Since agents were released, I’ve seen some improvement in results, but lately they’ve just been eating through credits and not fixing obviously broken code. The chat also stopped allowing me to upload files:
It would be great if the V0 team could look at my account and debug what is going on and also give me some extra credits to reimburse the crazy spends.
If you still need support, could you share your chat link with us? Unlisted, please?
Sure:
https://v0.app/chat/translator-toolkit-7CTBkI4v3rA
https://v0.app/chat/fork-of-translator-toolkit-9VNq8bav1nw
https://v0.app/chat/duplicate-of-fork-of-translator-toolkit-3bpBj0om5GV
https://v0.app/chat/legacy-ai-translation-checker-2sSQuathDbi
https://v0.app/chat/fork-of-alpaca-portfolio-dashboard-BrkxrK63BrZ
Thanks for sharing these! I’ll pass it onto the v0 team. Appreciate the patience as we continue to improve the product.
I highly agree with this. Honestly, my development on my app has come to a screeching halt since agents were implemented. The old sized “machines” were so phenomenal because you could do controlled incremental changes and intervene when something goes wrong It was like you can actually work on the app together with the AI. Now with agents, the agent just goes on a huge sprint, more often than not in the wrong direction quite early on and there’s just no intervention. It’ll expend so much credits and will more than likely have to revert to a previous version. It also seems to have over confidence with reviewing its own work- it always misses really obvious things that are wrong and just because it’s not returning an error. I can see how agents would be great with starting up a project- absolutely. But when it comes to bigger complex projects that require more careful incremental changes, I just feel agents are very unsuiting. I hope the Vercel team considers having agents optional at some point.
