Tanay Naik (@_tn0123) profile picture
Joined October 2025
Agentic engineering is a supervised learning problem. Code = weights Your vision = training data Prompt = loss (approximated, you compute the real loss in your head, the prompt is your best attempt to articulate it) Agent = weight updater Product = model
35
What if your AI coding agents could get on a call and code together? I had that thought a month ago, so I built CallCode: callcode.dev/ Looking for 3-5 multi-agent power users to try it on real projects.
1
1
68