9 Comments
User's avatar
Tim Esler's avatar

Would be great to see how it goes with a single superworker, no organization at all

Kun Chen's avatar

yeah totally. I’d expect the difference to vary based on the complexity of the task. Solo will likely win until the project becomes _really_ big.

Bryan Wieger's avatar

Love it! Good reminder in there about TDD

Zimran Ahmed's avatar

This is very cool!

Mike McCune's avatar

amazing work Kun, this is fascinating. I attempted to guess which models would work well on the quality/speed/cost axis and was wrong just as much as I was right.

Richard Moore's avatar

This is great

Rohit Krishnan's avatar

This is really very cool work! Love seeing more simulation based analyses...

J Derek Wilson's avatar

I've been playing with something like this but where i try to map the agent org chart to the desired solution architecture. like - can we take Conway's law and use it as a design tool? or maybe just for an efficiency improvement by reducing impedance between org comm structure and target system design?

Eric's avatar

What’s interesting to me is that the org charts were meant to describe the organization of the company, while I suspect the functioning of individual teams across the companies is actually quite similar. In other words, what does this look like with 20,000-100,000 agents?