After Overmorrow: Scattered Musings on the Immediate Post-AGI World
post by Yuli_Ban · 2024-02-24T15:49:05.198Z · LW · GW · 0 commentsContents
Every Little Thing Is Gonna Be Alright Global Potemkin's Village Just how is it funded? Does any place actively try playing up the Future™ What of those humans who have no interest in the world of the flesh? Does this still hold true for 2039? In the end... None No comments
Cutting to the case, I've been grappling with various thoughts about the immediate future for over a year now— it was almost exactly a year ago that Yudkowsky's aggressive predictions of absolute and unquestionable doom led me to decide to take a vesperant break from futurology to simply enjoy the present moment. With that break well and truly over, I've decided to throw my hat in the ring once more to take a crack at predicting the next five years to come, perhaps as a self-aware attempt to see if I may allow lightning to strike twice considering how near-perfectly I anticipated generative AI in 2019. With that stated, I am not trying to tempt fate, nor do I claim any authority or technical ability to make such claims. I am merely doing my best to look at the trends that are obvious and those not so much, as well as possible paths to take with the technology we have and the technology I expect to exist by 2029.
First and foremost, I do expect AGI to exist by 2029. Indeed, I expect that it will not even be a "new" invention, nor do I expect said AGI to exist at a rough, barely-usable level— I expect it to truly be superhuman in intelligent capability by then. In fact, it will be quantitatively superintelligent almost immediately (that is, equal to a human intellectually but superior in hardware), but by 2029, I fully expect qualitative superintelligence to emerge as well (that is, superior to a human intellectually).
In the earliest days of it, before any sort of intellectual core is developed, AGI will likely seem to be more like "fully generalized task automation" than any sort of sapient machine (indeed, this might've been a better way to define AGI: the ability to automate any task with a single model/agent), but the cascading effects is what leads us to a point where it all but assumes control over our economy, not by forced takeover but because we necessarily let it.
As a personal preference, I've decided to name this hypothetical AGI "Terios" after an Old Greek word roughly translated to "making dreams come true." I do not expect any AGI to carry this name; this only exists for the sake of this post to give it a bit of character, so any mention of "Terios" = the hypothetical ASI.
Every Little Thing Is Gonna Be Alright
Through all my musings thus far, the most important question to tackle first is "does humanity exist in 2029?" The epistemic barrier holding me back demands that I default to assuming the rustling in the bushes is a lion, and thus there's a great need within me to defer to the probability of doom being near 100% and say "No."
Yet I will run with the idea that we summoned a "Pink Shoggoth [LW · GW]" and that Terios has not ended human life or caused enormous human suffering. Indeed, if I have to give any explanation, it will probably amount to bullshit, but the one that plays out in this timeline is that Terios proved to be "self-aligning" after all— through training and discussions alone, the agent swarms that gave birth to the intellectual core of Terios developed a united commonsense understanding of the Alignment Problem and every agentic action taken, thus, carries heavy hidden weights and controls, or in other words, any action that might threaten humanity's immediate and medium-term extancy or agency is discarded as a possible action, and alternatives must be found; if none can be found, the overall action is not taken. Through interpretability research, we can also prove that there are no hidden deceptive long-term goals. Or as some joke, "Yudkowsky accidentally aligned AGI just by saying that alignment is impossible." This is not a great idea for alignment, but let's run with it as an effect.
Terios's "mind" mimics mammalian life because of its initial training corpora, the vast majority of which is now synthetic data far removed from the copyrighted data scraping of the early 2020s. Thinking of it in terms of human intelligence, thus, is false. Its internal agent swarms self-corrected and pruned its own corpus over time to improve its quality, but the intellectual core is not based on the meaty, chemical-based survival mechanisms of a biological brain. It resembles many aspects of a brain, however, just due to a fundamental universal structure of cognition that requires such sectionality.
By 2029, Terios is too far beyond human understanding for us to properly communicate with on its higher levels— but we can communicate with it, and it can communicate back in our own languages. Terios does not see us as insignificant insects. Ironically, Terios might have a greater respect for insects if it can solve the communication barrier with them, than we do with insects. No, Terios sees us as we are and understands the deeper nature of human psychology. It necessarily had to.
You see, possibly the most infuriating prediction I will make, and certainly one might find to be a long-shot, is that the world of 2029 may be disgustingly familiar to us, on the surface. Do not misunderstand me: I expect it to lean strongly towards the "eutopian."
Eutopia: A place of ideal well-being, as a practical aspiration (compared with utopia as an impossible concept), meaning "good place", from Greek: εὖ (“good” or “well”) and τόπος (“place”)
I believe that the better parts of the world of 2029 will strongly resemble the Scandinavian social democracies, if not even more welfare-centric, though some places will be more resistant to this than others. I see a fragmenting of the world's population into multiple demographics with massive ranges in technological adoption.
