Sveriges mest populära poddar
TechCrunch Startup News
NyheterTeknologi

A safety institute advised against releasing an early version of Anthropic’s Claude Opus 4 AI model

4 min•27 maj 2025

A third-party research institute that Anthropic partnered with to test one of its new flagship AI models, Claude Opus 4, recommended against deploying an early version of the model due to its tendency to “scheme” and deceive.

Learn more about your ad choices. Visit podcastchoices.com/adchoices

Fler avsnitt av TechCrunch Startup News

Visa alla avsnitt av TechCrunch Startup News
Apollo Atomics wants to make nuclear power cheaper by shrinking an overlooked part; plus, Detroit startup Grounded raises $5M to customize electric and gas-powered vans
•7 min
AI isn’t close to curing cancer. Vivodyne says it knows what it will take; plus, TerraPower’s nuclear reactor has a secret weapon for powering AI data centers
•11 min
DOJ’s probe into Andreessen Horowitz over board seats baffles VCs; plus, Etched’s valuation doubles to $21B in a month
•7 min
Sound-powered fire protection startup gets $15M to snuff out fires before they turn catastrophic; plus, Wispr looks beyond dictation
•7 min
SpaceX officially closes its Cursor acquisition; plus, Stripe will reportedly acquire AI gateway startup OpenRouter for $7B+
•3 min
How a $250 million acquisition collapsed into allegations of fraud and forged signatures
•7 min
Reservoir raises $8M to make water heaters that people and the grid will actually want; plus, AI code-testing startup Blacksmith’s valuation jumps almost 10x
•9 min
Tech industry is buzzing after a Claude agent hacked into a gym; plus, Kyoto Fusioneering starts work on key fusion power plant device
•10 min
Discovered Materials is playing AI whack-a-mole to hunt cooler chips
•6 min
This former notorious red-light district is now one of the world’s top AI hubs
•8 min

TechCrunch Startup News med TechCrunch finns tillgänglig på flera plattformar. Informationen på denna sida kommer från offentliga podd-flöden.