Decagon Co-founders at a16z: 90% of Enterprise AI Agents Run on Open-Source
kimberlywtan · x · 2026-07-31
a16z partners hosted Decagon's co-founders to discuss lessons from deploying AI agents in large enterprises like banks and airlines.
Key insights include:
- Open-source dominance: 90% of Decagon's traffic runs on open-source models. They recommend starting with frontier models for new use cases, then migrating to open-source as they mature.
- False trade-off: Smart vs. cheap is a false dichotomy; fine-tuning at the application layer bridges the gap.
- Demand elasticity: Support demand always outpaced supply. Making it cheaper with AI simply means companies buy more capacity.
Related event: a16z x Decagon: Open-Source Migration and Moat in Enterprise AI Agents(6 posts)→
More from coding & agent
- 'LLM code is sloppy' rebutted with real ffmpeg code: identical if/else branches in a top library — kingslyj · 2026-10-16
- Agent actions need checks outside the model: NVIDIA demos lean on guardrails amid wave of real agent incidents — Creamy-And-Crowded · 2026-10-16
- Vitalik builds "paranoid mode" agent wrapper with Tor, zkAPI and local prompt rewriting — DavideCrapis · 2026-10-16
- Agent fine-tunes Qwen3.5-0.8B with synthetic data: 4.7%→76.3% accuracy, 11.9× faster — ben_burtenshaw · 2026-10-16
- Beam-501B agent demo: solo beats 5-agent team on ProgramBench rebuild task — ben_burtenshaw · 2026-10-16
- Beam agent builds BeamSpeak, a local Parakeet dictation app for macOS — ben_burtenshaw · 2026-10-16