




we might be responsible 😬
fixed your slide @nvidia

can confirm we're pretty efficient
Why use many token when few do trick? @UseCaveman started as a joke to get my AI bill under my grocery bill and it just passed Gemini CLI in GitHub stars. Today it becomes a coding agent. Caveman Code is out now. Same model, a day of work in connected apps: Claude Code $7.65 Caveman Code $1.83

we built caveman to stop AI from making unnecessarily expensive decisions jensen, we may need to expand into humans

Caveman (YC F26) passed @supabase in GitHub stars today. Oh, and we train models now? trim-1: a 1.2 MB model that reads a long document and keeps only the lines your agent needs. Every kept line verbatim. No GPU, about 7 ms for 5,000 tokens on one CPU core. Lighter, Leaner, Meaner. More coming.

40% higher worldwide Google search interest in “caveman”. six months after launch. i just wanted AI to stop yapping and somehow became CMO of the stone age

My coding agent wrote the same script 394 times in one week. Same heredoc. 32 sessions. Write it, run it, throw it away, write it again. So we made Blocks. Agent carve script once, keep it in your repo, check it every PR. Next agent just call it. Why many script when few do trick. github.com/caveman-ai/blocks

Why use many token when few do trick? Caveman is backed by @ycombinator (F26). It started as a joke to get my AI bill under my grocery bill by having Claude speak like a Caveman. 111k GitHub stars later, it's a company. Today: Caveman 4.0 and Caveman Cloud.

6000 downloads in under a week. caveman happy, caveman tribe strong

I just became the youngest person to hit 82k GitHub stars...

v1.9.1 of Caveman is out. Hermes native support.