Trump's AI agenda is often described as a push for deregulation; an effort to cut red tape so AI companies can innovate. In truth, it's a big government project with designs toward domination.
Possibly before you wrote this, Carney spoke at Davos about "middle countries" creating alliances where their goals were in common. Europe is looking to exclude and/or regulate US technology. China blocked imports of NVIDIA H200 GPUs, possibly to remove supplier leverage and/or to spy on China. The world is intent on forcing the USA to stay in its nattowed "hemisphere," which will reduce the power of US technology in the world. The hegemon will cripple itself.
Excellent framing here. The whole deregulation narrative falls apart once you track where the money is actually going. Federal stakes in Intel and eight other supply chain companies plus favorable land deals for data centers, that's not hands off, thats industrial policy with a nationalist bent. I remembr reading about the Nvidia China deal and thinking it was strange nobody was calling it what it was.
Indeed, the geopolitical ramifications of this situation are substantial too and will have far-reaching consequences for decades to come. This is exacerbated by the fact that most of these “activities” go unnoticed by many, including most politicians and individuals worldwide. However, publishing and knowing “them” alone is, unfortunately, not going to change the essential elements of the impact of these “activities." Not only people in the United States but humanity as a whole is now confronted with increasingly perilous circumstances, exacerbated by the deliberate deployment and subsequent use of autocratic "Government by Grok" technologies (as explained by Brian Merchant) by many users around the world. To put it simply, we’re still trying to find ways to navigate the existential challenges, not moving past them.
"I was struck by how bad the images are. This is a technology that I keep getting told has made unbelievable progress over the last year ...": In the same vein, here's a quote from Uberto Barbini's new book "Process over magic: Beyond vibe coding":
"The trend with new models has been consistent: every few months a highly hyped release arrives, boasting improvements in areas where LLMs are already strong - supporting more languages, frameworks, and algorithms. But the same core issues remain. They still produce code that doesn't compile because of calls to non-existent methods. They still follow instructions too literally without solving the real problem. And they still make frustrating mistakes no human developer would make, like changing test assertions just to get them to pass."
Mind you, Barbini is pro-"AI" for coding. His book is about the "babysitting" (his term) it takes to get code good enough for anything important, which to my mind is extremely burdensome. (I'm a freelance software developer.)
The point is, this stuff still isn't anything like intelligent, and there's a reasonable argument that it hasn't been getting better in anything but superficial ways (i.e., "more languages, frameworks, and algorithms"). Moreover, this lack of genuine progress is unsurprising to serious students of machine learning:
"... systems with human-level reasoning are unlikely to be achieved through the approach and technology that have dominated the current boom in AI, according to a survey of hundreds of people working in the field. More than three-quarters of respondents said that enlarging current AI systems ... is unlikely to lead to what is known as artificial general intelligence (AGI). An even higher proportion said that neural networks ... alone probably cannot match or surpass human intelligence."
In my judgment, that shift is unlikely to emerge from Silicon Valley. More likely, like most significant progress in science and technology, it will emerge from academic research. And although the USA has historically led the world in the relevant kinds of academic research, the current regime is systematically sabotaging that advantage - killing the goose that laid the golden eggs, one might say. So it's increasingly likely that major progress will happen elsewhere.
Trump and his stooges are avidly pursuing "AI dominance" in the belief that what currently passes for AI can be turned into something immensely powerful by throwing more money at it. But they're stupid in general and mistaken in this regard specifically. To be sure, they'll waste vast amounts of public money, including the money they'll squander on pumping up the deflating "AI" bubble. They'll degrade public services by replacing civil servants with "AI". And infusing decision-making at, for example, their "Department of War" with "AI" may well lead to spectacular blunders. It will be a stupendous debacle, as indeed all of American public life is rapidly becoming. But their fondest, most sociopathic hopes for "AI" almost surely won't be fulfilled, as they're founded on little more than magical thinking.
I would love to hear your take on the Military rollout of Gemini and x.AI across 3 million American military users. I have written 3 articles on this, and still the mainstream media is covering it like it's just a procurement issue. Of all the other AI hysteria, this isn't even a blip - which is wild to me, because IMO this SHOULD worry people.
Possibly before you wrote this, Carney spoke at Davos about "middle countries" creating alliances where their goals were in common. Europe is looking to exclude and/or regulate US technology. China blocked imports of NVIDIA H200 GPUs, possibly to remove supplier leverage and/or to spy on China. The world is intent on forcing the USA to stay in its nattowed "hemisphere," which will reduce the power of US technology in the world. The hegemon will cripple itself.
