All news

Zuckerberg admits to Meta staff: AI-agent progress lags behind plan

Anthropic vs Pentagon: Amodei's correspondence revealed dispute over military AI control

JetBrains Rethinks IDE: The Next Primary User Is an AI Agent

LLM·Anthropic Blog

Anthropic Revealed Cybersecurity Details of Fable 5 and Proposed a Jailbreak Severity Scale

Anthropic explained what exactly Fable 5's safety classifiers block and published the first draft of an industry-wide jailbreak criticality

Jul 3, 2026·3 min

Study of 21,000 Companies: AI Creates Jobs Rather Than Destroys Them

A large-scale study of 21,000 American companies showed: after implementing AI, enterprises begin to hire more employees — although the effe

Sysdig documents first autonomous AI-powered ransomware attack

Sysdig has documented an incident in which an LLM independently executed a complete attack cycle—from system infiltration to data destructio

Jul 3, 2026·3 min

LLM·TNW

Kling AI raises $2 billion as Kuaishou spins off video generation as separate business

Kuaishou spins off AI video generator Kling AI into an independent company: first tranche of the funding round closed at $2 billion, with to

01Wisk Aero employee sues: fired after challenging autonomous taxi software testing cutsJul 3, 2026·2 min 02New South Wales officials removed 'absolute delight' about OpenAI office due to Terminator jokesJul 3, 2026·2 min 03Anthropic in talks with Samsung to develop its own AI chipJul 3, 2026·2 min 04Oracle investors sue over concealment of risks in $300 billion OpenAI dealJul 3, 2026·2 min 05Australian musicians demand Albanese protect their work from AI scrapingJul 3, 2026·3 min

Top themes

LLM·Guardian

Meta launched AI search mode on Facebook: answers from posts, groups, and Reels

Facebook launched AI Mode, a Meta AI-powered search that pulls answers from public posts, groups, Reels, and Marketplace listings across the

LLM·Guardian

Iason Gabriel: how a philosopher at Google DeepMind is trying to make AI safe

Since 2017, political philosopher Iason Gabriel has worked at Google DeepMind, trying to make sense of AI's consequences — but can an ethici

ByteDance shifts to in-house AI chips: rollout to begin next year

ByteDance is preparing to deploy AI chips of its own design next year — this will reduce the company's dependence on US suppliers amid ongoing sanctions.

LLM·TNW

Anthropic Meets With the U.S. Department of Commerce: The Fable 5 and Mythos 5 Crisis Turns Political

LLM·Habr AI

Alena from Sber: why evaluating an LLM on 20 examples is more dangerous than not evaluating it at all

LLM·Habr AI

Transformers at the wheel: how Yandex uses AI to control self-driving vehicles

Jun 30, 2026·3 min

LLM·MarkTechPost

Anthropic Releases Claude Fable 5 and Mythos 5: One Model, Two Access Tiers

Jun 30, 2026·3 min

LLM·AI News

AI Red Teaming — How Companies Test AI Systems for Vulnerabilities Before Release

LangSmith Added End-to-End OpenTelemetry Support for LangChain and LangGraph

LangChain released Promptim, a library for automatic prompt optimization

Sarah Guo of Conviction: why LLM won't replace work that can't be measured

LLM·LangChain Blog

New Computer Raised Agent Memory Search Accuracy by 50% with LangSmith

Startup New Computer leveraged the LangSmith observability tool to achieve a 50% increase in memory retrieval completeness, improving prompt