GLM-5.3: How Chinese labs keep stride with the frontier
Z.ai announced GLM-5.3, only in the coding plan, will soon be on its API and Hugging Face, showing an astounding performance boost.
Daily edition
Z.ai releases GLM-5.3 while Anthropic adds watermark detection API and OpenAI launches Computer History
Daily edition
50 of 50 feeds responded
181 distinct stories
82 previously issued, held back
9 worth your time
Collection and selection
181 candidate events; 39 shortlisted stories; 33 with article text available; 9 selected. Counts describe different stages; clustering, validation and refills can change the shortlist.
Active filters: none. Sort: Ranked.
All 9 stories, in ranked orderMore filters
Anthropic announces watermark detection API that will let third parties detect Claude's AI texts
Anthropic says its watermark detection API lets developers identify Claude‑generated text using a variant of Google’s SynthID method without affecting output quality.
GLM-5.3 is here with advanced cyber capabilities — and reportedly already found a 'serious vulnerability' in Cursor
Z.ai developer advocate Lou reports GLM-5.3 already uncovered a potentially serious vulnerability in Cursor, the SpaceX‑acquired coding startup.
OpenAI's Computer History turns your clicks and keystrokes into a searchable ChatGPT memory timeline
OpenAI says Computer History records local clicks, keystrokes, and app switches as unencrypted Markdown, providing searchable context for ChatGPT and Codex.
DeepSeek API Pricing Update
DeepSeek announced the GA release of V4‑Pro with major agent upgrades, flexible reasoning effort, and native OpenAI API support.
🔮 The market misread Google’s AI exodus
The analysis argues Jeff Dean and Sanjay Ghemawat’s exit signals an earlier AI investment cycle than the market perceives.
The Next Big Influencer Is This 4-Foot-Tall Robot From China
Unitree’s G1 robot, a low‑cost 4‑foot humanoid, has gained online fame but its practical job suitability remains uncertain.
Don't classify. Hallucinate!
Simon Willison notes that prompting an LLM to output tags without a predefined list enables vector matching to existing blog taxonomy.
Redacted Risk Report August 2026
Anthropic’s August 2026 Redacted Risk Report outlines the company’s latest safety assessments and risk mitigation strategies.
No stories match. Active filters.