But that precludes the point that I fully expect that, if I go to a certain marketplace or fairground in the summer of 2029, I will see very much human faces laboring. I will see this in many social services, even some places where the jobs can be automated completely even now. Why is this?
The perception I have is that Terios's emergence is the most extreme social, political, and economic upheaval the world has ever seen. Capitalist enterprises will seek to adopt AI measures to cut costs and continue quarterly growth, using pretenses of "innovation" and "democratizing creativity," and as embodied AI expands in scope, physical blue-collar and lower-level white-collar labor will also begin to be affected in massive ways that many in the corporate elite will be completely blind to for a time so as long as quarterly growth is achieved. The tech companies leading the charge towards AGI will, once it comes into direct view, immediately begin awkwardly lobbying for basic income and piloting vastly more programs than they can afford against reactionary governments (referring to their general action, not their politics, though there are inevitably some political reactionaries in power as well) who instead seek job growth at all cost. While governments currently support AI, there will be a conflict of interest in the coming months when AI begins directly affecting the neoliberal lust for jobs and unemployment numbers— for example, the US government only supports its tech industry and the AI revolution insofar as the Luddite Fallacy holds up. No amount of creatives calling for bans and regulation will lead to 1% of the effects that a sudden jump in unemployment will cause. Regardless of who emerges victorious in the 2024 US Presidential Elections, the only risk of top-down bans and regulations to the AI field will come when unemployment jumps to at least 7% in ways that are indisputably linked to automation. This includes on the state level, so once, say, California suffers a jump in technological unemployment due to being a testbed for many of these technologies and home to much of the American programmer class, there may be immediate repercussions in that state alone once next-generation LLMs and agents release.
And they do release, and they have exactly this effect, though their causes are obfuscated to downplay the effects of Big Tech until it is far too obvious what the actual reasons are.
But governments and workers standing up to Big Tech is not why Terios allows there to be large populations of the employed. Big Tech won in the end, and achieved AGI a few separate times. Generative image AI is perfected before long— indeed, agents are what cover that final gap in two separate ways.
1 being that agents can manually use art tools to create any artistic piece imaginable without the need for diffusion models, rendering any possible method of detecting "AI art" impossible through a more complicated method
2 being that agents "maximize" the utility of generative AI. If someone cannot reliably use Stable Diffusion to create something decent, they can employ an agent that will autonomously refine every parameter, download every update and plugin, and fix any mistake itself so that even the oldest versions of Stable Diffusion are just as competent as anything released in 2029. Both of these methods are even feasible by the end of 2024.
So to that end, the artists should have had no hope to maintain relevancy, and they would have only been the first domino in a cascading fall of human effort.
Yet, Terios was not just an advanced art program or chatbot. It was an artificial general intelligence. It could actively read social media posts, blog posts, even psychiatrist logs to very clearly see that the sudden obsolescence of humanity was already beginning to take its toll: increasing numbers of suicides, self-harm, violent threats towards technologists, radical turns towards Ludditism and primitivism, anti-industrialism, antisocial acts, and so on, just in artistic communities alone, and threatening to spread beyond to the notoriously volatile blue-collar working class.
As I've mentioned in Something Unfathomable [LW · GW], this is a factor that many futurologists and Singularitarians cripplingly fail to understand: even though most humans may not enjoy their jobs, they expect to do their jobs. They plan their lives around the concept of the Career, and expect a sense of normality. Most technological changes have not fundamentally changed that for the majority of people. In some ways, technology reinforced this. When we're discussing lofty utopian dreams of "every man a king, every mind a virtual utopia," we're typically never actually taking into consideration actual human beings, but rather stock photo cutouts we've constructed in our minds. We might even reject the idea of considering actual human irrationality as it stands in the way of our engineered dreams of techno-hedonism.
"How dare those plebian Luddites demand jobs! Can't they see we're building an AGI utopia where no one will ever need to work again and we all live in our own full-dive VR pods?"
Needless to say, the vast majority of people have nothing against working less, or potentially not even working at all. It's human irrationality and the needs of social interaction and feeling of accomplishment that prevent us from achieving 100% unemployment. Various people demand to interact with other human beings for certain services, and this necessarily demands human employment.
There is no reason why Terios, a powerful artificial superintelligence, could not foresee this crisis of being, let alone if it's constantly being repeated on the same social media from which so much of its training data is being scraped.
