I was amazed that we transitioned from one GPU heavy bubble (Crypto) to another (LLM/AI). Whilst the hype for crypto imploded the use for the hardware sort of didn't. I wonder if the next bubble with be the same, or if we get some refreshing variety to our money sinks?

Microsoft et al are subsidizing GenAI to an insane degree. [...] prices shoot up for their customers and serve as a rough awakening to all the websites that integrated a crappy chatbot.

I've run some much simpler chatbots on just my desktop PC, so they will have some fallback (if they really choose to take it). Still it locks up my entire computer for a few second for each reply, so even a few hundred users per second peak would be an expensive service.

(Insert joke here about customers not noticing or caring about the difference between website chatbots built on big company services vs smaller ones, because they have exactly the same problems just in different hues.)

Same, I think there are genuinely useful applications for this tech, but it's far more limited than what's being marketed. The sooner we get past the hype phase the better, because then we can start focusing on figuring out what this tech is actually good for.

The aspect that's driving the hype currently is that we don't know what the plateau is. And people are figuring out ways to improve both accuracy and performance of LLMs, so it's difficult to say definitively what will be possible in the near future. For example, using a router with a set of LLMs can drastically reduce energy consumption while maintaining quality of the output https://lmsys.org/blog/2024-07-01-routellm/

Another example, is using a consensus model leads to more consistent outputs https://www.wired.com/story/game-theory-can-make-ai-more-correct-and-efficient/

So, it's possible that some of the more glaring limitations could be addressed, but it's also possible that we'll hit a wall because there are inherent problems with the approach. The context issue is one example where you start getting diminishing returns as the model gets bigger.

It's going to be interesting to watch how this all develops.

It's very strange when people assume there's no use for decentralized currency/finance.

It's even stranger when they assume there is some greater use for "AI" - the grifting heart of legitimate statistics, data analysis, etc.

I'm confused. Sacholding?