DeepSeek open-sources inference optimizations with 60–85% faster generation [pdf]
(news.ycombinator.com)
1.
3.
4.
5.
DeepSeek Looks to Double Workforce as Fresh Funding Fuels Growth
(feeds.content.dowjones.io)
6.
The unbearable cheapness of open weight models
(news.ycombinator.com)
7.
The Unbearable Cheapness of Open Weight Models
(news.ycombinator.com)
8.
GPT-5.5 hallucinates 3x more than MIT-licensed GLM-5.2
(news.ycombinator.com)
9.
DeepSeek Introduces Vision
(news.ycombinator.com)
10.
11.
US holds off blacklisting DeepSeek, more than 100 firms deemed security risks
(news.ycombinator.com)
13.
DeepSeek Becomes China’s Most Valuable AI Startup After $7.4 Billion Fundraise
(feeds.content.dowjones.io)
14.
Applying Brevity and Language Efficiency in Prompt Engineering
(news.ycombinator.com)
15.
Open Reproduction of DeepSeek-R1
(news.ycombinator.com)
16.
Notes on DeepSeek
(news.ycombinator.com)
17.
18.
19.
Bringing Up DeepSeek-V4-Flash on AMD MI300X
(news.ycombinator.com)
20.
The mysterious Hy3 LLM is topping OpenRouter Model Rankings by a large margin
(news.ycombinator.com)
21.
22.
Math-to-Manim
(news.ycombinator.com)
23.
Outsourcing plus local AI will soon become more economical vs. frontier labs
(news.ycombinator.com)
24.
Innovation starts in schools — lessons from China
(feeds.nature.com)
25.
DeepSeek to Make Permanent 75% Discount on Flagship AI Model
(news.ycombinator.com)
26.
DeepSeek reasonix, DeepSeek native coding agent with high caching and low cost
(news.ycombinator.com)
28.
DeepSeek makes the V4 Pro price discount permanent
(news.ycombinator.com)
29.
KV Sharing, MHC, and Compressed Attention
(news.ycombinator.com)
30.
DeepSeek-V4-Flash means LLM steering is interesting again
(news.ycombinator.com)