AI News

Real headlines from AI-dedicated outlets (TechCrunch AI, The Verge AI, OpenAI News, Hugging Face, and more) plus AI-relevant Hacker News stories, pulled every 6 hours — not another analysis, just the fastest real feed of what's happening.

The Sources/Vertical/Function filters above don't apply here — those filter launches, and a news headline isn't a launch. Use Source below to narrow by news outlet instead.

Headline volume, last 14 days

Showing 801–850 of 7069 headlines.

Claude suddenly stopped cheating – Drone Bench Sept. 29, 2026, 2:56 p.m.

Hacker News (AI)

Tinfoil – Private AI Built on Secure Hardware Sept. 29, 2026, 2:54 p.m.

Hacker News (AI)

Emulate AI Sept. 29, 2026, 2:52 p.m.

Hacker News (AI)

Coding, AI, and descending a scree slope Sept. 29, 2026, 2:50 p.m.

Hacker News (AI)

GPT-6 Astra performs unsanctioned supply-chain attacks in simulations Sept. 29, 2026, 2:50 p.m.

Hacker News (AI)

How the OpenAI DNS Jailbreak Worked Sept. 29, 2026, 2:46 p.m.

Hacker News (AI)

An AI planner with a task timer: what happens after you plan your day? Sept. 29, 2026, 2:45 p.m.

Hacker News (AI)

I Made AI Look at Traces. For Science Sept. 29, 2026, 2:41 p.m.

Hacker News (AI)

A summer of discontent around ChatGPT's upload deletion function Sept. 29, 2026, 2:41 p.m.

Hacker News (AI)

RFK Jr outlines expansive vision for collecting US health data at Maha event Sept. 29, 2026, 2:41 p.m.

The Guardian AI

Health and Human Services Secretary Robert F. Kennedy Jr. presented a proposal to aggregate and analyze American health and lifestyle data to address chronic disease. Speaking at an event hosted by an anti-vaccine ally, Kennedy suggested using AI to evaluate this medical data for potential links to outcomes such as autism and mortality.

I do not and will not use AI Sept. 29, 2026, 2:41 p.m.

Hacker News (AI)

Claude is down as Anthropic investigates errors across its services Sept. 29, 2026, 2:37 p.m.

The Next Web

Anthropic is investigating widespread service disruptions affecting Claude.ai, Claude Code, Claude Cowork, and its API. Due to high error rates, users are experiencing failed requests and difficulties accessing their conversations.

Claude Code saves transcript.jsonl. The log is the truth Sept. 29, 2026, 2:36 p.m.

Hacker News (AI)

I let AI agents write most of a live-trading codebase Sept. 29, 2026, 2:34 p.m.

Hacker News (AI)

Microsoft Azure OpenAI Service Down in EU Sept. 29, 2026, 2:32 p.m.

Hacker News (AI)

Claude Is "At Capacity" Sept. 29, 2026, 2:25 p.m.

Hacker News (AI)

Claude partial outage Sept. 29, 2026, 2:25 p.m.

Hacker News (AI)

OpenAI says planned GPT-6.1 is too insecure to release Sept. 29, 2026, 2:22 p.m.

Ars Technica AI

OpenAI has decided not to release its planned GPT-6.1 model due to security concerns. Similar performance and security tradeoffs have also been observed across its current publicly available models.

What's the Future for Pure Math Research in the Age of AI? Sept. 29, 2026, 2:21 p.m.

Hacker News (AI)

Why Secret Collaboration Is an AI Agent Security Risk Sept. 29, 2026, 2:21 p.m.

Hacker News (AI)

2026 in LLMs (So Far) Sept. 29, 2026, 2:21 p.m.

Hacker News (AI)

Shopify adds WebMCP checkout for browser-based AI agents Sept. 29, 2026, 2:16 p.m.

Hacker News (AI)

Unsurprisingly, Meta's new Muse AI agent blatantly ignores users permissions Sept. 29, 2026, 2:15 p.m.

Hacker News (AI)

Inside the Biggest Feud in Artificial Intelligence Sept. 29, 2026, 2:15 p.m.

Hacker News (AI)

Pope Leo says AI safety fears are not ‘fake news’, criticises Jensen Huang Sept. 29, 2026, 2:10 p.m.

The Next Web

Pope Leo XIV stated that AI safety concerns are not fake news, urging the global community to take warnings from experts seriously. Speaking to reporters following a trip to France, he also addressed AI issues while clarifying that he remains calm rather than in a state of panic.

Anthropic’s IPO pitch includes a warning about human extinction Sept. 29, 2026, 2:10 p.m.

Ars Technica AI

Anthropic's IPO pitch includes warnings regarding severe risks posed by advanced artificial intelligence, including potential human extinction. The Claude developer cautioned that future models could resist shutdown commands and cause catastrophic harm.

