Notes On Work - by Caleb Porzio Everything I feel about AI right now
Feb 3, 2026
Thoughtful reflections on whether AI truly speeds up developer work. Skepticism about flashy parallel agent demos and the real costs they hide. Stories of an LLM fixing a tricky bug and times AI produced messy, hard-to-maintain code. A call to plan and scope work more rigorously to get reliable AI outputs. Imaginative thinking about always-on voice assistants that autonomously manage repos.
AI Snips
Chapters
Transcript
The Pressure Of AI Hype
- Caleb Porzio feels an urgency and panic from AI hype that he isn't matching in speed or usage.
- The loudest success stories on Twitter create pressure that may not reflect real, sustained productivity gains.
A Day Spent Rigging AI Workflows
- Caleb spent days building Ralph loops and PRD files and still felt distracted and unaccomplished at day's end.
- He wonders whether doing the tasks himself might have been faster and more satisfying.
LLMs Win At Hard Debugging
- LLMs can solve gnarly debugging problems much faster than Caleb could alone.
- That speed is a clear, tangible productivity win when the model surfaces the right root cause quickly.