I think we do, except, AI is just another (big, powerful, broad reaching implications) tech, like the internet, or computers, or nuclear fission, or spaceflight (artificial satellites), or powered flight, or the steam engine, or steel making, or circumnavigation capable sailing ships, or gunpowder…
None of those things switched on all at once in an instant, and AI has been creeping up on us for 60+ years. The past 2 or 3 have been a rather dramatic acceleration, fulfilling much of the promise and expectation of the past 50 years, but it’s still not as great as people imagine it could be - a lot like everything else on the list above. Most people “on the cusp” of those technologies had very different visions for what they would bring to the world compared with what actually happened.
Can we “shoot ourselves in the foot” with this one? Yes, but you could do the same with a rock in a bamboo pole with a little gunpowder, too.
But that’s just one guy’s poor foot, or several volunteers tragically crashing to their deaths, etc. These guys are terraforming the planet to be hostile to ours and many (most?) other species as a side effect of their technology.
Even if this technology is twice as earthshattering as the Internet or gunpowder or you name it, it is hard to see the point if most or even many of us have to suffer and die, similarly condemning our heirs, just to force it to market before its time.
This is not some remote future possibility or granola eating hippie concern. There are permit requests on file all over the North America, among others, for AI data facilities that propose to use more water and power than their entire province/state they reside in does, and are securing sweetheart deals for those resources, passing the costs onto the community. The aquifers are draining, the cities are sinking, and their power bills are rising already, and they’re not even built all built yet.
These technologies have a lot of promise and in some non-LLM applications, have already produced incredible results, but to unleash it for mass consumption is irresponsible on a global scale. I can’t help but see current cloud LLMs as large scale basic answer and structured data generators also capable of the not useless ability to perform hundreds of “bad programmer” man-hours in minutes, but they run on a furnace fueled by metaphorical countless burning orphans, and our business and political leaders are taking the stance that is the non-negotiable cost of progress, and it isn’t.
I think we do, except, AI is just another (big, powerful, broad reaching implications) tech, like the internet, or computers, or nuclear fission, or spaceflight (artificial satellites), or powered flight, or the steam engine, or steel making, or circumnavigation capable sailing ships, or gunpowder…
None of those things switched on all at once in an instant, and AI has been creeping up on us for 60+ years. The past 2 or 3 have been a rather dramatic acceleration, fulfilling much of the promise and expectation of the past 50 years, but it’s still not as great as people imagine it could be - a lot like everything else on the list above. Most people “on the cusp” of those technologies had very different visions for what they would bring to the world compared with what actually happened.
Can we “shoot ourselves in the foot” with this one? Yes, but you could do the same with a rock in a bamboo pole with a little gunpowder, too.
But that’s just one guy’s poor foot, or several volunteers tragically crashing to their deaths, etc. These guys are terraforming the planet to be hostile to ours and many (most?) other species as a side effect of their technology.
Even if this technology is twice as earthshattering as the Internet or gunpowder or you name it, it is hard to see the point if most or even many of us have to suffer and die, similarly condemning our heirs, just to force it to market before its time.
This is not some remote future possibility or granola eating hippie concern. There are permit requests on file all over the North America, among others, for AI data facilities that propose to use more water and power than their entire province/state they reside in does, and are securing sweetheart deals for those resources, passing the costs onto the community. The aquifers are draining, the cities are sinking, and their power bills are rising already, and they’re not even built all built yet.
These technologies have a lot of promise and in some non-LLM applications, have already produced incredible results, but to unleash it for mass consumption is irresponsible on a global scale. I can’t help but see current cloud LLMs as large scale basic answer and structured data generators also capable of the not useless ability to perform hundreds of “bad programmer” man-hours in minutes, but they run on a furnace fueled by metaphorical countless burning orphans, and our business and political leaders are taking the stance that is the non-negotiable cost of progress, and it isn’t.