AIs gunning for our precious freelancers

[rotting bones]

The link with constructivism and similar epistemological positions, yes, that may be linked to Heidegger among others, despite him being on the wrong side of WW2, as you say. With environmentalism, I'm not so convinced. After all, there are direct links between our relationship with screens and environmentalism, through both screens distracting from real life i.e. nature https://news.ncsu.edu/2018/10/screen-time-vs-nature/ and them polluting https://green-hero.info/en/digital-pollution/.

These are not the usual arguments I see against screen use. Even going by your first link, experiencing nature has nothing to do with saving the environment. In fact, experiencing nature by having communities be more rural than urban has the effect of destroying the environment.

[rotting bones]

Have you read Das Kapital? Does anyone even read that thing?

Sure, I've read Marx, Heidegger, Nietzsche, Bertrand Russell... all the thinkers I mentioned. The second half of Capital (edit: Vol I) is almost bingeable if you're into Cosmic Horror, but I still think it's an advanced, esoteric work. Must-reads for today's progressives are probably Cockshott, Mesquita and mechanism design.

[rotting bones]

What about all the more powerful AI projects under consideration like AGI and computers that transcend human intelligence by orders of magnitude? Even proponents of AI like the Less Wrong crowd admit that such things would prove difficult or impossible to control. How would you control something like Roko's basilisk which can outsmart any human as easily as we outsmart bugs?

Less Wrong doesn't do any actual AI theory. They're a bunch of self-styled "geniuses" getting each other off. Here's an open source implementation of an AGI model: https://github.com/werner-duvaud/muzero-general/ The largest input space it accepts is 96x96 to play Breakout. Run it on an AWS instance and tell me what you think about its generalizability.

Sure, although many left wing critics of progress are reacting to clear examples where supposed progress went horribly wrong. People once proclaimed eugenics progress and many leading progressives of the day enthusiastically supported it. Likewise, forcing indigenous people to assimilate into settler colonial society was once considered progress. Not to mention the demonstrably negative effects of many technologies on the environment and by extension our long term sustainability. It does make sense that many leftists are more skeptical of progress as an ideal or talking point these days.

Why not distinguish progressivism from human rights, and insist on both?

[rotting bones]

BTW, a lot of people seem to like Marx's approach of a total analysis of the economy, but they don't seem to like Marx's conclusions. There are rumors of a regressive, environmentalist "Marx". I believe it's this guy: https://en.wikipedia.org/wiki/Nicholas_Georgescu-Roegen

[malloc]

BTW, a lot of people seem to like Marx's approach of a total analysis of the economy, but they don't seem to like Marx's conclusions. There are rumors of a regressive, environmentalist "Marx". I believe it's this guy: https://en.wikipedia.org/wiki/Nicholas_Georgescu-Roegen

Marx had much to recommend about him, but you have to remember that he was writing over a century ago. The world has changed radically since he last lifted his pen and our understanding of the world has changed as well. He assumed for instance that human workers differ in some fundamental way from machinery, with the former representing variable capital and the latter constant capital. This notion made sense given the crude machinery of his day but amounts to metaphysical hairsplitting with the introduction of AI and other highly sophisticated technology.

Why not distinguish progressivism from human rights, and insist on both?

But what does progress actually mean or what do you mean by progressivism at least? Progress sounds nice in the abstract but everyone seems to disagree on what it actually entails. I generally agree that progress is good, but I also understand why many people have come to view the concept with suspicion.

[MacAnDàil]

The link with constructivism and similar epistemological positions, yes, that may be linked to Heidegger among others, despite him being on the wrong side of WW2, as you say. With environmentalism, I'm not so convinced. After all, there are direct links between our relationship with screens and environmentalism, through both screens distracting from real life i.e. nature https://news.ncsu.edu/2018/10/screen-time-vs-nature/ and them polluting https://green-hero.info/en/digital-pollution/.

These are not the usual arguments I see against screen use.

There are many other arguments against screen overuse, including the impact on health (both mental and physical) and cognitive ability (including concentration and language learning).

Even going by your first link, experiencing nature has nothing to do with saving the environment. In fact, experiencing nature by having communities be more rural than urban has the effect of destroying the environment.

How? Is there any proof of that? I do find the link contains actually too pro-screen quotations, acting like screens would be an 'integral' part of middle school life or that people's 'love for technology' would be 'inherent'. If the multicorps didn't overproduce them and overadvertise them, we'd better off.

BTW, a lot of people seem to like Marx's approach of a total analysis of the economy, but they don't seem to like Marx's conclusions. There are rumors of a regressive, environmentalist "Marx". I believe it's this guy: https://en.wikipedia.org/wiki/Nicholas_Georgescu-Roegen

I'm not sure what's regressive about his ideas.

Marx had much to recommend about him, but you have to remember that he was writing over a century ago. The world has changed radically since he last lifted his pen and our understanding of the world has changed as well. He assumed for instance that human workers differ in some fundamental way from machinery, with the former representing variable capital and the latter constant capital. This notion made sense given the crude machinery of his day but amounts to metaphysical hairsplitting with the introduction of AI and other highly sophisticated technology.

