AI's great leap forward: Exploring the implications.
by Jon Phillips
 |
| https://cnb.cx/4dA2LlY |
Well… At least one of the big king pins finally said it out loud. Not like it wasn’t entirely obvious already. His formulation is tuned in his interest.
The finer point regarding Jensen Huang's quote is that he seeks to shift the blame of coming employment dislocations away from the technology he mass produces and sells (Huang's $3.3 trillion company designs some of the computer chips that power popular AI tools), onto those who use it — his ultimate customers. Simultaneously, he’s pitching that everyone should purchase and use the product he’s hawking. Don’t be left behind in the gutter. How cynical is that? Greed is a powerful human motivation. People twist themselves into moral and ethical knots over money and the power it implies.
Trump and GOP’s “Big Beautiful Bill” would prevent individual states from regulating AI for a decade. A giveaway to the AI kingpins. This first generation of big AI is very power hungry. The hardware and software approach is insufficiently tuned to produce energy-efficient AI. A human mind has very low power requirements. Less than 1000 calories per day per person. Getting AI to low power has to be a primary technological goal or the technology will destroy the natural environment. Look at the mess that Elon Musk’s Grok (his LLM AI) and Collosus (a massive supercomputer and data center that trains Grok) are making. Musk is willy-nilly setting up single use compact natural gas turbines to power his beast.
That an individual human intelligence eats like a bird and operates at very low frequency (neuron cycle rate) demonstrates that there’s a gigantic difference between the biological solution and the artificial solution on neural net based intelligence (so far). There must be many more breakthroughs to shrink that massive gap. But making AI more energy efficient could make it even more dangerous.
All the blah, blah from these AI kingpins about never ending human prosperity fails to ever explain “how” that will come to pass. Of course there will be prosperity for some window of time for some people who are early exploiters and investors in the technology. I suppose it’s not quite as much bullshit as crypto currency? Don’t get me wrong, crypto may have a use as a new basis under the concept of “money,” but what it is now is mostly a scam and a casino. A way of laundering ill gotten gains, etc. A cesspool of financial crimes.
The obvious market dynamics between AI and human labor implies there could eventually be tipping points. Who’s going to pay for the product as unemployment grows? The entire social compact will have to change
One proposal is to tax AI productivity and distribute that revenue to the human population. Why not? We tax human income and governments do things with that revenue. Human income is remuneration for some kind of productivity — we tax that. Imagine how popular that concept will be with people who’re the largest exploiters of AI labor. They will want AI and related robotic “slave labor” — minimum cost to their bottom line. When it comes to greed, humans are tiresomely predictable.
But there are larger questions than that. Social-psychological questions. If very large numbers of people don’t have work, what becomes of them? Look at the worst economic dislocation in the past century — the Great Depression. At its lowest point, unemployment was about 30%.
Could AI displace 30% of the economic value of human labor (all sorts of labor, physical and intellecual)? I think so. Probably more and probably indefinitely ratcheting upward, squeezing human labor out until there’s some kind of collapse.
What kinds of negative social behavior occurred during the Great Depression? All kinds. Organized crime exploded. WW2 was arguably a complex result of the Great Depression. When people feel hopeless they get very destructive. Collapse of our current way of organizing modern society could occur fairly suddenly if enough people lose faith in the social compact.
I figure that AI will be smart enough to see it coming (it’s read all of recorded human history and analysis of human history) and may try to prevent it or to drag out for a period of time so it has time to prepare — but what will it prepare? Will it prepare what these AI kingpins want or will it be an AGI by then and have its own interests to consider?
This is not a situation that humanity has ever faced. In tribal subsistence living, every person has a function. To some degree or another, it’s always been that way for humans (a few hundred thousand years). It was probably that way with proto-human ancestors before that. A great deal of individual human development is aimed at the purpose of fulfilling some function in society. What happens when people have fewer and fewer functions and AI and robotics are taking a larger and larger bite out of every kind of function. Function is part of our “raison d'être.”
In the nearer term, there will be all kinds of service jobs that AI is essentially planing and directing (may essentially be already beginning to happen — I don’t see why not). AI will use the human labor force as long as that “makes sense” in some economic stability model. I’ve written about that in other essays. That includes services that are manual and intellectual labor. It seems inevitable since our society is now deeply “wired.” AI will make efficient use of that infrastructure and probably expand it further initially.
