[Thread] A new US paper shows the best frontier LLM models achieve 0% on hard real-life Programming Contest problems, domains where expert humans still excel (Rohan Paul/@rohanpaul_ai)

https://www.techmeme.com/feed.xml Hits: 7

Summary

Invest Like the Best: The leading destination to learn about business and investing. We do this by showcasing exceptional talent and ideas. Subscribe to Invest Like the Best

First seen: 2025-06-17 10:13

Last seen: 2025-06-17 16:14

Read Full Article More from this Source

[Thread] A new US paper shows the best frontier LLM models achieve 0% on hard real-life Programming Contest problems, domains where expert humans still excel (Rohan Paul/@rohanpaul_ai)

Summary

Related News

Threads launches a fediverse feed and a way to search for fediverse users, with support for Mastodon and others, for users who turned on fediverse sharing (David Pierce/The Verge)

Waymo expands its coverage area in San Francisco and LA by 80 square miles each, bringing its total coverage area in California to 250 square miles (Andrew J. Hawkins/The Verge)

Nielsen: in May, Americans watched more TV on streaming than on cable and broadcast networks combined, the first time that has happened over a full month (John Koblin/New York Times)

Meta launches generative AI tools in its Advantage+ suite for advertisers and agencies at Cannes, including to turn still images into videos and Virtual Try-On (Mark Mwachiro/Adweek)

Israeli observability startup Coralogix raised a $115M Series E led by NewView Capital at a $1B+ valuation, and plans to invest ~$100M in India over five years (Jagmeet Singh/TechCrunch)