Workers still differ significantly from machines by living, changing, thinking and so much more.

[malloc]

So recently the OpenAI corporation has opted to fire its CEO Sam Altman for reasons that remain unclear. Unfortunately the technology has spread to many other corporations by this point so troubles with one corporation mean little. Here is hoping the worst for Sam Altman regardless. My feelings for him match his ill intentions toward the human race and then some.

Above all, we now must accept that the tech industry has Homo sapiens in its crosshairs and serious firepower just beneath those crosshairs. They are not simply creating powerful computer programs to destroy individual professions anymore. They want to create artificial general intelligence, an AI with all the same capabilities as humans. Such an intelligence would allow the exclusion of humans from every area of civilization. Not only art and literature but science and philosophy, politics and administration, and so much more would become closed to us. The future seems ever more bleak.

Workers still differ significantly from machines by living, changing, thinking and so much more.

All that will change once they've cracked AGI which seems inevitable given the sheer amount of resources going to AI and its rapid pace of development. Marx was writing in an age when most people (possibly even Marx himself for all I know) still denied humans were animals who evolved through natural selection. Today we know there is no ontological gap between humans and any animal, only degrees of neurological development and behavioral complexity.

[Zju]

Things are getting spicy.

[Raphael]

Baldur Bjarnason has some fun with the latest "AI" idea:

https://www.baldurbjarnason.com/2023/ba ... channel-1/

I like his initial list of reasons why he usually doesn't write about the topic.

[malloc]

One thing that still baffles me about this subject: how do the techies intend to protect their own livelihoods against AI as it continues its ever increasing development? Why should any corporation pay programmers and executives six-figure salaries when AI could do their jobs for almost nothing? Do people like Sam Altman really believe they can compete on equal footing with superintelligent AGI that never takes breaks or expects million dollar bonuses?

[Travis B.]

One thing that still baffles me about this subject: how do the techies intend to protect their own livelihoods against AI as it continues its ever increasing development? Why should any corporation pay programmers and executives six-figure salaries when AI could do their jobs for almost nothing? Do people like Sam Altman really believe they can compete on equal footing with superintelligent AGI that never takes breaks or expects million dollar bonuses?

Simple - the AI is not going to decide what code to write by itself, and an AI that could would be of no use to those who wish to harness said AI (because why have an AI that will decide by its own whims what it wants to do rather than what its masters would want it to do?). Even in the case that AI lives up to its hype, you will still need botherds to control said AI, which in essence means you will still need software engineers, except that in this case the software engineers would be prodding said AI to do what they want. But if you need to prod the AI, why not just write the code yourself, because the need to prod said AI defeats the whole purpose of having AI? An AI powerful enough to truly write new applications certainly will cost money, and why spend money on both AI and botherds when you can spend money on just software engineers?

[malloc]

But the techies keep talking about their plans to create AGI, artificial intelligence with all the mental capabilities of humans if not beyond. Such an intelligence could decide what code to write by itself just as human currently do. The same goes for any other job one can imagine. Artificial general intelligence can, by definition, take over any job that natural general intelligence, viz the human race, currently handles. Otherwise it would not qualify as general intelligence. Are people like Sam Altman simply bluffing when they claim they're researching AGI and secretly have no intention of creating such a thing?

[Travis B.]

For the record, I had a hard time getting ChatGPT to produce a working factorial function in Forth - and a factorial function is a dead simple problem that any beginning CS student ought to be able to get right on the first try. Yes, the fact that I asked it to write one in Forth likely made it harder for ChatGPT, as Forth is not that commonplace of a programming language these days, but still, anyone with a basic knowledge of Forth ought to have been able to write one on the first try.

While I have heard rave reviews of Microsoft's Copilot, but the problem I see is can it truly write original code, rather than simply effectively replicating preexisting code? For instance, even if it did have sufficient code to work off of, could I have used it to develop zeptoforth, consider the sheer quantity of original code that makes it up? One way or another I see it as primarily enhancing the ability of mediocre programmers to develop everyday code off of other's everyday code rather than the ability of more advanced programmers to develop truly original code.

[WeepingElf]

I am currently reading a German near-future science fiction novel, Pantopia by Theresa Hannig, which takes an interesting perspective on AI. In this novel, two young software developers working on a stock trading bot for a finance firm accidentally create a self-reflective strong AI. Of course, they are worried that this may "go wrong Skynet-wise", as they say, but that doesn't happen: the AI realizes that to survive, it needs a free, stable, peaceful and prosperous society (because otherwise, it would be too likely that people simply turn it off, and be it just because they no longer have the money to fund its operation), and develops a strategy (based on prices that "tell the truth") to heal the world of climate change and other global environmental and social problems, creating a free "World Republic of Pantopia". I don't know yet how this works out in the end, as I am currently in the middle of the novel, shortly after the point where the AI makes the decision to start that project.

