The Sweet Spot of Open Source: Why a 6x Profit Margin Is Just Right

Everything has its difficulties, so this is a sweet spot for a company our size. If we were larger, we might have other issues; if smaller, our resources would be insufficient. As for open source, I think we should set a price. We should realize that we won't be overly aggressive. With our pricing model, I won't charge a very high fee; I might set it based on a ten-month payback period. With a ten-month payback, we can already make it unprofitable for competitors...

With a ten-month payback, we can make it unprofitable for third parties to deploy independently. They can't match this cost; they simply can't. So open source doesn't affect my revenue. Of course, if I wanted to make a 100x profit, then open source would... I hear you, but the video seems to have dropped, boss. Maybe a call came in.

Open source, I believe, has no impact on our business model. The premise is that we only make a 6x profit, with a ten-month payback period roughly corresponding to a 6x profit margin. If we only make a 6x profit, open source won't have any impact. But if you want to make a 100x profit, then open source will indeed affect that, because third parties will deploy it, possibly at a 20x cost, lower than yours.

Is this model sustainable in the long run? I think it is. Under our vision, I believe open source is sustainable long-term, and we intend to do it. You could say it's both restrained and gives you long-term benefits. This strategy allows us to have more opportunities at the forefront of technology, increasing our probability of achieving AGI. We are more at ease.

Think about it, we don't even need to work overtime because it's not that hard. But for other companies, it might be very hard because they overthink things. Actually, it's not hard at all. Initially, it seemed like we chose a difficult model-doing research, tackling the hardest problems-like a hard mode. But in reality, we gave up a lot in other areas, which makes us very strong and allows us to do things very easily.

So regarding open source, my judgment is that it's sustainable. There is no conflict between open source and commercial payment, as long as it's under a 6x profit margin. A 6x profit might seem high, but it's actually not. Given how efficient AI is today, a reasonable profit margin might be around that. In the future, it might drop to, say, 4x or 3x. I think that's already... it can't go much lower than that. But it will still have significant profit.

Just looking at selling APIs, I don't think that's very attractive. But you can see there's no conflict. I'm not worried about others deploying our models and competing with us-not at all. In fact, we hope they can deploy them.

We do our best to help the open-source community deploy our models. I'm not worried about them taking business away from us because the market is big enough. I'm only worried that they might not deploy successfully, get some details wrong, and end up with poor performance or higher costs. Yes, there is no conflict here.

Last year, when we had the B2B business, a common question was: our consumer side is open source, so won't that conflict with our consumer side? Since we don't have an advantage in traffic, or for example Tencent has a lot of traffic, they could deploy our open-source model and take all the consumer users away from us. But that doesn't actually happen, for many reasons.

Then the question: is the open-source model we provide the same as the one we deploy ourselves? Yes, it's the same. We don't open-source an inferior model and use a better one ourselves-no, it's the same. This also shows there's no conflict. All of last year, on the consumer side, we basically open-sourced, and we saw no conflict in consumer services-truly no conflict. So that's the open-source part.

Then there's also the company's long-term vision. I think our goal should be AGI. Everyone's definition of AI may differ, but that doesn't stop us from taking AGI as our goal. From a technical roadmap perspective, the path to AGI is quite clear. Compared to the current generation of AI technology, if you can describe a problem very clearly and give it complete context and instructions, it already surpasses humans. But there is a definition and a prerequisite: you give it complete context and complete instructions.

This definition is hard to achieve. For example, in today's meeting, we have a long and strong context-maybe decades of context among us-which AI doesn't have. What AI can do is, within a limited context, perform better than humans. But it still cannot replace humans. What's missing is continuous learning. Humans can continuously learn. When you hire an employee, they might spend two months getting familiar with the company environment and their work, and after those two months, they can get started.

They can do many things. They can understand what you say-for example, if you say 'get Xiao Wang over here,' they know who Xiao Wang is. But for AI, because it lacks that two months of learning context, if you tell it to get Xiao Wang, you'd have to tell it who Xiao Wang is, what their position is, where they are, how to find them, and what to pay attention to. You'd have to give AI all the context. In that case, AI could do it, but you can't give it all the context-it's not realistic.

So AI cannot replace your employees. But if AI had the ability for continuous learning, like your employees, and it could spend two months learning at the company, then it could replace all humans. So we are still one step away from the next stage: learning to learn. We can understand AI development as a staircase. Last year's step was CoT, or chain-of-thought. Because we discovered that chain-of-thought reasoning allows intelligence to reach a higher level. By thinking on its own, it raises the ceiling and enables AI to do more.