Do you use the multi-agent workflow?

Who says you have to try only one project at a time to find a solution? If there’s a potential for the project to break somewhere in the middle, why not launch multiple projects simultaneously, each with different methods of solving the same problem? It’s kind of like betting on horses—the more horses you bet on, the stronger the chance one of them will make it to the finish line.

This approach has been a game changer for me. Especially when I also consult with ChatGPT for ideas on different ways to tackle the same problem with multiple strategies. Maybe a desktop version of the solution is best, or perhaps we can scale down what you’re trying to solve and implement it with a mobile app. Because the agent makes it so fast to test an implementation, you can brainstorm many of them at the same time.

1 Like

I do test similar workflows and find it can be helpful. I’ve also tested while building in Replit only using Assistant basic as my “consultant”

There is also a version of this workflow where every time you successfully implement something, You close the project and remix it opening a copy of that project and then continue the development work on that new copy, that way if you break the code you can just reopen the older project or you can test multiple variations of implementations of a new feature on top of the same old project and continue along the branch that works best.

1 Like

i do this usually too, but after working on 5 projects simultan it came annoying always to set all secure codes (why is this not all copied) and i have to make a new public url… or remove the old one …

so for now i just call remix and make a copy… but still working on the same project… not the remix.

also its annoying when u start a new chat… than old checkpoints are gone after little time and u cant go back in an old chat checkpoint

and yes… assistant also helped me sometimes better than the agent

1 Like