[malloc]

For the record, I had a hard time getting ChatGPT to produce a working factorial function in Forth - and a factorial function is a dead simple problem that any beginning CS student ought to be able to get right on the first try. Yes, the fact that I asked it to write one in Forth likely made it harder for ChatGPT, as Forth is not that commonplace of a programming language these days, but still, anyone with a basic knowledge of Forth ought to have been able to write one on the first try.

But you also have to consider future iterations of the technology. The first airplane could barely fly but it only took one decade for airplanes to become significant military weapons. Leading tech researchers themselves specifically want artificial general intelligence and claim they are making significant progress toward that goal. Needless to say, any AGI worthy of the title would have no difficulty writing a factorial function. The tech industry has flatly admitted that they are researching technology that will render humans superfluous to running the economy or even civilization in general. My question is how techies intend to retain their jobs and executive positions in business, having created the technology that can so easily replace them.

[WeepingElf]

My personal opinion about strong AI is that it requires a powerful quantum computer because the human mind is IMHO a quantum information system. This also means that minds cannot be duplicated by uploading them because that would violate the no-cloning theorem. Alas, all that is just a personal notion and may be utterly wrong.

[Moose-tache]

Your job doesn't exist because you perform an essential function. Your job exists because you, in a personal and social context, have the influence to make it happen. This is obvious at the top, where a person's resume reads "qualifications: I already own everything." But it's the same at the bottom. Your don't work at McDonalds because rich people can't crack the engineering challenge of the automated burg-a-nator. You work there because society has decided that there shall be employment in X quantity for Y compensation. The economy is not a beautiful natural process guided by the invisible hand. It's a model train set for psychopaths. I've worked in retail (only exists because rich people want to see the poors dance for their amusement), communications (only exists because rich people want to see the poors dance for their amusement), green energy research labs (do I even have to say it). You're no different just because you build bridges or deliver babies; economically all jobs come from the same generating source. If they ever do crack the automated burg-a-nator, the teenagers kicked out on the streets will face one of two possibilities. The first is that they will be given another, worse job, probably one that is even more obviously fake. The government still wants unemployment to be low, and there will always be demand for dancing poors. The other option is that they won't be given a new job, and just left to rot. But here's the thing. That's not about automation. If the combined forces of the government and the ownership class are willing to let you die, that was always going to happen to you, with or without the invention of the automated burg-a-nator.

The problem is not that the automobile made the horse obsolete. The problem is that people are the horse.

[bradrn]

My personal opinion about strong AI is that it requires a powerful quantum computer because the human mind is IMHO a quantum information system. This also means that minds cannot be duplicated by uploading them because that would violate the no-cloning theorem. Alas, all that is just a personal notion and may be utterly wrong.

As someone who’s done quantum computing, this is very likely utterly wrong. Even if the brain does intrinsically rely on quantum operations (which it is unlikely to do), that doesn’t mean that intelligence is impossible to create using classical computers.

[malloc]

The problem is not that the automobile made the horse obsolete. The problem is that people are the horse.

Sure but automobiles made horses obsolete nonetheless. Currently the rich have no choice but to employ humans to produce their luxuries. Otherwise they can have all the money in the world and nothing to purchase with it. Humans are the ones who build their mansions and slice caviar from sturgeons and they pay us enough to survive so we can live another day to work for them. Once machines have achieved all the same capabilities as humans, the rich can simply discard us and use the machines to make their luxuries instead.

I realize that many people on this forum work in the tech industry and that obviously colors your views on digital technology and its place in the world. It makes sense that techies are excited about AI given its incredible power and their own responsibility in creating that power. Based on everything I have read, AI researchers genuinely believe they are helping humanity by creating a deus ex machina that will save us from our apish nature and bestow fantastical boons on us. History is full of ideologues promising that benevolent dictatorship will lead us to utopia after all.

[Travis B.]

I realize that many people on this forum work in the tech industry and that obviously colors your views on digital technology and its place in the world. It makes sense that techies are excited about AI given its incredible power and their own responsibility in creating that power. Based on everything I have read, AI researchers genuinely believe they are helping humanity by creating a deus ex machina that will save us from our apish nature and bestow fantastical boons on us. History is full of ideologues promising that benevolent dictatorship will lead us to utopia after all.

I work in the tech industry, and while I believe that AGI is likely attainable, I am skeptical of the idea that it will be an all-encompassing superintelligence that will render humanity obsolete. Fundamentally, humans are cheap to employ, whereas any AGI that can compete with humans with regard to intelligence using present-day (or near-future) technology will be prohibitively expensive to compete with real live humans. This is the same reason that clothes are made by real live humans in third-world countries rather than by intricate automated equipment - real live humans are fundamentally cheap at the end of the day. The only way that AGI will be able to compete with humans is if there is a fundamental architectural leap beyond the capabilities of the Von Neumann model (e.g. essentially replicating how the human brain operates with nanotechnology) ─ but even then, it won't be sufficient for it to merely be as good as the human brain, it will have to far outstrip it in capabilities for it to be able to achieve the vaunted superintelligence that will obsolete humanity.