4 ms·
It's nothing like bitcoin mining. Bitcoin mining is intentionally designed so that it gets harder as people use it more, no matter what. With LLMs, if you have
by lukeschlather 2y ago
It's nothing like bitcoin mining. Bitcoin mining is intentionally designed so that it gets harder as people use it more, no matter what.
With LLMs, if you have a use case which can run on an H100 or whatever and costs $4/hour, and the LLM has acceptable performance, it's going to be cheaper in a couple years.
Now, all these companies are improving their models but they're doing that in search of magical new applications the $4/hour model I'm using today can't do. If the $4/hour model works today, you don't have to worry about the cost going up. It will work at the same price or cheaper in the future.
- Mistletoe 2y agoBut OpenAI has to keep releasing new ever-increasing models to justify it all. There is a reason they are talking about nuclear reactors and Sam needing 7 trillion dollars. One other difference from Bitcoin is that the price of Bitcoin rises to make it all worth it, but we have the opposite expectation with AI where users will eventually need to pay much more than now to use it, but people only use it now because it is free or heavily subsidized. I agree that current models are pretty good and the price of those may go down with time but that should be even more concerning to OpenAI.
- kergonath 2y ago> But OpenAI has to keep releasing new ever-increasing models to justify it all. There seems to be some renewed interest for smaller, possibly better-designed LLMs. I don’t know if this really lowers training costs, but it makes inference cheaper. I suspect at some point we’ll have clusters of smaller models, possibly activated when needed like in MoE LLMs, rather than ever-increasing humongous models with 3T parameters.