I listened to an interview of him talking about this. There was also a note about allowing people to use their own set up. Thanks for the link. I’ll check this out
Haha. Yes, 1, Yes, 1. I thought I moved out of the dumb zone when I learned to manage context windows to keep my agents out of the dumb zone.
Having moved away from my terminal and cursor I now find myself back in the dumb zone as I copy/paste feedback from one LLM to another. I'm now the bottleneck.
…i put your prompt in my a.i. and it sent me a pizza…that is a good prompt…
I will never steer you wrong
Your point about the terminal being the computer is spot on. It's a deliberate choice by Boris Cherny, who built Claude Code. He spent years doing type inference at Meta and his whole career has been about making powerful formal tools approachable to devs who just want to ship. Covered his background and the design philosophy here: https://reading.sh/who-is-boris-cherny-the-engineer-behind-claude-code-c2d9c6753f29?sk=139c376109a3af3db5110a3767d009c6
I listened to an interview of him talking about this. There was also a note about allowing people to use their own set up. Thanks for the link. I’ll check this out
Haha. Yes, 1, Yes, 1. I thought I moved out of the dumb zone when I learned to manage context windows to keep my agents out of the dumb zone.
Having moved away from my terminal and cursor I now find myself back in the dumb zone as I copy/paste feedback from one LLM to another. I'm now the bottleneck.
The force is strong with this one.