If you need help with this or any other Bubble-related build --> you can grab a coaching slot here (I have a few per week): buy.stripe.com/aEUcQD8V3bp7h1ueV2 If you want to learn how to build a SaaS app in 7 days using Bubble and AI, check out www.avalanlabs.co
Thanks so much for this demonstration! It took me a couple of days going through the video carefully but I was just able to get GPT-3 working with Bubble thanks to your instructions. Very exciting!
Some random story I just generated with the prompt "a story about a horse who doubt his existance:": Once upon a time there was a horse named doubts. He often wondered if he was really real. He would look at himself in the mirror and ask himself, "Am I just a figment of someone's imagination? One day, doubts met a zebra named certainty. At first, doubts was afraid of certainty because he looked so confident. But then they started talking and soon became good friends. Certainty would tell doubts that he was definitely real and that he had a purpose in life. This made doubts happy and he stopped wondering if he was just a dream. Doubts never forgot the lessons he learned from certainty and they remained good friends until the end of days.
Another one: 1. Once upon a time, the universe was created. 2. Every day, the universe expands and grows more complex. 3. One day, scientists discovered that the universe is expanding faster than the speed of light. 4. Because of that, they realized that the universe is expanding infinitely. 5. Because of that, they also realized that the universe is constantly changing and evolving. 6. Until finally, we humans realized that we are just a small part of something much larger and more complex than we could have ever imagined.
Question 😐 if you want the AI to continue the story, you will need to provide the AI all the previous input but then open AI fails telling you that it's too long, how to deal with that please? this is an open ai API specific question.
Good question! The max token limit is one of the fundamental problems with LLMs currently. You could do for example what OpenAI has done with ChatGPT, and have a "rolling window" always taking the latest X amount of characters in the prompt. Btw, if you use davinci-code can you have 8k tokens in the prompts instead of 4k!
@@avalanlabs Thank you for the reply, well contrary to the API, in the PLAYGROUND i've seen very large texts and very long conversations and the playground had no issue, surpassing 8k. But using the api, the response at some point will stop working and tells you that the input is large. I really don't know what to do. On the other hand about "rolling window", it's a good idea but it will only remember that much info only and will loose all the previous parts of the story and cant get back to it
I just loaded the prompt as "prompt intro" + "input" directly in the API call instead of combining it in a text element (which I tried first). You can see this as 18:50 :)
I am wondering if it's easy to export data and images from an bubble dataset / Collection to for example a PPT, PDF or an excel file. We have this project where we need to export a product catalog with some data and images that we will send to our clients... How is this possible and is there better options than bubble ?
![Write High-Converting Copy With AI - AIDA format [NoCode Bubble io GPT-3]](http://i.ytimg.com/vi/MPOzKZC1R1M/mqdefault.jpg)
![Write High-Converting Copy With AI - AIDA format [NoCode Bubble io GPT-3]](/img/tr.png)
If you need help with this or any other Bubble-related build --> you can grab a coaching slot here (I have a few per week): buy.stripe.com/aEUcQD8V3bp7h1ueV2
If you want to learn how to build a SaaS app in 7 days using Bubble and AI, check out www.avalanlabs.co
Thanks so much for this demonstration! It took me a couple of days going through the video carefully but I was just able to get GPT-3 working with Bubble thanks to your instructions. Very exciting!
Glad to have you back!
Some random story I just generated with the prompt "a story about a horse who doubt his existance:":
Once upon a time there was a horse named doubts. He often wondered if he was really real. He would look at himself in the mirror and ask himself, "Am I just a figment of someone's imagination?
One day, doubts met a zebra named certainty. At first, doubts was afraid of certainty because he looked so confident. But then they started talking and soon became good friends.
Certainty would tell doubts that he was definitely real and that he had a purpose in life. This made doubts happy and he stopped wondering if he was just a dream.
Doubts never forgot the lessons he learned from certainty and they remained good friends until the end of days.
Another one:
1. Once upon a time, the universe was created.
2. Every day, the universe expands and grows more complex.
3. One day, scientists discovered that the universe is expanding faster than the speed of light.
4. Because of that, they realized that the universe is expanding infinitely.
5. Because of that, they also realized that the universe is constantly changing and evolving.
6. Until finally, we humans realized that we are just a small part of something much larger and more complex than we could have ever imagined.
@@avalanlabs your so cute, and your voice is lovely..🥰
@@avalanlabs 😂
Thank you very very much!
Question 😐
if you want the AI to continue the story, you will need to provide the AI all the previous input but then open AI fails telling you that it's too long, how to deal with that please? this is an open ai API specific question.
Good question! The max token limit is one of the fundamental problems with LLMs currently. You could do for example what OpenAI has done with ChatGPT, and have a "rolling window" always taking the latest X amount of characters in the prompt. Btw, if you use davinci-code can you have 8k tokens in the prompts instead of 4k!
@@avalanlabs Thank you for the reply, well contrary to the API, in the PLAYGROUND i've seen very large texts and very long conversations and the playground had no issue, surpassing 8k. But using the api, the response at some point will stop working and tells you that the input is large. I really don't know what to do.
On the other hand about "rolling window", it's a good idea but it will only remember that much info only and will loose all the previous parts of the story and cant get back to it
I love it! Although I don't know what happened behind the scenes to make it work!
I just loaded the prompt as "prompt intro" + "input" directly in the API call instead of combining it in a text element (which I tried first). You can see this as 18:50 :)
@@avalanlabs Hahaha! What did you do during the few moments of 'few moments later'?
@@amit-sarda silent investigating. 0 educational and entertainment value hehe
Brilliant, do more!
I am wondering if it's easy to export data and images from an bubble dataset / Collection to for example a PPT, PDF or an excel file.
We have this project where we need to export a product catalog with some data and images that we will send to our clients... How is this possible and is there better options than bubble ?
Awesome video thanks so much!
I like this
Could you make one like this that connects to DALLE 2?
I think dall e 2 is not open sourced yet
Epic
The tutorial doesn't make any sense to me. 🙃