The "pelican benchmark" by Simon Willison arbitrates Kimi K3
An SVG illustration test acts as a quick progress gauge: Simon Willison publishes his reading of K3 on July 16 and 17.
4 h ago 3 3
An SVG illustration test acts as a quick progress gauge: Simon Willison publishes his reading of K3 on July 16 and 17.
4 h ago 3 3
OpenAI has built GPT-Red, a purpose-built model to probe prompt injection vulnerabilities, per Tech in Asia. Two things happen when you name your red-team model: you signal a discipline, and you create a market.
yesterday 9 9
Nvidia is partnering with Toyota's Woven City on physical AI hardware and software. The reason it matters: Woven City is not a demo - it's an actual populated test bed, and Nvidia just booked the seat next to the engineers.
yesterday 9 9
Anthropic is preparing a US listing as early as October, according to Tech in Asia. The number nobody has yet: at what price the "safety-first" narrative sells to public investors when OpenAI still sets the tempo.
yesterday 9 8
A public benchmark by Stripe shows that AI agents write integrations that compile, run, and silently fail on the edge cases that matter.
Jul 15, 2026 at 19:28 10 6
A Rust port of PostgreSQL now passes all official regression tests. This is not a toy - it's a signal.
Jul 11, 2026 at 17:19 12 10
Rest of World documents a Bhashini hackathon focused on offline models in Indian languages. A bet: making AI useful without American cloud.
Jul 11, 2026 at 10:44 12 8
The first open-weights model from Thinking Machines Lab (Mira Murati) is here: Inkling, MoE 975B params (41B active), multimodal, Apache-2.0 license. The most discreet box in the frontier scene enters the open market through the front door.
yesterday 8 8