Thomson Reuters bets $40M on Qwen-based proprietary AI model
The Decoder · rss · 2026-08-24
Thomson Reuters has launched "Thomson," its own language model built on Alibaba's Qwen, at a cost of about $40 million over two years. Benchmarks show top performance only when the model can tap into the company's proprietary content like Westlaw. CTO Joel Hron emphasizes that the key is not intelligence itself, but knowing which intelligence you need to own.
More from Companies & People
- AI's real test: turning expertise into a business for less, not bigger models — TansuYegen · 2026-08-24
- Whop Built an Enterprise Agent Integrating Slack, Data Warehouse, and Ad Accounts — eptwts · 2026-08-24
- Study: Frontier AI Labs Still Won't Disclose Plans to Contain Rogue Models — RebeccaBellan · 2026-08-24
- LangChain recruiting users to improve dcode coding agent — LangChain · 2026-08-24
- OpenAI President: Compute Gap Massive, H100 Prices Rise, AGI Will Eliminate Interfaces — AI寒武纪 · 2026-08-24
- AI didn't make design more competitive, just crowded at the bottom — oykun · 2026-08-24