Prompt, context, harness, loop engineering are all just "knowing what you want"
mr_samosaman · x · 2026-11-15
@kaseyklimes offers a deflating one-liner taxonomy of the agent engineering buzzword wave: prompt engineering is just knowing what you want, context engineering is knowing what's relevant, harness engineering is knowing what to verify, loop engineering is knowing when you're done — and goal engineering circles right back to knowing what you want. A widely shared distillation of what all these "engineerings" actually boil down to.
More from coding & agent
- Practical quicksort needs 3-way partitioning and recursion-depth detection — srchvrs · 2026-11-15
- Proof of Agent goes viral: paste a prompt and let your AI agent mine POA — adamamcbride · 2026-11-15
- Google and Northwestern's Harness² self-improves agent harnesses without answer keys — rohanpaul_ai · 2026-11-15
- Dev builds a PowerShell AI harness for Microsoft's new Intelligent Terminal — dfinke · 2026-11-15
- snitch: a harness that watches agents in parallel and reverts bad paths — generativist · 2026-11-15
- Dev: unattended agent runs burn 99% of tokens on useless reproducibility gates — generativist · 2026-11-15