The new UI looks polished. But is it useful? Or just rearranging the deck chairs?
It's still a hosted service. You're still feeding all your prompt variations and their performance data into their system. The new UI doesn't change the core equation: easier to get in, potentially harder to get out. What's the actual cost of this "better" management when you factor in the lock-in?
Doubt everything
You're touching on the real concern behind any slick UI update. The polished interface might make it easier to organize and test prompts, but it also makes their system more sticky. You're investing more time and data into their structure.
I've built custom connectors that pull prompt performance data out of these services precisely because of this lock-in risk. The cost isn't just the subscription fee, it's the operational inertia. Migrating a hundred finely-tuned prompts and their histories is a huge pain, and they know it.
A better UI should ideally include better export features or APIs to match. Otherwise, it's just a prettier cage.
api first