Remember when we loaded system prompts with death threats to get LLMs to write React?
granawkins · x · 2026-08-25
X user granawkins reminisces about the wild early days of 2023/24, when practitioners would stuff their system prompts with death threats and fantastical hostage narratives just to make LLMs write React components more reliably.
The nostalgia post captures how far prompt engineering has come — from literally scaring models into compliance to today's structured instructions and tooling — a bit of shared industry lore from the frontier's early chaos.
More from Fun
- Humanoid robot sees cheerleaders, instantly forgets how to climb stairs — basedjensen · 2026-08-25
- One person built a 3D fishing game in 4 weeks with Claude Code and Godot for ~$300 — jocarrasqueira · 2026-08-25
- Unitree Robot Gets Lost During 400m Obstacle Race at Humanoid Games — saibharadwaj · 2026-08-25
- Anthropic employee reportedly stopped ADHD meds to better manage AI agents — MatthewBerman · 2026-08-25
- DeepMind Scientist Uses $10 Clamp to Prove AI, Not Hardware, Is the Robot Bottleneck — cen6wkf · 2026-08-25
- Claude Thinks for 27 Minutes, Then Hits Session Limit — damnGruz · 2026-08-25