DeepSeek's staff mostly work on data; Liang Wenfeng labels data himself, LatePost reports
teortaxesTex · x · 2026-11-06
According to a new LatePost piece on China's AI data industry: the vast majority of DeepSeek employees work on data-related tasks, and founder Liang Wenfeng occasionally takes part in data labeling himself. After taking over Hunyuan, Yao Shunyu re-cleaned existing data and improved infrastructure, helping Hunyuan 3 Preview significantly outperform its predecessor; the official release further refined post-training data and scaled up RL for additional gains. The report also reveals gray-market practices: startups pool accounts with model providers, resell API access, and sell the resulting data. The quoting poster argues that as models grow more capable, each training round can begin sooner — the data flywheel accelerates.
More from Companies & People
- Free Cryptography and ML Winter School in Milan opens registration for Jan 2027 — andrew_ilyas · 2026-11-06
- Trae Stephens on going from Palantir to Founders Fund to Anduril — adityaag · 2026-11-06
- Founders must hide their fear while financiers flex vulnerability — katieruthmishra · 2026-11-06
- Lambda CEO is extremely bullish on neoclouds as compute demand compounds — gharik · 2026-11-06
- Parallax founding team assembles in UK to build white-box belief auditing for AI evaluators — gsarti_ · 2026-11-06
- OpenAI Sets Nov 12 'Cyber Day' With Livestream and Security Announcements — gdb · 2026-11-06