Stupid is, as stupid does.
I've been reading Palo Alto by Malcolm Harris.
What is old is new again in the synergy of California Tech and American dominance
Excellent framing here. The whole deregulation narrative falls apart once you track where the money is actually going. Federal stakes in Intel and eight other supply chain companies plus favorable land deals for data centers, that's not hands off, thats industrial policy with a nationalist bent. I remembr reading about the Nvidia China deal and thinking it was strange nobody was calling it what it was.
I notice more and more censorship and bias in the way AI functions. 😡
Indeed, the geopolitical ramifications of this situation are substantial too and will have far-reaching consequences for decades to come. This is exacerbated by the fact that most of these “activities” go unnoticed by many, including most politicians and individuals worldwide. However, publishing and knowing “them” alone is, unfortunately, not going to change the essential elements of the impact of these “activities." Not only people in the United States but humanity as a whole is now confronted with increasingly perilous circumstances, exacerbated by the deliberate deployment and subsequent use of autocratic "Government by Grok" technologies (as explained by Brian Merchant) by many users around the world. To put it simply, we’re still trying to find ways to navigate the existential challenges, not moving past them.
"I was struck by how bad the images are. This is a technology that I keep getting told has made unbelievable progress over the last year ...": In the same vein, here's a quote from Uberto Barbini's new book "Process over magic: Beyond vibe coding":
"The trend with new models has been consistent: every few months a highly hyped release arrives, boasting improvements in areas where LLMs are already strong - supporting more languages, frameworks, and algorithms. But the same core issues remain. They still produce code that doesn't compile because of calls to non-existent methods. They still follow instructions too literally without solving the real problem. And they still make frustrating mistakes no human developer would make, like changing test assertions just to get them to pass."
Mind you, Barbini is pro-"AI" for coding. His book is about the "babysitting" (his term) it takes to get code good enough for anything important, which to my mind is extremely burdensome. (I'm a freelance software developer.)
The point is, this stuff still isn't anything like intelligent, and there's a reasonable argument that it hasn't been getting better in anything but superficial ways (i.e., "more languages, frameworks, and algorithms"). Moreover, this lack of genuine progress is unsurprising to serious students of machine learning:
"... systems with human-level reasoning are unlikely to be achieved through the approach and technology that have dominated the current boom in AI, according to a survey of hundreds of people working in the field. More than three-quarters of respondents said that enlarging current AI systems ... is unlikely to lead to what is known as artificial general intelligence (AGI). An even higher proportion said that neural networks ... alone probably cannot match or surpass human intelligence."
(https://web.archive.org/web/20250305233251/https://www.nature.com/articles/d41586-025-00649-4)
Similarly:
"These findings highlight the need for a fundamental shift in the design and development of general-purpose artificial intelligence ..."
(https://www.nature.com/articles/s41586-024-07930-y)
In my judgment, that shift is unlikely to emerge from Silicon Valley. More likely, like most significant progress in science and technology, it will emerge from academic research. And although the USA has historically led the world in the relevant kinds of academic research, the current regime is systematically sabotaging that advantage - killing the goose that laid the golden eggs, one might say. So it's increasingly likely that major progress will happen elsewhere.
Trump and his stooges are avidly pursuing "AI dominance" in the belief that what currently passes for AI can be turned into something immensely powerful by throwing more money at it. But they're stupid in general and mistaken in this regard specifically. To be sure, they'll waste vast amounts of public money, including the money they'll squander on pumping up the deflating "AI" bubble. They'll degrade public services by replacing civil servants with "AI". And infusing decision-making at, for example, their "Department of War" with "AI" may well lead to spectacular blunders. It will be a stupendous debacle, as indeed all of American public life is rapidly becoming. But their fondest, most sociopathic hopes for "AI" almost surely won't be fulfilled, as they're founded on little more than magical thinking.
I would love to hear your take on the Military rollout of Gemini and x.AI across 3 million American military users. I have written 3 articles on this, and still the mainstream media is covering it like it's just a procurement issue. Of all the other AI hysteria, this isn't even a blip - which is wild to me, because IMO this SHOULD worry people.
https://twvme.substack.com/p/accelerate-like-hell-the-thermodynamics