I am aware that the West’s best time was essentially at the expense of the rest of the world - but where can one find opportunity and prosperity of those times today? Was this a great anomaly of history? Is it time to hedge your bets on living out the Chinese dream?
Is it time for the West to accept that what our parents experienced will never exist again?
Will the world overall need to accept that the western living standard of that time will never be reached again? (It’s too unsustainable - for the planet, and because superprofits won’t flow as they once did)
What has actually happened that’s caused it to collapse?
Is it tho? The first few iterations of LLMs were pretty impressive and showed promise but it seems to me like progress is slowing down, also pretty much all AI companies are operating at a loss with seemingly no end in sight, these things suck up all the hardware and electricity you can get and even the best models are not even all that useful right now, they still make mistakes and can’t really do anything original. The only real winner in the AI race is Nvidia lmao.
It’s definitely a bubble that’s gonna pop, some of it is going to survive but I don’t at all expect society to fundamentally transform due to these things.
Yeah it really is. I was the world’s #1 hater and without getting into this much detail I ended up being in a position where like it or not I do this for a living. The technology is real. You can draft contracts without a paralegal, you can process invoices without accountants. As the technology progresses more and more will be subsumed.
Yeah writing boilerplate documents/emails/code/whatever is what it’s good at, not much more.
I’m a software developer by trade and in my experience those things aren’t all that good at coding unless you’re really really specific about what you want and it’s not super complicated. I like using it for some utility scripts and some boring boilerplate stuff but beyond that it’s just not all that useful.
So a good example is we had to refactor some models to use pyspark, and our data scientists were able to speed up the process probably 3x by using Claude. Like it won’t take you start to finish, this process that is being sold is a definitive lie, but it will dramatically increase productivity when properly applied. And as people figure out how to improve the results through proper context application, the use cases will grow. That shit destroys jobs. Because now a data scientist can do 3x as much.
The things can’t think, but they are basically intellectual mech suits, and the societal implications are profound. A lot of folks on hexbear I think dismiss the tech prematurely without really understanding what it is or how it works