I don’t see how they are different. Their uses, yes, but the fundamentals look the same, don’t they? Everything still is broken down into quanta, discrete steps. They haven’t invented a new AI or a chip that goes along with it that “flows”.
LLMs are just one branch of a much wider field of AI research, and analog circuits that deal with flows instead of discrete values tend to be more efficient for modelling dynamic systems such as traffic flows.
Honestly, the article is crap. It doesn’t mention any of this. Thanks for linking it.
I remember reading about analog neural circuits somewhere but wasn’t aware this was what the author was referring to. He just went around in circles describing how they were used.
I thought it was interesting conceptually even if light on details. In the west, most application for AI has been in the realm for content generation like making images, documents, writing code. Meanwhile, in China AI systems are used for stuff like monitoring traffic systems, maintaining high speed rail networks, and other types of dynamic systems management. I haven’t really heard much about AI being applied in this way in western countries, and it seems like a far more practical use to me.
You don’t see the difference between an AI that can design + 3D print rocket engines, or manage robotic arms on an assembly line, & Grok? I’m not big on the I Ching unless novelists are using it, I think the OP article buries the lede on the AI race (the only one who wants to race is the US, & its losing)
The real battle is between primarily proprietary US models & the state-corralled primarily open source Chinese scene
That isn’t the difference the author is describing. They are talking about discrete vs continuous and bring it up a number of times. That is a fundamental difference in operation. Nowhere does it state how this is achieved. A neural network? LLM? Some advanced stochastic model? What?
Right the article sucks. Anyone who tries to make this about ideals rather than meticulous management of the quant & datacenter firms by the CPC is a bonehead. If you want good writing from Asia Times check this guy out. I met him on twitter and thought he was just a horny insane old man but he’s actually quite intelligent!
I’m pretty sure the article itself is generated. This sort of refusal to commit to saying anything concrete is a hallmark of AI generated articles. The closest it comes to saying how AI is being used differently and it hedges with an “imagine this” sort of statement without references:
Now imagine AI in a different role. Instead of answering questions, it adjusts traffic lights as congestion builds. It balances electricity across a grid as demand shifts. It predicts equipment failures before they occur, reroutes freight around disruptions or continuously updates a digital model of an entire city as millions of people move through it.
Imagine A.I. deciding to fuck your commute home because you and 2 other cars are turning but there’s 20,000 cars going straight so it trolley problems your ass into traffic purgatory
I don’t see how they are different. Their uses, yes, but the fundamentals look the same, don’t they? Everything still is broken down into quanta, discrete steps. They haven’t invented a new AI or a chip that goes along with it that “flows”.
Not sure what to make of this article…
Not really, China has been experimenting with a number of different approaches using analog neural circuits
LLMs are just one branch of a much wider field of AI research, and analog circuits that deal with flows instead of discrete values tend to be more efficient for modelling dynamic systems such as traffic flows.
Honestly, the article is crap. It doesn’t mention any of this. Thanks for linking it.
I remember reading about analog neural circuits somewhere but wasn’t aware this was what the author was referring to. He just went around in circles describing how they were used.
I thought it was interesting conceptually even if light on details. In the west, most application for AI has been in the realm for content generation like making images, documents, writing code. Meanwhile, in China AI systems are used for stuff like monitoring traffic systems, maintaining high speed rail networks, and other types of dynamic systems management. I haven’t really heard much about AI being applied in this way in western countries, and it seems like a far more practical use to me.
You don’t see the difference between an AI that can design + 3D print rocket engines, or manage robotic arms on an assembly line, & Grok? I’m not big on the I Ching unless novelists are using it, I think the OP article buries the lede on the AI race (the only one who wants to race is the US, & its losing)
The real battle is between primarily proprietary US models & the state-corralled primarily open source Chinese scene
That isn’t the difference the author is describing. They are talking about discrete vs continuous and bring it up a number of times. That is a fundamental difference in operation. Nowhere does it state how this is achieved. A neural network? LLM? Some advanced stochastic model? What?
Right the article sucks. Anyone who tries to make this about ideals rather than meticulous management of the quant & datacenter firms by the CPC is a bonehead. If you want good writing from Asia Times check this guy out. I met him on twitter and thought he was just a horny insane old man but he’s actually quite intelligent!
https://asiatimes.com/author/han-feizi/
I’m pretty sure the article itself is generated. This sort of refusal to commit to saying anything concrete is a hallmark of AI generated articles. The closest it comes to saying how AI is being used differently and it hedges with an “imagine this” sort of statement without references:
Imagine A.I. deciding to fuck your commute home because you and 2 other cars are turning but there’s 20,000 cars going straight so it trolley problems your ass into traffic purgatory