@george_nqu I get people telling me that we developers use claude code to build us everything while we do nothing. I was adding to what you are saying about people thinking ai will magically solve our problems.
-Noticeably faster overall
-the diagrams it creates for logic makes decision making much better for complex logic
- edits are far less likely now to produce errors, so it’s much more reliable in build mode
-it can see the big picture of a page and can identify gaps. I feel it understands context much better than before.
2 small enhancements for future updates:
-sometimes it makes a workflow with “this element is clicked” it have a problem with referencing the right element/button even when i give it screenshots.
- i think it does not account for race conditions as it always prefers to make create/update workflows in steps not with different workflows. I found myself always correcting it to avoid race conditions like when a button can both create and update data.
Buildprint can now test your @bubble app right in the browser.
This is rolling out to everyone now, and allows Buildprint to identify issues, test your work, and test its own work.
Here, it identifies a real issue whilst testing out a new flow. I can then ask it to investigate and fix it.
Much more cool stuff coming soon.