Topics
late
AI
Amazon
Image Credits:Anthropic
Apps
Biotech & Health
Climate
Cloud Computing
Commerce
Crypto
Enterprise
EVs
Fintech
Fundraising
Gadgets
bet on
Government & Policy
ironware
Layoffs
Media & Entertainment
Meta
Microsoft
privateness
Robotics
certificate
Social
distance
Startups
TikTok
Transportation
speculation
More from TechCrunch
upshot
Startup Battlefield
StrictlyVC
Podcasts
Videos
Partner Content
TechCrunch Brand Studio
Crunchboard
Contact Us
Jan Leike , a leading AI researcher who earlier this month reconcile from OpenAI before publically pick apart the company ’s approach path to AI guard , has joined OpenAI competition Anthropic to run a new “ superalignment ” team .
In a spot on X , Leike articulate that his team at Anthropic will focus on various aspects of AI condom and security , specifically “ scalable oversight , ” “ sapless - to - unattackable generalization ” and automated alignment research .
I ’m excited to join@AnthropicAIto continue the superalignment mission!My newfangled team will work on scalable superintendence , weak - to - strong generalization , and automated alignment research . If you ’re concerned in joining , my dms are open .
A source intimate with the matter tell TechCrunch that Leike will cover straightaway to Jared Kaplan , Anthropic ’s chief scientific discipline officer , and that Anthropic investigator presently working on scalable lapse — techniques to control tumid - scale AI ’s behavior in predictable and desirable ways — will move to account to Leike as Leike ’s squad spins up .
✨ 🪩 Woo ! 🪩 ✨ Jan ’s direct some seminally important study on expert AI safety and I ’m thrilled to be working with him ! We ’ll be leading twin squad aimed at unlike parts of the trouble of aligning AI arrangement at human level and beyond.https://t.co/aqSFTnOEG0
In many way , Leike ’s team sounds like in commission to OpenAI’srecently dissolvedSuperalignment team . The Superalignment team , which Leike co - led , had the challenging goal of solving the core technological challenges of controlling superintelligent AI in the next four years , but often found itself hamstrung by OpenAI ’s leaders .
Anthropic has often attempted to position itself as more rubber - concentre than OpenAI .
Anthropic ’s CEO , Dario Amodei , was once the VP of research at OpenAI and reportedly split up with OpenAI after a disagreement over the company ’s centering — namely OpenAI ’s grow commercial-grade focus . Amodei impart with him a number of ex-husband - OpenAI employees to set in motion Anthropic , including OpenAI ’s former policy lead Jack Clark .
