5 ❤️’s if I could. Fantastic articulation and insight regarding my concerns. This problem first occurred to me as a mathematician when students were no longer taught basic computing skills and simply received results from their omnipresent hand held calculators. I would be in meetings and whenever a problem requiring rudimentary math skills was discussed no one could figure out why I calculated the result more quickly mentally than the time it took to retrieve their calculators and enter a few numbers. Results without effort and understanding are why many people will fail to perceive when their AI is hallucinating.
All AI outputs we see, need a superfine awareness based judgement for its usefulness. Not practicing such skills will quickly drown them in problems they or AI output wont know! Much like the clueless fake- traveller through a sand storm... sad
I agree with this framing, but I don't think it's fair to say this is coming from Silicon Valley. Silicon Valley reflects the logic of the market, which rules our day because it reflects what many people want: to live by getting to destinations with minimal effort, and then to wonder why life seems to lack meaning, which ChatGPT can then talk to you about!
The vision in Requarth’s piece of humanity being trampled over by technology rather than traveling with it leaves out a brutal physical substrate: these systems are literally drinking the West dry.
Hyperscale data centers can each consume on the order of millions of gallons of water per day for cooling, usage comparable to a town of tens of thousands of people, and most of that is drawn from already stressed rivers and aquifers. In the High Plains, projects tied to AI and cloud build‑out are being sited over the already‑declining Ogallala aquifer, turning a slow‑motion agricultural crisis into a digital land rush where farmers watch wells fail while server farms secure priority access to what’s left.
Using this train metaphor should force us to notice what is being shoveled into the boiler.
In the U.S., AI‑driven data centers are already large enough loads to visibly deform regional power markets and planning timelines. Single facilities are being designed at gigawatt scale where one campus draws as much power as a major city, while grid upgrades take five to ten years longer than it takes to announce and build the compute, leaving ordinary ratepayers to absorb higher bills and more brittle infrastructure.
Analysts now project that data centers dominated by AI workloads could consume a mid‑teens percentage of U.S. electricity within the next decade, in some states pushing toward a majority of total demand, making blackouts and renewed reliance on fossil generation a certainty.
All this argument is academic. This is moving forward, and will play the same role in our extinction as the Union Pacific did for the buffalo.
There's truth in what you say, but it's a narrow truth. Tools always eliminate the need for some capabilities, while allowing us to develop others. Something is lost and something else is gained. Pianos, calculators, cameras, electric saws, cars -- would we be better off without them?
I use AI 50 times a day to write better by leaning on it, but also to improve my writing abilities faster than I've ever been able to. I learn new physics, biology, history and programming, faster than ever before.
And I've known since the '60s that AI was the most dangerous thing humans would ever invent. As an undergrad in '69 I took a course in "Finite State Machines," the abstract concept of a computer to learn about AI.
If you want to learn a bit about why it's so dangerous, you can start here: DareToKnow.stoft.com . I've just started posting on this.
Perhaps a better comparison is sailing ships and steamships. The advent of steam powered ships reduced the importance of understanding the interaction of wind, sails and rigging except now for the hobbyist. However, the advent of steam powered ships created the need for new areas of expertise involving engineering, manufacturing, operations and maintenance.
Why do you think Silicon Valley is blind to this? Presumably you're not against locomotives, and you don't think they were overhyped. What's the claim you're making?
5 ❤️’s if I could. Fantastic articulation and insight regarding my concerns. This problem first occurred to me as a mathematician when students were no longer taught basic computing skills and simply received results from their omnipresent hand held calculators. I would be in meetings and whenever a problem requiring rudimentary math skills was discussed no one could figure out why I calculated the result more quickly mentally than the time it took to retrieve their calculators and enter a few numbers. Results without effort and understanding are why many people will fail to perceive when their AI is hallucinating.
All AI outputs we see, need a superfine awareness based judgement for its usefulness. Not practicing such skills will quickly drown them in problems they or AI output wont know! Much like the clueless fake- traveller through a sand storm... sad
I agree with this framing, but I don't think it's fair to say this is coming from Silicon Valley. Silicon Valley reflects the logic of the market, which rules our day because it reflects what many people want: to live by getting to destinations with minimal effort, and then to wonder why life seems to lack meaning, which ChatGPT can then talk to you about!
The vision in Requarth’s piece of humanity being trampled over by technology rather than traveling with it leaves out a brutal physical substrate: these systems are literally drinking the West dry.
Hyperscale data centers can each consume on the order of millions of gallons of water per day for cooling, usage comparable to a town of tens of thousands of people, and most of that is drawn from already stressed rivers and aquifers. In the High Plains, projects tied to AI and cloud build‑out are being sited over the already‑declining Ogallala aquifer, turning a slow‑motion agricultural crisis into a digital land rush where farmers watch wells fail while server farms secure priority access to what’s left.
Using this train metaphor should force us to notice what is being shoveled into the boiler.
In the U.S., AI‑driven data centers are already large enough loads to visibly deform regional power markets and planning timelines. Single facilities are being designed at gigawatt scale where one campus draws as much power as a major city, while grid upgrades take five to ten years longer than it takes to announce and build the compute, leaving ordinary ratepayers to absorb higher bills and more brittle infrastructure.
Analysts now project that data centers dominated by AI workloads could consume a mid‑teens percentage of U.S. electricity within the next decade, in some states pushing toward a majority of total demand, making blackouts and renewed reliance on fossil generation a certainty.
All this argument is academic. This is moving forward, and will play the same role in our extinction as the Union Pacific did for the buffalo.
Brilliant! Bravo! 👏
There's truth in what you say, but it's a narrow truth. Tools always eliminate the need for some capabilities, while allowing us to develop others. Something is lost and something else is gained. Pianos, calculators, cameras, electric saws, cars -- would we be better off without them?
I use AI 50 times a day to write better by leaning on it, but also to improve my writing abilities faster than I've ever been able to. I learn new physics, biology, history and programming, faster than ever before.
And I've known since the '60s that AI was the most dangerous thing humans would ever invent. As an undergrad in '69 I took a course in "Finite State Machines," the abstract concept of a computer to learn about AI.
If you want to learn a bit about why it's so dangerous, you can start here: DareToKnow.stoft.com . I've just started posting on this.
Perhaps a better comparison is sailing ships and steamships. The advent of steam powered ships reduced the importance of understanding the interaction of wind, sails and rigging except now for the hobbyist. However, the advent of steam powered ships created the need for new areas of expertise involving engineering, manufacturing, operations and maintenance.
Why do you think Silicon Valley is blind to this? Presumably you're not against locomotives, and you don't think they were overhyped. What's the claim you're making?