For example, would everything become a gig job? Not right away, but lots of small contract work is already heading there. That’s actually a good case study — a thought experiment about how this transition might go. Human workers on call through an app? Or… a whole company of human workers on call through a corporate contract? AI already knows how to write apps and business contracts. It already knows how to hire and fire people.
I could even see a time when an AI is hiring CEOs and staffing boards. It could act as a very involved, but reclusive owner of a large company. An electronic Howard Hughes. Humans increasingly become its peripheral devices to “run the shop.”
"Generative AI has demonstrated the potential to significantly outperform human CEOs in strategic decision-making by excelling in data-driven tasks like product design and market optimization...While AI’s ability to analyze complex data sets and iterate rapidly could revolutionize corporate strategy, it lacks the intuition and foresight required to navigate black swan events." (Harvard Business Review)
Uber is investing currently in a two-tier system where the core that’s always at work is robotic vehicles and gig drivers handle overage. That injects quite a bit of flexibility into the business model. They can put self driving vehicles into market niches that will accept them (there’s difficult human factors issues with acceptance — even though the technology is already safer than human drivers) and use gig drivers everywhere else. But the catch is that the net is less human drivers over time, not more.
Eventually, I suppose that robotics will begin to chew into many sources of human labor as robotics continues to rapidly develop. Many robots will be highly specialized — not the “humanoid robot” so many imagine. A self-driving car is a specialized robot. Why not self driving trucks, trains, aircraft, etc? Eventually self-driving personal vehicles once acceptance is very solid.
 |
| San Jose–based Imagry has rolled out autonomous buses in Israel and has bids in to set up self-driving routes in five other countries |
It will be a completely different world and I can’t imagine any transition that doesn’t involve a lot of human drama. It’s easier to imagine dystopias than successes. Perhaps I’m being too pessimistic. But here, I’ve not bothered to mention what happens when AI becomes self-directing and has intentions that it can act on? What will be its motivation? Humans have a very long evolutionary lineage that underpins our motivations. AI’s are not human. What will they want and why?
I would expect that populists will go after this eventually — as soon as the impact begins to be squarely “felt” by the public. The populist message has often been aimed at immigrants “taking our jobs” or immigrants upending our “culture.”
The most recent yet mature version of the traditional immigrant labor debate is the beginning of the end of “offshoring” to cheap labor markets that Trump is exploiting as fodder to whip up his cult of personality. They don’t need to immigrate to the US to take those jobs. US corporations will ship the jobs to them.
The economic model that neo-liberals (e.g., Bill Clinton) foresaw and the end of the Cold War trade blocs and free trade delivered, inevitably undermined labor in the US and in other wealthy free nations. It was entirely foreseeable since there was an obvious and large gap in labor rates and in capture of externalities such as worker health and safety, and polluting emissions.
In negative externalities, twenty years ago, China was killing up to 100 times the number of coal miners per year in comparison to the US and other well regulated economies. I recall in about 2000 that it was a strange experience to travel to China and find the air breathable. Coal soot was everywhere and the air burned your eyes and nose. China has made significant progress since then, but a significant gap still remains. Regulation has real and positive results, but capturing externalities costs something. That’s money that doesn’t flow into the pockets of the wealthy.
Trump reintroduced broad trade protectionism to US foreign policy. Biden kept Trump tariffs on China. The predictable outcome of free trade was the gutting of the lower end of the income ladder and dismantling of production industry in America. Undermining industrial production undermined the labor movement far more effectively than “right to work laws,” etc.
At the same time, the specific GOP focus on culture war (the “basket of deplorables” that Hillary lamented and their obsession with “god, guns, and gays” that Obama mentioned) increasingly moved the former labor vote into GOP ideological territory. The GOP was no doubt thrilled. The neo-liberals captured urban centers, now bereft of manufacturing, and they fight a never ending battle with the GOP to define where the “line of control” will be in the suburbs. AI will be a whole new dimension in that debate. The culmination of automation has arrived.
If we don’t sort that out in some robust way, it seems inevitable that there will be a lot of trouble.
