LLAMA2 is out and it’s pretty fancy, and there’s a news report that they’re making a way bigger one. But if it’s just open source stuff, how do they justify the massive costs to make it? It seems like everyone can just use it for free.
In: 156
The answer, that you’ve already received, is they don’t (sort of).
Now, other people have speculated on *why* they are willing to take a loss on it, and most of them seem to agree that it’s to harm OpenAI and their ChatGPT, but that skips over a very important piece of information that is, they want to catch up.
Many companies were seemingly blindsided by ChatGPT and are in a rush to catch up with the technology. OpenAI clearly figured out a good method to design and train their language models, and now other companies need to do the same. As it stands, Facebook can’t compete with OpenAI, so, in order to entice people to use, and consequently test, their product, they take a loss. In some ways, yes, they’re losing money on this, but in reality, they’re actually just investing it, as this is the cost of catching up to OpenAI.
It’s important to remember that companies are made of people, and these people need to learn how to do things (and do them better) for the company to innovate. Taking a loss on the training and operating cost of their model is really just the cost of their employees learning what OpenAI employees have already learned, but faster.
Latest Answers