Teknium Questions Why OpenAI Trains Models for Only One Specific Harness
Teknium · x · 2026-09-07
Teknium pushes back on OpenAI's harness-specific training, asking why models would be trained for a single workflow when orchestration patterns are in flux. He details Hermes' setup: subagents can't talk to each other, only submit final work to an orchestrator, with fresh context and skill references supplied per subagent.
Related event: Teknium Details Hermes Sub-Agent Design and Questions OpenAI's Harness Ties(3 posts)→
More from coding & agent
- Kolab Benchmark Uses Pro Screen Recordings to Show AI Agents Overestimate Collaboration Quality — minnesotanlp · 2026-10-27
- Dev argues Claude burns far more tokens than GPT models for the same coding work — haider1 · 2026-10-27
- Stop building agents that can't change models: stable contracts, external state, real evals — rseroter · 2026-10-27
- Cheap models can cost more in agent harnesses than frontier ones, researcher warns — vasuman · 2026-10-27
- Kanzo: A 5k-Line Rust Kanban Binary With MCP for Claude Code and Codex — allmudi · 2026-10-27
- 'A human approves everything' is a trap: approval fatigue and how to actually design safe agents — Efistoffeles · 2026-10-27