The Unbearable Cheapness of Open Weight Models
(news.ycombinator.com)
31.
32.
GPT-5.5 hallucinates 3x more than MIT-licensed GLM-5.2
(news.ycombinator.com)
33.
DeepSeek V4 Pro beats GPT-5.5 Pro on precision
(news.ycombinator.com)
34.
Bringing Up DeepSeek-V4-Flash on AMD MI300X
(news.ycombinator.com)
35.
37.
KV Sharing, MHC, and Compressed Attention
(news.ycombinator.com)
38.
DeepSeek-V4-Flash means LLM steering is interesting again
(news.ycombinator.com)
39.
A Few Words on DS4
(news.ycombinator.com)
40.
41.
DeepClaude – Claude Code agent loop with DeepSeek V4 Pro
(news.ycombinator.com)
42.
DeepClaude – Claude Code agent loop with DeepSeek V4 Pro, 17x cheaper
(news.ycombinator.com)
43.
DeepSeek V4–almost on the frontier, a fraction of the price
(news.ycombinator.com)
45.
DeepSeek-V4 on Day 0: From Fast Inference to Verified RL with SGLang and Miles
(news.ycombinator.com)
46.
Three reasons why DeepSeek’s new model matters
(technologyreview.com)
47.
49.
DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence
(news.ycombinator.com)
50.
DeepSeek v4
(news.ycombinator.com)
Today's top topics:
openai
anthropic
apple
ai safety
google
ios 27
dario amodei
iphone 18 pro
nvidia
microsoft