Why do LLMs always estimate task time like sequential human work?

ColleenMBrady · x · 2026-09-24

Developer Colleen Brady flags a recurring pain point when planning with OpenAI and Claude models: their time estimates assume human-style sequential work — one task, then the next — ignoring parallelization.

Her current workaround is essentially giving the model a pep talk, which she admits wastes time and tokens; she's asking the community for a better fix.

Original post →

More from coding & agent

coding & agent channel →