New Apple study challenges whether AI models truly “reason” through problems

Ars Technica
By Benj EdwardsJun 11, 2025, 5:56 pm339 pts
In early June, Apple researchers released a study suggesting that simulated reasoning (SR) models, such as OpenAI's o1 and o3, DeepSeek-R1, and Claude 3.7 Sonnet Thinking, produce outputs consistent with pattern-matching from training data when faced with novel problems requiring systematic thinking. The…

Read Article Share
Share Article
- email
- twitter
- facebook
- pocket
- reddit
- tumblr
- linkedin
- pinterest

Trending Today on MacHash

ASUS ProArt 6K Display for Mac users will be available in August, at a compelling price 280

Indie App Spotlight: ‘Nunch’ is an elegantly designed calorie and macro tracker 238

Some Early Tests and Notes on ChatGPT Agent 232

Apple to Upgrade iPhone 17 Displays in These Two Ways 181

Amazon slashes up to $300 off M4 iPad Pros this weekend 179

UK to target Apple, Google ahead of Trump visit 170

Corning's EU compliance changes won't affect iPhones, iPads 170

Puck: ‘Was Colbert’s Cancellation Really “Economic” for CBS?’ 160

About MacHash

MacHash is your real-time Apple news aggregator, delivering the latest headlines on Apple, Mac, iPhone, iPad, and iOS from top sources across the web.

As a powerful content discovery platform, MacHash continuously curates breaking news, product announcements, software updates, reviews, and industry insights related to Apple Inc. and its ecosystem.

MacHash helps you stay informed on everything from macOS and iOS developments to Apple Watch, AirPods, and the latest in tech and app innovation.

Access MacHash from your desktop or mobile device to explore, follow, and share the most trusted Apple news all in one place.