Why AI Swarms Cannot Build Architecture

> You can’t instruct individual agents to contribute to global knowledge

Of course you can. What makes you think you can’t? They can use task trackers like anyone.

Also you can have review on prs, stages where you ask for review and then have them create tickets for improvements (we’re using two http crates, refactor to just use this one).

this is well written and thought out

there's also a part 2: https://jsulmont.github.io/swarms-ai/part2