Clojurians Log v2

the crazy part is that money could be applied almost anywhere else and it would be more beneficial to all

Er, "Go Nvidia!" I guess? 🤯

Nvidia is there because it's providing the chips for all those companies hehe

They need to add more than NVIDIA, all the RAM makers, the companies that make the machines to build the chips, and so on

especially interesting with the Anthropic and OpenAI plans to IPO... they will need to grow or show some path (even if unbelievable) to profitability

Ed Zitron has been very vocal about this for some time already https://www.wheresyoured.at/the-subprime-ai-crisis-is-here/

The whole point of AI is not generating revenue: 1. it makes required hype for the stock market (just a money game like 2. It is the only industry/chance western world could out compete China (& Global South) See it through this lens 📈

perhaps, but you can't sustain a model where you lose $5 for every $1 you make forever

Yes, but they not pay the price for the cost, government give them money which they get from middle & lower class, so it is a successful tool for making rich richer Something like TulipMania https://en.wikipedia.org/wiki/Tulip_mania which a Tulip in 1634 in Dutch in which one Tulip cost become equal to a house! Was it madness? yes But it works for speculation & making rich richer, and the only thing matter is making rich richer. Investing in public health does not making rich richer, it helps middle & lower class

I think tulip mania comparison was more apt against the previous hype (NFT / crypto stuff)... but LLMs do actually have use cases they are good for, so it isn't just a completely useless asset

but there sure seems to be market irrationality regarding AI hype as well... interesting to see if/when the bubble pops

The big labs, Anthropic, OpenAI, Google, the question is, can they increase price and not lose customers? Can they monetize through ads and not lose customers? If yes, then they have a path to profitability. It's just like Uber and other tech companies otherwise, get people hooked with subsidized cost, than increase price.

One indirect use case for LLM providers is intelligence , so in the booking it seems they are not profitable, but their real use case is something darker, exactly like social networks , They could be used for gathering intelligence!

NVIDIA selling shovels during a gold rush

@didibus I think their biggest problem is that they overhyped the Davos crowd. The kings then ordered their dukes to AI ASAP, the dukes tasked the marquises, the markuises delegated to earls, etc. The artisans are currently burning tokens and scratching their heads, while peasants are getting pushed out of farms and into the cities. The tech is useful, but it is not THAT useful, and even if it was THAT useful who will have the money to pay for it when everyone's out of their jobs...

@didibus the old american recipe for a "disruptive" tech company: profit on something legally grey (preferentially morally grey as well) and them put the money on lobbying so it stays legal (and in PR if people insists in calling it immoral) is remarkable how fast AI labs got a hold of america, in a sense they own the country

I suspect AI is a strategical tool for politics, their goal is beyond business. Even Pope Leo is worried.

This ^ It's called Capacity Building and Dual Use Technology