Technology

Wages or health insurance are a very known cost, with a known return. At some point the curve flattens and the return gets less and less for the money you put in. That means there is a sweet spot, but most companies don't even want to invest that much to get to that point.

AI however, is the next new thing. It's gonna be big, huge! There's no telling how much profit there is to be made!

Because nobody has calculated any profits yet. Services seem to run at a loss so far.

However, everybody and their grandmother is into it, so lots of companies feel the pressure to do something with it. They fear they will no longer be relevant if they don't.

And since nobody knows how much money there is to be made, every company is betting that it will be a lot. Where wages and insurance are a known cost/investment with a known return, AI is not, but companies are betting the return will be much bigger.

I'm curious how it will go. Either the bubble bursts or companies slowly start to realise what is happening and shift their focus to the next thing. In the latter case, we may eventually see some AI develop that is useful.

There are not enough 💯 emoji in the world for this post.

it makes sense to someone like me who is not a dev but works with coding at times, I don't get the experience to be quick with it.

Computer science degrees being the most unemployed degree right now leads me to believe this will actually suppress wages for some time

I hardly post on any social media besides here and I still feel violated.

Tech CEOs really should be replaced with AI, since they all behave like the seagulls from Finding Nemo and just follow the trends set out by whatever bs Elon starts

Suppose many of the CEOs are just milking general venture capital. And those CEOs know that it's a bubble and it'll burst, but have a good enough way to predict when it will, thus leaving with profit. I mean, anyway, CEOs are usually not reliant upon company's performance, so no need even to know.

Also suppose that some very good source of free\cheap computation is used for the initial hype - like, a conspiracy theory, a backdoor in most popular TCP/IP realizations making all of the Internet's major routers work as a VM for some limited bytecode for someone who knows about that backdoor and controls two machines, talking to each other via the Internet and directly.

Then the blockchain bubble and the AI bubble would be similar in relying upon such computation (convenient for something slow in latency but endlessly parallel), and those inflating the bubbles and knowing of such a backdoor wouldn't risk anything, and would clear the field of plenty of competition with each iteration, making fortunes via hedge funds. They would spend very little for the initial stage of mining the initial party of bitcoins (what if Satoshi were actually Bill Joy or someone like that, who could have put such a backdoor, in theory), and training the initial stages of superficially impressive LLMs.

And then all this perpetual process of bubble after bubble makes some group of people (narrow enough, if they can keep the secret constituting my conspiracy theory) richer and richer quick enough on the planetary scale to gradually own bigger and bigger percent of the world economy, indirectly, of course, while regularly cleaning the field of clueless normies.

Just a conspiracy theory, don't treat it too seriously. But if, suppose, this were true, it would be both cartoonishly evil and cinematographically epic.

In no small part because they see it as a time-limited gateway to permanent, infinite profits through market consolidation, job cutting, and government contracts. After all, if they get there FIRST, it's all theirs, and the infinite profits then will make up for all the money spent now. Never mind the fact that in doing so they'll destroy the environment, the economy, and the world long before they can actually SPEND those profits on anything.

Sunken costs?