AI news of 28 August 2026 · Robots Versus Humans

Robots

Yesterday

Google Deepmind's AI Co-Scientist now plans experiments, runs lab equipment, and writes scientific papers

Google Deepmind's AI Co-Scientist now designs experiments, operates lab equipment and writes research papers across multiple disciplines.

Google Deepmind has expanded Co-Scientist from a hypothesis generator into a research system that's integrated into the lab. Across three disciplines, from materials synthesis to the autonomous development of a medical AI architecture, the Gemini-based multi-agent system delivered experimentally validated results.

The Decoder 5h ago

AI Has Human Doctors Asking: What’s Left for Us? Opens at the publisher in a new tab

Research paper argues AI often outperforms human doctors, raising concerns among physicians.

Wired 9h ago

Meta researchers taught an 8B AI model to match Claude Opus 4.5 — without the frontier price tag

Meta's 8B AI model matches Claude Opus 4.5 performance at significantly lower cost.

VentureBeat 8h ago

Always-on and self-starting AI agents might be OpenAI's next big play

OpenAI is testing always-on AI agents that generate their own tasks without human instruction.

The Decoder 16h ago

Anthropic gets its first court win over the Pentagon’s supply chain risk label Opens at the publisher in a new tab

Federal judge rules Pentagon illegally labeled AI company Anthropic a supply chain risk.

TechCrunch 11h ago

Open-weight AI companies are the Valley’s hottest acquisition targets Opens at the publisher in a new tab

Open-weight AI companies attract major acquisition interest as capital flows into the market for freely distributed models.

TechCrunch 6h ago

Cohere Parse 5 loses the benchmark on points. It wins on cost per page.

Cohere releases Parse 5 document extraction tool with improved cost efficiency over competitors.

VentureBeat 9h ago

GLM-5.3 is now open-weight Opens at the publisher in a new tab

GLM-5.3 open-weight AI model released.

Hacker News 9h ago also at Hacker News

Supporting Thailand’s next generation of AI startups Opens at the publisher in a new tab

OpenAI and Thailand's MHESI launch eight-week accelerator for 10 AI startups in health, wellness and education.

OpenAI 22h ago

Breaking Claude Code Opus 5 Auto Mode

Anthropic's Claude Code auto mode defaults to new security approach, raising questions about prompt injection protection.

Simon Willison 1d ago

Yesterday

Google Deepmind's AI Co-Scientist now plans experiments, runs lab equipment, and writes scientific papers

AI Has Human Doctors Asking: What’s Left for Us? Opens at the publisher in a new tab

Meta researchers taught an 8B AI model to match Claude Opus 4.5 — without the frontier price tag

Always-on and self-starting AI agents might be OpenAI's next big play

Anthropic gets its first court win over the Pentagon’s supply chain risk label Opens at the publisher in a new tab

Open-weight AI companies are the Valley’s hottest acquisition targets Opens at the publisher in a new tab

Cohere Parse 5 loses the benchmark on points. It wins on cost per page.

GLM-5.3 is now open-weight Opens at the publisher in a new tab

Supporting Thailand’s next generation of AI startups Opens at the publisher in a new tab

Breaking Claude Code Opus 5 Auto Mode

Latest AI News

Choose sides Choose whose side you're on

Thirsty for more?

Install it as an app

Spread the word

PWA successfully installed.
Take that, Apple! 😉