Agent Architecture: Does Intelligence Live in the Model or the Harness?
lateinteraction · x · 2026-08-06
The author argues that in current AI systems, including typical coding agents, almost all intelligence resides inside the neural network, even though much of the capability relies on external tools and harnesses.
However, for Reinforcement Learning Models (RLMs) and beyond, the dynamic might shift. The author suggests that increasingly, intelligence will begin to live within the inductive biases of the external harness. Eventually, this learned system could make the harness just as fundamental to the architecture as activation functions inside a DNN.
More from coding & agent
- Claude Code Beats Codex in Long Tasks and Automations, Dev Reports — carlosdponx · 2026-08-07
- Hands-On: Claude Code's Automations Outperform OpenAI Codex — carlosdponx · 2026-08-07
- Podcast Explores Nous Research Hermes Agent: Executives Should Streamline Repeated Tasks First — msg · 2026-08-07
- Podcast: Automating Executive Workflows with Local AI Agents — msg · 2026-08-07
- Agent Memory Leaderboard Open-Sources Evaluation Code Across Multiple Datasets — AIwithGhotai · 2026-08-07
- Multi-Model Orchestration: GPT-5.6 Paired with Opus Cuts Costs by 46% — kevinkern · 2026-08-07