[bisonbear]

Intuitively makes sense, but in my experience, a more realistic workflow is using the main agent to sub-agent delegation pattern instead of straight 7x-ing token costs.

By delegating to sub agents (eg for brainstorming or review), you can break out of local maxima while not using quite as many more tokens.

Additionally, when doing any sort of complex task, I do research -> plan -> implement -> review, clearing context after each stage. In that case, would I want to make 7x research docs, 7x plans, etc.? probably not. Instead, a more prudent use of tokens might be to have Claude do research+planning, and have Codex do a review of that plan prior to implementation.

[languid-photic]

Yes, understandable.

The question is which multi-agent architecture, hierarchical or competitive, yields the best results under some task/time/cost constraints.

In general, our sense is that competitive is better when you want breadth and uncorrelated solutions. Or when the failure modes across agents are unknown (which is always, right now, but may not be true forever).

[girvo]

> straight 7x-ing token costs

You are probably right, but my work pays for as many tokens as I want, which opens up a bunch of tactics that otherwise would be untenable.

I stick with sub-agent approaches outside of work for this reason though, which is more than fair a point

[darkerside]

Maybe an evolution based approach does make sense. 3x instead, and over time drop the least effective agents, replacing them with even random choices.

Edit: And this is why you should read the article before you post!

[languid-photic]

Yes indeed, you get a big lift out of running just the few top agents.

We run big ensembles because we are doing a lot of analysis over the system etc