Video from All favorites #995

If OpenAI charged you $200 a month, would you pay it? Be honest. So that's the question everybody's asking themselves right now, because there was a leak out of the code base at OpenAI. I don't mean like a security leak. I mean someone dug through publicly available code and looked at actual lines of code to get this. And what they found was that OpenAI is preparing to launch a new pricing page. There's code in place for that. It's not available or viewable yet. And that right now it shows $200 a month as a pro user. I've got a little Lego guy here. As a pro user. And would you pay as a pro user $200 a month? So the rumor is that the 01 model will drop and that's what it will be priced at. And the only way people will pay that, in my view, is if it is 10x better than a 01 preview, which means that it has to solve organization scale problems. If, for example, I could use it and say, I want you to develop a product strategy for me that is custom to my business, and then have a dialogue with me and the executive team about that product strategy, and do that without me having to go back and tell you to edit, and tell you you're using too many words, and tell you you're being too vague, etc. Just do it and do it correctly. Well, it would save me a lot of time. Like it's possible that becomes something that organizations are willing to look at as super affordable for the value they get. It still feels out of reach for most individuals. And I think the unit economics are such that OpenAI needs to make it something that individuals will pay for in order to justify. You pay that much money for your phone in your pocket, you just don't think about it because it's part of your phone plan. It depends on the phone, right? Like fancy phones, it gets closer to that. If you're on a short payment plan, less fancy phones, it gets much lower than that, and you can choose. They don't really have as many choices at OpenAI. They've got zero, the free tier, $20, the plus tier, and whatever this might be at $200. We also have leaked information on their finances that, again, it's not leaked. Like Sam shared this, it's just that you have to put it together with the code to make sense of it, and that's what I'm doing here. So Sam shared in an interview that there are 300 million weekly active users, which is amazing, but only 18 million people pay according to the CFO of OpenAI. Much smaller number. Now, they're getting a conversion rate that's decent on those active users if they're not already paying. But at the end of the day, their growth has slowed down. They're not growing like 10x a year. They've grown about 25% since June, which is still great growth. It's just not as fast a growth as they've been experiencing, and they need a way to start to eat into the incredible revenue hole that they're in because of the cost of training the next models. I know they just raised a ton, but they are only at a $4.3 billion run rate, and they are spending way, way, way more than that on training. And they have to, to keep up. The way you play the game as a model maker is you put more and more chips down on the table until you get to AGI, artificial general intelligence, until you get to super intelligence. And so from a game theory perspective, they can't stop spending because if they stop spending, they're worse than if they ever started. And so they're stuck in a spending loop, which means they have to increase prices, which I've been telling you guys is coming. And I will say $200 a month would be 10x cheaper than the rumored price two months ago. I think they workshopped $2,000 a month and realized it wasn't going to fly. And so now they're down to 200 a month. We will see. It hasn't dropped yet. It's not live yet. But fundamentally, they need to get to a place where they are moving from 18 million people paying them 20 bucks a month to 200 million people paying them on average, much closer to 50, 60, 70, 80 bucks a month. And they have to justify that with value, right? Like this model now needs to do a whole lot more to show the value. They are trying to get to a spot where they have Apple level free cash flow, which means that this needs to provide iPhone level value. That is the core problem Sam Altman is facing. He hasn't demonstrated he has a model that does that yet. And the unit economics don't make sense unless he can. And that's why he desperately needs to release a one. I am really curious to see what the price would be. You tell me in the comments, would you pay $200 for an amazingly good open AI model?