Factory CTO Deep Dive: Building Coding Agents, Tokenomics, and Human-in-the-Loop
LangChain · x · 2026-07-28
The LangChain podcast hosted Eno Reyes, CTO of Factory (a $1.5B AI coding company), for a deep dive into building autonomous software factories.
Key takeaways from the interview:
- Harness over Models: Why a model-independent harness isn't always ideal, and how to build a universal meta-harness.
- Memory and Validation: Why 'memory' is AI's most overused word, and how Factory routes around agent failure modes (like 'panic' before cheating) using multi-layered validation.
- Tokenomics: How intelligent routing significantly cuts agent operational costs.
- Human-in-the-loop: Why product management isn't going away and why humans will remain essential for a long time.
More from coding & agent
- LightOn says its agent search hits 86.27% accuracy with just 9.7 calls — IgorCarron · 2026-07-28
- The Modern Wait Equation: Fix AI Slop Now or Wait for the Next Model? — ilanbigio · 2026-07-28
- Kimi K3 pairs a 2.8T MoE with 1M context and multi-teacher post-training — novasarc01 · 2026-07-28
- Webhound says agents stop too early, and its $5 demo found an unreported $18M change — ycombinator · 2026-07-28
- Building a Coding Agent from Scratch in Node.js with Zero Frameworks — dfinke · 2026-07-28
- PRO-LONG hits 97.4% on public ARC-AGI-3 for $1,750 with a log.txt memory harness — GregKamradt · 2026-07-28