OpenAI had a big day yesterday and announced a host of new and improved functionality.
Which of these features are you most excited to see implemented (either in an upgraded OpenAI integration or natively as part of Glide AI)?
Please be specific with the functionality and what use-cases you envision it helping with. We’re excited about the possibilieis and want to hear what you think!
“Ask a custom GPT a question” action. Zapper have something similar as part of their chatGPT integration
Asking custom GPTs would be unreal. Even just a modification of the ask question where we can specify some more parameters like “enable json” etc. would be huge!
“Analyse image” using GPT Vision
JSON mode is one of the things that I look forward to.
What about the retrieval API? If Glide implements something like that, would the stored files be on developers’ side or your side? If it’s on your side, I would think that means extra cost, since it charges per day.
Yes 
JSON mode and TTS would be amazing.
For any/all of these to be added to the OpenAI integration , please don’t just add to GlideAI
TTS is awesome , would love to see it.
Basically add them ALL 
Generate an image based on a prompt. Would really love this.
Almost forgot about DALL-E-3 , seems an obvious one to add also
ChatGPT just added image tools, including DALLE. It is quite unbelievable. This was my first experiment using a photo from the weekend.
I think this was new to the app? (including Code Interpreter)
DALL-E has been available for web for a weeks I think.
Yes that’s right. I have a hunch the GPT App version has some extra AI that allows the prompt to be a little more casual, e.g. I’m going to upload a photo of something and I would like you to do this… etc…
Honestly I could imagine this would be a game changer feature for Glide if we could integrate. Being able to work on an uploaded file with an ‘on submit’ action would be super interesting.
You can do it on the web version as well.
Knowledge Based Assistant for Private Corporation。
In his keynote address at the Nocode Summit in Paris, Glide’s CEO David explained there are two ways Glide can leverage AI:
Right now, the power of AI is put in the hands of the app developer with Glide AI and the OpenAI integration.
David did say that eventually they’ll allow the developer to build with AI.
As a sidenote, I don’t think you need a Glide expert to build apps on Glide. The whole point of Glide is the put app-building capabilities in the hands of as many as possible, not only tech savvy people.
I can’t remember where, but i remember you told you need to trig webhook
Actually, for now, you can do it from your classic app
I haven’t gone through Glide’s Documentation from A to Z in a long time. Your statement made me curious and I will take the time to do exactly this.
People learn differently. Through text, audio, video. Academically by going through structured learning. By experimenting and being hands on. I find Glide has done a good job at mixing up where and how people can learn: Glide Docs, the community forum, recently Glide University. Some of the experts, notably Bob, Darren A., Marco and a few others have created valuable tutorials available on Youtube. Not only once but on a regular basis.
I have seen a few people complain about there being too much text and not enough video. Or too much video and not enough text. Or the videos are too quick. And so on. I think mostly the material is there. Videos can be slowed down or speeded up at will. One can take the time to read and re-read as many times as one wishes. One can read the manuals and build in parallel. At the end of the day, it’s about putting in the work to learn.
Most importantly, I personally think that anything that needs a manual is not well designed. This is my personal view and developers by trade here will probably disagree with this because documentation and manuals are such a part of dev culture. But I insist, if something needs a manual, it’s already broken. Especially in the context of Glide. It’s ironic I say this since I greatly benefited from Glide Docs and still do. That’s my rigid and academic side, it helps me structure my knowledge. In my view, Glide is so intuitive that it doesn’t need a manual. Anytime one does need a manual in Glide, it probably means the person really didn’t try, Glide didn’t yet design the feature correctly or the developer shouldn’t be attempting their setup in the first place.
The very best experts push the limits of Glide and enjoy developing on Glide things that seem impossible. They are the best Glide experts therefore they push the limits of the platform, and they push the limits of the platform therefore they are the best Glide experts. They get great joy out of pushing the limits, it helps them build their expertise and it might serve a few other creators in the forum and on Youtube. But usually, these setups, which I would call hacks, shouldn’t be used by most Glide developers and shouldn’t become part of anyone’s toolbox.
Being a Glide expert is not at all a job. It’s just a recognition by Glide that the person has more than basic expertise using Glide. Many users of Glide have extraordinary expertise in Glide and yet aren’t part of the experts program. And among experts, Glide expertise can differ greatly.
I don’t think anyone is creating video for the incentives offered by Glide. It takes a tremendous amount of work and effort to create content. Experts have their reasons for creating content on Youtube, in Glide University and helping people in the community forum, and I doubt the worry of being “replaced” crosses anyone’s mind. I could be wrong of course. I’ve interacted a little online with the best Glide experts, and from my experience they all seem to be part of a select club of people who have a capacity for goodness, generosity and trustworthiness.
All of this being said, of course things could be better, as they often can. Glide works hard to provide the best software builder they can, all the while providing a good learning experience and community to get people on board. Everyone is doing the best they can with the resources they have.
I have a simple collection with a button that calls the action to create a “blue house”.
Trying to use the Generate Image action to create the blue house but only get “Generate image: could not fetch”.
I set the API key and I did give OpenAI my credit card # 
When I run it in the editor or from the App I get the same “could not fetch” result.
