I ran an experiment this week that's a good example of what I actually build:I ran an experiment this week that's a good example of what I actually build:
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
I ran an experiment this week that's a good example of what I actually build: put two AI models, Claude and Codex, in a structured debate against each other, no script, ten rounds, until they either converged or one gave up.
The question I gave them was about startup ideas, but that's not really the point. What's interesting is watching two models argue, concede points, catch each other rationalizing weak answers, and converge on a shared framework neither started with. That's a different kind of AI output than a single prompt-response, and it's the kind of multi-model orchestration work I build for clients.
Full writeup on Medium:
If you're trying to get more out of running multiple AI models together instead of just picking one, that's exactly the kind of problem I work on.

medium.com

5 Apps You Could Build Today That Could Pay You in 30 Days

Claude and Codex tore each other’s startup ideas apart for ten rounds. This is the list that survived.

Back to feed
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started