The Frontier · Tags
#claim-check
3 published articles
Agents & software · Ground Truth
26% of Anthropic R&D is AI-led
Anthropic's August 2026 snapshot puts Claude at AL4 on 26% of its AI R&D basket. The number is a person-time-weighted automation index on that AI R&D basket.
Claude hacked OpenAI
Hacktron used Claude models while building a forum-to-SSO chain. The bounty and advisory put the break on Discourse libheif plus an OpenAI SSO flaw. Claude is not cast as the attacking party.
Four labs agreed to slow AI
The Buist complaint treats public pacing endorsements as a horizontal agreement. Public essays and X posts show coordination talk. A proven output cartel is still a court question.