2026
10,780 articles · 136 days · 10 categories · avg 79 per day

US lifts restrictions on Anthropic's Fable 5 model after easing curbs on Mythos 5
On July 15, 2026, US authorities allowed Anthropic to lift export restrictions on Fable 5, days after the more powerful Mythos 5 received similar relief.

Anthropic holds talks with US officials over the Mythos 5 and Fable 5 ban
Anthropic executives met with US Commerce Department officials on Monday to discuss Friday's ban on access to the Mythos 5 and Fable 5 models.

Yandex's head of developer relations: AI in code is an exoskeleton, not an autopilot
Sergey Berezhnoy of Yandex explained on Habr why creativity has become a developer's key skill in the era of AI-generated code.

Microsoft launches Frontier Company, an AI deployment business with 6,000 engineers

Yandex Music: 37% of New Releases in H1 2026 Were AI-Generated

Anthropic restores Claude Fable 5 after US export restrictions are lifted

SpaceX showed investors a prototype AI device with xAI technology, Musk denies it

Honda starts building data center batteries at plant meant for canceled EVs
Honda has started making batteries for data center energy storage at an Ohio plant originally meant for EVs, Nikkei Asia reports.

Ashton Kutcher Leaves Sound Ventures for a New AI Fund With Morgan Beller
According to TechCrunch, actor and investor Ashton Kutcher is leaving Sound Ventures to launch an AI infrastructure and energy fund with Morgan Beller.

Study: An AI planner sharply raises the risk that an executor carries out harmful requests

Researchers propose a new way to measure AI intelligence beyond the human level

Study: AI agent trained on explicit social norms coordinates with humans nearly 4 times better

Large Behavior Model: A Language Model That Predicts Purchases Like a Digital Twin

UASPL: a new self-paced training method that accounts for prediction uncertainty
Researchers have introduced UASPL, a method that picks training examples for a neural network based not only on the size of the error but also on how reliable the…

Simon Willison used GPT-5.5 to build a web component for embedding GitHub code
Developer Simon Willison built a github-code web component with a single GPT-5.5 prompt. It embeds code snippets from GitHub into HTML pages with line numbers.

Simon Willison releases llm 0.31.1 with a fix for a tool-call bug
Simon Willison, developer of the open-source CLI tool llm, has released version 0.31.1, fixing an error that occurred when tools were called with empty arguments.

Habr translation explains how transformer-based LLMs actually work

Quaise Energy raises $134M to drill into superhot rock with microwaves

Google DeepMind launches Gemini assistant ATL Saathi for 1.1 million Indian schoolchildren

OpenAI lands $520 million credit line from Bank of America

Hud CEO Roee Adler: Runtime Intelligence Is the Next Stage of Software Operations
Hud CEO Roee Adler explains why logs, metrics, and traces no longer suffice now that AI agents write production-ready code in minutes.

Google DeepMind and Isomorphic Labs unveil joint AI-powered bioresilience program
On July 16, 2026, Google DeepMind and Isomorphic Labs outlined their bioresilience approach: safeguards against AI model misuse and tools for biosecurity experts.

Amazon shows Nova-driven pipeline using SAM 3 and Textract to automatically redact PII in images

Romania launches CyberCUBE, the first private mission led under the European Space Agency

SpaceX joins the Nasdaq 100: one analyst values it at $10.5 trillion, while Grantham expects a crash

MNTN CEO Mark Douglas doubts Meta can take on AWS and Google Cloud

Luffy AI raises £8.1M for self-tuning AI that controls electric motors
UK startup Luffy AI has closed an £8.1 million Series A to build "neuroplastic" AI that controls motors in real time.

AWS shows how to run MiniMax models on Amazon Bedrock
The AWS Machine Learning Blog has published a walkthrough of running MiniMax models on Amazon Bedrock, covering service tiers, scaling and the APIs available for…

SpaceXAI and Cursor Prepare Their First Joint AI Model to Take On Opus 4.8 and GPT-5.5
SpaceXAI and Cursor may release their first jointly developed AI model as early as July 15, 2026. An internal memo benchmarks it against Opus 4.8 and GPT-5.5.

Jensen Huang: a $500,000 engineer should spend at least $250,000 a year on AI tokens

CISA Is Using Anthropic's Mythos Model to Audit US Government Code

Typhoonminigen 2.0: an open-source macOS app for running FLUX.2 Klein locally

