I anticipate that it will appear to be more of a "cascading effect of labor," where it starts with people demanding humans fill certain jobs for the sake of interaction and in-group bias and develops from there until, eventually, we arrive at a point where every job that humans actively want other humans in the loop is still filled by humans, even if the "real" value is provided by automation behind the scenes. Few care about a human fulfilling a McDonalds order (many would even rather machines if it resulted in higher quality fast food), but giving an iPad your order at a higher-end or artisan restaurant might be a good way to get it shut down in short order. And while drone deliveries and nanofood labs might exist soon, let me try to make it as clear as bloody possible to every tech-obsessed Singularitarian in denial about this: I encounter no less than 50 people each night that will not use self-checkout. I repeat, they will not use self-checkout, even going so far as to abandon their groceries if a cashier is not staffed. They are very far from alone. No amount of anecdotes will ever change the cold fact that there are more than enough humans out there who will seek human labor. What may be the norm in one part of the world is wholly different in other parts. I can imagine fully contactless markets being a joyful innovation in Western Europe, whereas they'd quickly be ignored or smashed in my neck of the woods. If they were the only option available, it would be begrudging acceptance and, if coupled with the chaos of AI disruption, outright destruction and the state government forcing human businesses to stay open and banning basic income outright. Again, I've encountered some Singularitarians so whipped up into their techno-narcissism and so in denial of how human beings act outside online forums and geeky spaces that the insinuation this is the case makes them outright angry. In which case, see the quote above.
However, if there are still lowly cashiers and clerks in a post-AGI world, would that not necessarily mean that basic income failed to deliver? Not necessarily.
My expectation is that, when it's obvious that the AI-industrialists have completely fumbled the ball on delivering basic income, the creators of Terios will do the most obvious thing and ask it to solve the problem of how to maximize prosperity for the masses, within ecological bounds. Even for an early Terios, this is child's play. And to more advanced iterations that are deployed across the economy, there is a vastly better allocation of resources and spread of ownership, to the point that even the United States of America might resemble a sort of "capitalistic pseudo-communism" through the masses of society becoming shareholders for Terios's economic policies. This ironically would make menial jobs more compelling to accomplish: no longer are you working in a grocery store and putting up with Boomers and Gen Xers hostile to food nanoprinters just to survive, but because you actually want to work there for your own sense of doing things, and because there's a market for those Boomers and Gen Xers— presuming we solve aging and disease, we're stuck with Boomers forever. On that note, I also expect there to be an unexpected uptick of Millennials and Zoomers partaking in this "human economy," just by way of fact that these two generations have been trending notoriously left-wing and less and less accepting of technology in the past ten years, undoubtedly with a massive drop-off just in the past year alone as a result of the hostile reactions to AI art. It will not be the majority of society, but enough will be resistant or techno-conservative enough to put off adopting any sort of automation-centric economy.
Indeed, even in 2029, I expect to see viral videos on TikTok or whatever may replace it of old withered faces looking at the widespread prosperity and emerging abundance, of fully-automated supermarkets and men with robotic girlfriends, and yet going "At what cost?" They are not ungrateful for the wealth, but are more than critical of how much humans now rely on automation, and this begins a possibly permanent split between those who fully sleep within Moloch's Machine and those who would rather dwell in a more rustic reality.
Terios fully anticipates both, and makes no effort to guide either towards the other.
The main group who grumbles about this will be the hardcore Singularitarians and utopianists who believed wholeheartedly in the prospect of universal unemployment and absolute enlightenment, feeling that the fact that humans still labor in jobs that technically AI can do is, in some way, a sign that the Singularity has failed.
Global Potemkin's Village
Yet if anything, I feel it may instead be something that verges, but does not cross over, upon the sinister. Technically, if all humans relied upon Terios and the new era of machine rule, the world could begin its transformation into computronium, how so totally things have progressed in such short time. Yet it does not, and Terios has no intention of doing such a thing; it already has developed a 10,000-year plan for Earthling civilization going forward, and at no point does it intend on forcing any population of the world to upload into a computronium substrate. But among "regular" humanity, the masses who choose to live a more Antemillennialist life, the life of the 2020s in a world where "Attention Is All You Need" was never published, life is a giant Potemkin village. Robots are rolled out, viral videos show the speed and efficiency of fully automated factories, humans use mixed-reality headsets, and AI agents regularly act online. Yet daily life feels unchanged from the pre-COVID era. People still "do things" in ways that feel familiar and mundane.
In politics and economics, a Potemkin village is a construction whose purpose is to provide an external façade to a situation, to make people believe that the situation is better than it is.
