Deep Research took 682 minutes and 54 corrections for one question: is it broken?
big-fox- · reddit · 2026-10-20
A Reddit user details the pain of using Deep Research on hard questions: it nails the easy parts then stops halfway through the brief. One question (whether an Ethereum builder strategy could make durable money) took 682 minutes and 54 messages, mostly spent demanding better sources and fixing omissions. The output was fine; getting there was "a part-time job." The author also built a harness around Deep Research to do the babysitting and is soliciting test questions, with a self-promotional angle.
More from coding & agent
- One API Call to a Full Agentic Sandbox: Trying Gemini 4 Argon via Managed Agents — _philschmid · 2026-10-20
- AI Harness: Code Isn't Scarce Anymore — Human Attention Is, Argues Microsoft AI PM — leslysandra · 2026-10-20
- OpenAI Launches ChatGPT Team Tasks for Automating Recurring Engineering Work — OpenAIDevs · 2026-10-20
- OpenAI launches team tasks in ChatGPT to automate recurring engineering updates — OpenAIDevs · 2026-10-20
- LukeW Details His Personal AI Pipeline: Hourly Auto-Indexing of Audio, Video and PDFs — LukeW · 2026-10-20
- Delhi's DevDay Exchange community hackathon set for Nov 1, a full-day Codex build sprint — gabrielchua · 2026-10-20