How OpenAI Decided to Build ChatGPT

@technooptimist

17.6K followers · usually 1.1K views a post

Views

4.2K

Likes

Comments

SoundTikTok sound · original sound - technooptimist

Promoted product

OpenAI

OpenAI is a company that develops AI technology and offers an API for developers to interact with its models.

20 videos17 creators

Why it worked

The video explains the accidental discovery of a product-market fit for AI chatbots by observing user behavior and viral content, demonstrating the power of iterative development and responding to emergent use cases.

Summary

OpenAI initially struggled to find a product for GPT-3, releasing it as an API. When a user's tweet about using GPT-3 for copywriting went viral, it showed a clear demand for AI-powered chat applications. This led OpenAI to develop a better model, GPT-3.5, focusing on instruction following to create a chatbot, which eventually became ChatGPT.

Structure

1OpenAI develops GPT-3 but struggles to find a product.
2Releases GPT-3 as an API, initially with little traction.
3A viral tweet demonstrates GPT-3's potential for copywriting.
4

Original caption

ChatGPT is one of the fastest growing products in history. However, it's own creation was a complete accident for OpenAI. "We had made GPT-3 and we needed to make money because we wanted to go scale up to, you know, a billion and multi-billion dollar computers. And we had GPT-3 and it was kind of interesting. It was a cool demo, but we couldn't figure out a product to build around it. And so we knew the models were going to get better, but we also wanted to like start a revenue engine sooner. And we said, well, since we can't figure out what product to build, we're just going to put this into an API. And so we launched in like summer of 2020, the GPT-3 API. And initially it kind of got no traction at all. And then about a month later, randomly, as far as we can tell, it went viral on Twitter. And then like a lot of people started trying the API, but it was shockingly bad And really, the only business that people got to work in a significant way with GPT-3 was copywriting. And that was like not that great and not that exciting. And we said, well, we can build a good chatbot. People clearly want that. And we had a new model." —@OpenAI Co-Founder, Sam Altman at @Stanford #openai #samaltman #chatgpt #startup #techcompany

More from @technooptimist

On-screen text

How OpenAI Decided to Build ChatGPT We had made GPT-3 and we needed to make money because we wanted to go scale up to, a billion and multi-billion dollar computers. And we had GPT-3 and it was kind of interesting. It was a cool demo, but we couldn't figure out a product to build around it. And we had been thinking, we just couldn't do it. Sam Altman Co-Founder & CEO OpenAI We had tried a few things, they hadn't worked. And so we knew the models were can get better, but we also wanted to like start a revenue engine sooner. And we said, well, since we can't figure out what product to build, we're just gonna put this into an API. somebody else can figure out what product to build. OpenAI API June 11, 2020 Product OpenAI API We're releasing an API for accessing new AI models developed by OpenAI. Sign up Explore the API access in order to integrate the API into your product, develop an entirely new application, or help us explore the strengths and limits of this technology. initially it kind of got traction at all. And then about a month later, randomly, as far as we can tell, it went viral on Twitter. On the same day, a few different developers kind of found, got it to do something cool, Other people started trying. and then like a lot of people started trying the API. But it was shockingly bad. If you go back and use GPT-3 or 3.5, you will be astonished at how bad the models were then relative to the amount of excitement they generated at the time. So people tried all of these things and really the only business that people got to work in a significant way with GPT-3 was copywriting. and that was like, and not that exciting. And we were kind of like, you know, "Ah" "for a better model." But that was the only business that was working, developers had figured out how to like put in a prompt and be able to chat with it. And we saw this like more people were using they couldn't get for their business, but they were using their API key to just chat. And we said, "Well, we can build a good chatbot." "People clearly want that." And we had a new model. We actually had GPT-4 done, but we had a new model we were ready to release in between called 3.5. And we had figured out a new kind of post-training where we can get the models to do like instruction following. So it can make it easier And we said, "the API" "is not working great." Maybe it was like a $10 or $20 million run rate kind of business, but there is this thing that people love. And under the YC principle of see what your users love and do that, "we'll build a chatbot around it." And we put that out, it was gonna do that well. It was really meant as like a research demo to convince other people that they should build chat-like products and pay us for the API. But that went like crazy viral. And another thing I had learned from YC is when something really starts growing and it's not very good, you have like a guaranteed hit on your hands. And so we had like five days the traffic would shoot up, fall off, and everybody be like, "Well, that was just a hype cycle." By the fourth or fifth day, I know how this works. I know what's gonna happen. Like we have the potential here.

Transcript

We had made GPT-3 and we needed to make money because we wanted to go scale up to, you know, a billion and multi-billion dollar computers. And we had GPT-3 and it was kind of interesting. It was a cool demo, but we couldn't figure out a product to build around it. And we had been thinking, we just couldn't do it. We had tried a few things, they hadn't worked. And so we knew the models were going to get better, but we also wanted to like start a revenue engine sooner. And we said, well, since we can't figure out what product to build, we're just gonna put this into an API. somebody else can figure out what product to build. And so we launched in like, something in the summer of 2020, the GPT-3 API. Initially, it kind of got traction at all. And then about a month later, randomly, as far as we can tell, it went viral on Twitter. On the same day, a few different developers kind of found, got it to do something cool, other people started trying. And then like a lot of people started trying the API. But it was shockingly bad. If you go back and use GPT-3 or 3.5, you will be astonished at how bad the models were then relative to the amount of excitement they generated at the time. So people tried all of these things and really the only business that people got to work in a significant way with GPT-3 was copywriting. and that was like, and not that exciting. And we were kind of like, you know, "Ah" "for a better model." But that was the only business that was working, developers had figured out how to like put in a prompt and be able to chat with it. And we saw this like more people were using they couldn't get for their business, but they were using their API key to just chat. And we said, "Well, we can build a good chatbot." "People clearly want that." And we had a new model. We actually had GPT-4 done, but we had a new model we were ready to release in between called 3.5. And we had figured out a new kind of post-training where we can get the models to do like instruction following. So it can make it easier And we said, "the API" "is not working great." Maybe it was like a $10 or $20 million run rate kind of business, but there is this thing that people love. And under the YC principle of see what your users love and do that, "we'll build a chatbot around it." And we put that out, it was gonna do that well. It was really meant as like a research demo to convince other people that they should build chat-like products and pay us for the API. But that went like crazy viral. And another thing I had learned from YC is when something really starts growing and it's not very good, you have like a guaranteed hit on your hands. And so we had like five days the traffic would shoot up, fall off, and everybody be like, "Well, that was just a hype cycle." By the fourth or fifth day, I know how this works. I know what's gonna happen. Like we have the potential here.