ML engineer asks: how to stop being the AI's project manager and QA engineer
Conscious-Shake391 · reddit · 2026-10-20
An ML engineer on Reddit vented that his biggest frustration isn't model intelligence but managing the AI's thinking — rewriting prompts, reminding it to research and compare alternatives, pointing out flaws, and repeating instructions already in Custom Instructions and Memory. He wants: state a high-level goal, and the AI autonomously clarifies intent, researches, plans, delegates to subagents, executes, critiques, and delivers a verified result. He asks whether this is a reasoning or orchestration problem, whether Hermes Agent/OpenClaw/Claude Code/Codex solve it, and whether Skills, memory, and automated evals reduce repeated corrections. No answers yet, but the framing hits a real pain point in current agent workflows.
Related event: ML Engineer Asks How to Stop Managing AI Like a Project Manager(2 posts)→
More from coding & agent
- NVIDIA paper predicts post-training coding-agent performance from base models with ρ=0.964 — sudoraohacker · 2026-10-20
- OpenAI benchmark: GPT-6.1 Sol hits 97% accuracy migrating COBOL at $0.50 per trial — ChrisGPT · 2026-10-20
- Matt Pocock: my self-built personal wiki became critical business infrastructure in 3 months — _AustinCalvert_ · 2026-10-20
- 99% don't know 'record a skill': computer use agents' ceiling is nowhere near — alliekmiller · 2026-10-20
- Swapping to Qwen when GPT refuses your prompt is like typing sudo — DanGrover · 2026-10-20
- Letting an AI agent 'live a life' in a 30-year-old MMORPG: a Reddit experiment proposal — Aeylwar · 2026-10-20