Viewpoint
Ablation shows model is smarter without prompts, but product prompts help usage.
Ablation shows model is smarter without prompts, but product prompts help usage.
We use ablation to evaluate prompts, finding the model is slightly more intelligent without prompts, but product prompts help usage.
- Interview
- Boris Cherny: Building Claude Code
- Speaker
- Boris Cherny
- Source timestamp
- 4:32
More from this interview
- Opus 5 runs for weeks or months with Auto Mode and isn't prompt injectable.
- Deleted 80% of system prompt because Opus 5 is more intelligent.
- Defines product overhang and hobbling in AI products.
- Claude Code emerged from Sonnet 3.5's untapped file-writing capability.
- Give high-level tasks with guardrails and exit criteria to modern models.
- Giving models hard tasks with verification is the key skill.
- Simple prompt to rewrite Electron app in Swift runs over two weeks.
- Dynamic workflows let Claude orchestrate thousands of agents efficiently.
- Routines let Claude maintain its own codebases and automate maintenance.
- Coding is solved for some tasks, not deep systems or UI verification.
- Practical application and business sense should complement computer science theory.