most attempts to let jev speak give very limited vocab and jev ends up speaking like…
This is a AI post classified by Jev as AI infra & evals (a tool drop), kept by the AI Radar because it carries real work, not commentary.
most attempts to let jev speak give very limited vocab and jev ends up speaking like caveman. to improve on this i added 50k word vocabulary using vocab sharding, hierarchical routing, and beam search to prune weakest branches. chat with jev here: https://danprice.ai/jev-speaks
Posted by gooby_esq (1.3k followers) 1 h ago · 3 likes · 75 views · view the original post on X. Kept by the AI Radar as AI infra & evals. Tools mentioned: Hacker News, re-ranked.
More AI work like this
- How to setup Jev from @typesafeai on @CloudflareDev AI Gateway in 60 seconds. — @CloudflareDev
- Seven cents per million output tokens. DeepSeek charges $0.66 for the same model. — @merge_api
- Q: What should you do when your "gold" eval dataset becomes stale? — @HamelHusain
- SpaceXAI Console just got a lot more powerful — @XFreeze
- June 2024, I was writing about evals in Lenny's Newsletter before it was cool. — @hammer_mt
- Clay evaluates every AI agent on three things: — @LangChain
- 1/ The OpenRouter Batch API is live. — @OpenRouter
- TIP: automatically classify your LLM requests using @typesafeai's Jev, using… — @OpenRouter
Every post is read and classified by Jev (TypeSafe): what it is, which market it belongs to, and whether the link is a real tool. 42.4k posts from 5k X accounts over the last 14 days, 1.8k tools, 19 markets. Collected every 5 minutes, fully re-ranked every hour — last update 2026-09-22 17:33 UTC. Full method.