It’s been almost two years since GPTs were released, but it’s hard to say they’re being fully utilized by most users. I’d like to ask OpenAI to seriously consider a major update to GPTs.
Here are my requests. Builders, please feel free to add your own suggestions as well:
-
Country-specific rankings in the GPT Store (Explore GPTs):
Right now, rankings are based on global conversations. Many Japanese users have told me they don’t know which GPTs to try because they can’t read English. And of course, this isn’t unique to Japan. There are many users around the world who don’t read English, so I believe GPTs need stronger localization support. -
Looser system prompt character limits:
Since the December 2023 release, system prompts for GPTs have been capped at 8,000 characters. On social media, I often see people saying that AI resource consumption keeps dropping. If that’s really the case, I’d love to see the limit raised. For example, 16,000 or even 24,000 characters might be a good starting point. -
Personal pages in the GPT Store:
Some users specifically want to access GPTs I’ve built, so I currently publish a profile page on an external AI portal. My guess is that when GPTs were first released, no one expected Builders to have “fans.” But in reality, who built a GPT is a crucial factor. I imagine even enterprise users would want company-branded Builder pages. I hope OpenAI considers adding features that showcase Builders more directly. -
Bookmarking for GPTs:
Users need a way to collect their favorite GPTs. Simply pinning them to the sidebar isn’t enough, because it doesn’t allow for organization or proper management.
To be frank, the current GPTs offer a terrible UX for general users: they’re hard to understand, hard to discover, hard to choose, and hard to manage. But GPTs are also one of the main offerings inside ChatGPT Enterprise. If more everyday users could find and effectively use great GPTs, I believe more companies would adopt ChatGPT Enterprise.
Please consider updating GPTs. I believe it would make a real difference.