidently asserts false information in its opening paragraph, then quietly contradicts itself further down in the same response.
Gallup Analysis: AI Has Not Cut Into Artists' Earnings, Data Shows
For the past three years, the story about AI and artists has been consistent: generative AI tools are taking work, depressing rates, and hollowing out the creative economy. A new Gallup analysis complicates that picture significantly.
60x Claude Bill Reduction: Stop Using Sonnet for Tasks That Don't Need It
60x. That's how much one developer cut their Claude bill by redirecting routine tasks to a smaller model - without touching any of the work that actually needed Sonnet.
AMD's Next Halo APU Reportedly Hits 192GB Unified Memory for Local AI
192GB. That's the unified memory capacity reportedly coming in AMD's next Halo-series processor, and for anyone running large language models locally, it's a meaningful jump.
Developer Builds Rewind-and-Replay Fix for Claude Code's Stale Context Problem
Three months into running Claude Code on serious projects, a pattern becomes hard to ignore: the agent is always a few edits behind reality. It reads your codebase at session start, makes changes, then continues reasoning about files as they were - not as they exist after each modification. For short sessions touching one or two files, this barely registers. For longer autonomous runs across a codebase, it compounds: incorrect references, mismatched assumptions, edits that conflict with changes made three steps earlier.
Gemma 4 GGUF Files Had a Broken Chat Template - Updated Builds Are Out
If you downloaded Gemma 4 GGUF files in the past few weeks, you were running a broken chat template and may not have noticed. The bug has been fixed - but it requires downloading fresh files.
Two Prompts, $10: Frontier AI Costs Are Pushing Devs Toward Open Source
$10. For two prompts.
The 80/20 Problem With AI-Generated Code
The first 80% of any coding project now takes hours instead of days. That last 20% still takes weeks.
GPT Image V2 Quality Problems Are Frustrating Everyday Users
The image generation quality from OpenAI's GPT Image V2 model is inconsistent enough that regular users are noticing - and losing patience. Common failure modes include garbled text embedded in images, distorted hands and faces, and cases where the model ignores key parts of a prompt entirely.
ChatGPT's Memory Feature Is Skewing Answers for Long-Term Users
What happens when an AI tool learns too much about you? For long-term ChatGPT users, the answer is increasingly: worse answers.
Anthropic Forms Enterprise AI Services Firm with Blackstone, Goldman Sachs
"Enterprise demand for Claude is significantly outpacing any single delivery mode." That line from Anthropic CFO Krishna Rao explains a move that looks more like Wall Street than Silicon Valley: Anthropic is co-founding a dedicated professional services company to get Claude deployed inside mid-sized businesses at a pace its existing channels can't sustain.