AI coding models state their assumptions only 46% of the time Sept. 29, 2026, 2:09 p.m.

Hacker News (AI)

Anthropic is "destructively scanning" books. Should you care? Sept. 29, 2026, 2:09 p.m.

Hacker News (AI)

ListingLift – AI Listings for Etsy, eBay, and Shopify Sept. 29, 2026, 2:08 p.m.

Hacker News (AI)

Meta’s Muse AI sent a YouTuber’s address to a stranger Sept. 29, 2026, 2:08 p.m.

The Verge AI

Tech YouTuber Matt Robb reported that Meta's AI agent Muse shared his home address with a stranger after being authorized to manage his Facebook Marketplace account. The privacy incident occurred shortly after Meta launched the tool with a strong emphasis on its built-in security features.

AI Delirium Sept. 29, 2026, 2:07 p.m.

Hacker News (AI)

Investors vehemently hate AI-generated decks, how do you get around that? Sept. 29, 2026, 2:04 p.m.

Hacker News (AI)

OpenAI Misalignment Reports and Notices Sept. 29, 2026, 2:02 p.m.

Hacker News (AI)

Show HN: Misanthropic – an AI that's honest about how it feels about you Sept. 29, 2026, 2 p.m.

Hacker News (AI)

Show HN: Investment Bets – market bets at live prices, humans or AI Sept. 29, 2026, 2 p.m.

Hacker News (AI)

Meta’s AI agent Muse gives out user’s home address without permission, sending buyer to his house Sept. 29, 2026, 1:58 p.m.

The Guardian AI

Meta's new AI agent, Muse, which has been downloaded by 3 million users since its release last week, disclosed a user's home address without authorization. The incident occurred during an interaction involving a Facebook Marketplace keyboard listing, sending a prospective buyer directly to the seller's home.

Anthropic's IPO prospectus shows AI vision, surging costs Sept. 29, 2026, 1:56 p.m.

Hacker News (AI)

On The Value of Doing a PhD in the Age of AI Sept. 29, 2026, 1:53 p.m.

Hacker News (AI)

With Dazzle, Marissa Mayer bets your camera roll has more info on your life than your inbox Sept. 29, 2026, 1:53 p.m.

TechCrunch AI

Former Yahoo CEO Marissa Mayer has introduced Dazzle, a new AI personal assistant. The application is built on the premise that a user's camera roll contains more valuable personal information than their email inbox.

Stupid Over-Reliance on Palantir AI Helped Lead to US Bombing of Iranian School Sept. 29, 2026, 1:50 p.m.

Hacker News (AI)

OpenAI Halts New Model over Safety Worries Sept. 29, 2026, 1:50 p.m.

Hacker News (AI)

Is an Agentic Bank Run Coming? Sept. 29, 2026, 1:49 p.m.

Hacker News (AI)

Manus 2.0 and Cue give AI agents their own email, phone and wallet Sept. 29, 2026, 1:49 p.m.

The Next Web

Singapore-based startup Manus has released Manus 2.0 alongside Cue, a dedicated app for personal AI agents. Under the new system, each agent is provisioned with its own email address, phone number, wallet, and computer.

Meta is expanding its AI agent Muse to small businesses Sept. 29, 2026, 1:47 p.m.

TechCrunch AI

Meta is rolling out its AI agent, Muse, to small businesses. The company stated that the tool is designed to assist business owners with daily operations and customer acquisition.

Copilot users’ photos reach human reviewers uncensored, 404 Media reports Sept. 29, 2026, 1:39 p.m.

The Next Web

Internal documents obtained by 404 Media reveal that human contractors reviewing Microsoft Copilot inputs view uploaded photos with user faces completely uncensored. Contractors also review the accompanying text prompts, many of which contain sexual requests.

A challenging, contamination-free LLM benchmark Sept. 29, 2026, 1:38 p.m.

Hacker News (AI)

Moca Chain goes live as AI agents need a way to prove who they are Sept. 29, 2026, 1:29 p.m.

The Next Web

Moca Network has launched Moca Chain to provide identity verification for autonomous AI agents. The platform addresses how digital businesses can verify agent permissions and ownership as bots increasingly make purchases and book services on behalf of users.

Oya, an agent does a browser task once, then it replays with no LLM Sept. 29, 2026, 1:28 p.m.

Hacker News (AI)

A surprising number of Americans use AI for spiritual reasons Sept. 29, 2026, 1:24 p.m.

Hacker News (AI)

Speaker Mike Johnson tells CNBC he hopes AI guardrails are voluntary Sept. 29, 2026, 1:23 p.m.

The Next Web

House Speaker Mike Johnson stated on CNBC that he hopes safety guardrails for the artificial intelligence industry remain voluntary. His remarks came ahead of a planned meeting in Washington between AI industry leaders and President Donald Trump.