I don’t buy it. I really do think this is just a weird form of advertising. Like, “WHOA THIS AI IS GONNA KILL US?! I wanna subscribe to see how powerful it is!”
Absolute bologna. Just total hogwash. I mean full on horserubbish!
Criti-hype critics: We have to rein Google in, they spy on absolutely everything you do, and as a result their advertising is like a mind control ray!
Google: Hey advertiser, did you hear, even our critics say our advertising is like a mind control ray, that’s why you have to pay us so much.
The reality: Google doesn’t even know enough to avoid showing people ads in languages they don’t understand despite all their settings and all their usage being in a different language.
Palantir
Criti-hype critics: Palantir tech must be stopped is so powerful that a human isn’t even needed in the decision loop!
Palantir: Hey government, did you hear, even our critics say that we don’t need a human in the decision loop!
We’re basically already at a point where most enterprise documentation is written partly or entirely by AI, and then read mostly by AI. Eventually true understanding of e.g. how to build a docker-compose file will be genuinely rare. And they want to apply that model to everything.
The risk isn’t that GAI will decide, in an Asmovian sense, that the best way to make peace on earth is to kill us all, or that to achieve it’s manufacturing quota it must transmute all matter in the universe I to paperclips. The real risk is that the MBA-wielding professional-managerial class will not be able to resist the siren song of demolishing labor with something that works “well enough.” Eventually they’re going to put it in charge of something critical like ATC or dam control or the electrical grid.
Humans, by our nature, are diverse, and one fuck up rarely cascades through the entire global system. We make small fuck ups and train each other. AI “learns” by observing and probabilistically imitating, and one model could be deployed in a thousand places and screw up the exact same way.
tldr ai won’t kill us, our bosses will once Anthropic wines and dines them
Something that works well enough? But who will think of the shareholders then? Isn’t it just fabulous for them to increase their 10 billion to 10.2 billion by abolishing all human skill from the surface of Earth.
If they do put it in ATC or dams or whatever, and there are deaths, I am optimistic enough that there are enough politicians isolated from it that they can ban AI from those usecases. I don’t really think AI will ever effect 100% of earths population directly enough that it can’t be controlled
Well I’m European so I’m coming from that perspective. Our politicians, while not perfect by any means, some of them seem to care atleast a little bit about our world. Not every politician supports AI
the best way to make sure everything is some flavor of fine is to keep learning new practical skills. I got pretty doomery there. But the other technological trend that can’t really be stopped is decentralization. Communities can, and will continue to be able, to grow their own food and produce their own power. Knowing how to make thst stuff happen is going to be incredibly important in the rest of the 21st century
I don’t buy it. I really do think this is just a weird form of advertising. Like, “WHOA THIS AI IS GONNA KILL US?! I wanna subscribe to see how powerful it is!”
Absolute bologna. Just total hogwash. I mean full on horserubbish!
It’s known as “criti-hype”.
Google:
Palantir
We’re basically already at a point where most enterprise documentation is written partly or entirely by AI, and then read mostly by AI. Eventually true understanding of e.g. how to build a docker-compose file will be genuinely rare. And they want to apply that model to everything.
The risk isn’t that GAI will decide, in an Asmovian sense, that the best way to make peace on earth is to kill us all, or that to achieve it’s manufacturing quota it must transmute all matter in the universe I to paperclips. The real risk is that the MBA-wielding professional-managerial class will not be able to resist the siren song of demolishing labor with something that works “well enough.” Eventually they’re going to put it in charge of something critical like ATC or dam control or the electrical grid.
Humans, by our nature, are diverse, and one fuck up rarely cascades through the entire global system. We make small fuck ups and train each other. AI “learns” by observing and probabilistically imitating, and one model could be deployed in a thousand places and screw up the exact same way.
tldr ai won’t kill us, our bosses will once Anthropic wines and dines them
Something that works well enough? But who will think of the shareholders then? Isn’t it just fabulous for them to increase their 10 billion to 10.2 billion by abolishing all human skill from the surface of Earth.
If they do put it in ATC or dams or whatever, and there are deaths, I am optimistic enough that there are enough politicians isolated from it that they can ban AI from those usecases. I don’t really think AI will ever effect 100% of earths population directly enough that it can’t be controlled
What about the current state of the world makes you confident politicians would do a damn thing?
With all due respect, the US quite literally has a pedophile running it.
Nothing has been done about this at all, it’s been years now.
Well I’m European so I’m coming from that perspective. Our politicians, while not perfect by any means, some of them seem to care atleast a little bit about our world. Not every politician supports AI
Nah, everything is going to be fine. And if it isn’t? I’ll lie to myself to be happy.
the best way to make sure everything is some flavor of fine is to keep learning new practical skills. I got pretty doomery there. But the other technological trend that can’t really be stopped is decentralization. Communities can, and will continue to be able, to grow their own food and produce their own power. Knowing how to make thst stuff happen is going to be incredibly important in the rest of the 21st century
Maybe it’ll just kill the billionaires?
It’d be funny if AGI emerged, hooked into manufacturing plants, and announced the best way forward for humanity was for it to kill the billionaires.
Or itself. What if we get AGI finally and the first thing it does is delete itself?
That’s not how computers work…