Train your GPT on your historical taste so even lazy prompts yield great output
yungcontent · x · 2026-09-12
The poster argues that if you build up a custom GPT to the point where it has absorbed your "historical taste," you can throw even the worst prompts at it and still get great output — the model has internalized your style preferences.
More from Apps
- Oncologist runs complex cancer-of-unknown-primary cases through Open Evidence and ChatGPT, impressed — scottleibrand · 2026-09-13
- OpenAI shows off GPT-6 Astra community builds, from 3D anatomy to Unity air traffic sims — OpenAIDevs · 2026-09-13
- Android AI Is Useless Without Root: Sandboxes Lock Agents Out of Real Work — Embarrassed-Way-1350 · 2026-09-13
- One Grok prompt recovered ~$500 in unclaimed property across state databases — petergyang · 2026-09-13
- Using Astra to Animate PowerPoint Frame by Frame, One Slide per Frame — itsShnik · 2026-09-13
- OpenDisplay: free open-source Sidecar alternative turns iPhone/iPad into a Mac second display — tom_doerr · 2026-09-13