The Fight over Whose AI Monster Is Scariest
wsj.com
wsj.com
You see, this idea did travelled to the past. Maybe not with a time machine, but by means of us humans imagining it before it existed.
Now we're laying the public discourse of a real technology on top of a landing strip made on the popular culture by movies and sci-fi authors.
The plane: the actual technology we're just developing.
The landing strip: the decades of imagining and popular culture.
If anyone truly believes LLMs are a monster, they're mistaken. The monster is within this brahmanian representation of it. That is the dangerous stuff. And it's the stuff media outlets are also playing with a lot.
We're kind of dumb, technical people. We can't see that stuff very clearly. Culture and metaphor and ideas. I mean, we can, but we're naive, idealistic and shallow. What if we weren't?
It's been over a decade, it's years past the point it was supposed to put all of us out of a job. And the same problems they had on day 1 are still there, mathematically proven to be impossible to eliminate. Meanwhile ai companies are planning literally over $1 trillion to build datacenters that would require multiple nuclear reactors each. And they plan to do it in only a couple of years. Meanwhile none of them have even a concept of a plan towards profitability.
How anyone can be hyped by this situation is beyond me. Hoping really hard and having religious faith in the prophecies of the only people who stand to gain anything from any of this, doesn't change the fact that every single promise they've made has been a dud. And then they double down with another promise.
How is it even possible to be hyped?