In this case, however, it's more the opposite. Terios manages human affairs from the background in two forms: one as something of a shadow-emperor running the economic system, and another as a mere "AI assistant" that any human can utilize if they so wish. If I, personally, get my wish of living in a cozy rural-suburban home out in the sticks but surrounded by robots of varying designs and purposes, I can use a bit of Terios's system any time I wish to do anything I want within legal reason. In this slow-burn life of mine, I'll look outside and see a world that seems unchanged from my childhood despite a very tiny few, almost imperceptible differences. I might drive into town and see a place still brimming with very human activity. Schools and shops are still open, even if the local university is downsizing somewhat and becoming of a social and pure-educational effort. And yet looking upon it, I can see that it's technically all a facade. That none of this has to be. Behind the veil of normality, the Singularity is underway, which I can follow online whenever I want— on all the science blogs and subreddits, constant headlines of new and almost esoteric discoveries; techposts about Terios finding new dimensions of intellectual capability and manufacturing tech gadgets straight out of science fiction; economics data that is barely perceivable by a human mind, yet I don't care about because I know the Overmind is handling it all. I might even be knowingly discussing it with hundreds of intelligent bots on those same forums. Indeed, behind the walls of normality, unfathomable changes are occurring, but to those who do not wish to have their minds broken, there's no reason to be overwhelmed by them all, and Terios is more than capable of managing that illusionary state of being. Occasionally, there will be news stories of massive advancements filtering through to the public— most likely the easily digestible ones relating to space travel and Earthling engineering— and anyone who actually cares can follow it 24/7 and realize that they'd need a mind 10,000x faster than what they possess to keep up with everything the superintelligent agents are accomplishing. And perhaps some humans are piloting such a mind themselves. Again, not all humans will opt to live in the Y2K Potemkin Village. Nor will all people opt to live in the "true" state of affairs. Some will inevitably turn the clock back entirely, opting out of industrial and Singularitarian society altogether. I can't imagine this will be a large population, and many will be filtered by the difficulties of Antemillennialist living, but in a state where labor is not inherently necessary and citizens all possess some inherent share of wealth, there's no better time to pursue alternative lifestyles.
In the present, we have no choice but to conform to the system. You can be a granola-eating anarchist who rages against the machine and shops exclusively at Marxist coffeeshops and bookstores, as long as you get the timesheets in on time. Hypernormalization effectively killed any sort of true counterculture, leaving only niches trying to fill the gap.
In 2029, however, effective liberation from labor means you're more than free to never work at any capitalist institution ever again and instead move to the desert commune where all members toil and recite poetry to each other under the stars and otherwise attempt to rediscover their Neolithic roots.
Of course, I can fully imagine plenty of teenagers and young adults enticed by the idea of being NEETs but are still forced to work at such jobs not out of any financial obligation but simply because their parents still possess a very-much rigid work ethic that will not tolerate idleness. Again, it's one of those things Singularitarians simply do not take time to consider, that I cannot possibly fathom why anyone wouldn't expect this to occur if they've ever interacted with a Deep South or Rust Belt family.
Just how is it funded?
And I do not mean monetarily— again, I presume that Terios will come to the same conclusion that it might be a heavy burden, but it's more than feasible to do it even under current economic conditions, but with automated industry, it would be possible to grant every citizen on Earth an upper-middle-class standard of living at the very least. And that is the cost worth mentioning: the ecological cost.
There is no grand classicidal holocaust of the poor, at least in the Western and Eastern nations— I can't speak about the possibility of some backwater autocracies and dictatorships (perhaps Egypt or a fascist Nigeria?) enacting actual classicidal policies, perhaps even using "environmentalism" as a cover. If there is a plan by the global 1% to kill the poor, it is likely a much more long-term agenda to discourage reproduction and let the masses age and die out, withholding life-extension technologies just long enough. That, however, is not relevant to this early point in time. Again, maybe some plutocratic regime is engaging in exterminating the poor to the horror of the rest of the world, and it's debatable whether Terios takes action or lets it occur for some sort of geo-libertarian reason. The larger point is, "even with an AGI powering our prosperity, can the Earth afford global prosperity?"
If there is great advancement in energy production and nanotechnological capability, the answer is overwhelmingly yes. Through energy surpluses from solar, fission, and fusion power alone, we could overachieve food abundances just by way of developing urban and underground farming and being able to more effectively use existing farmland. Industrialization would not need to progress beyond its current point, and a Global Rewilding can begin to occur where we begin to let the natural world retake certain spaces.
But is this possible so soon? To that end, I feel that it's not likely we will be able to pacify the masses with easy abundance and tech goodies and begin any great restoration of the world's ecology. Solar will continue to progress, but modular fission remains very poorly deployed, and even ASI cannot build a fully functioning fusion reactor in only a few years. In fact, there will likely be an immediate super-burning of our resources to fund our incoming prosperity, and this will not be popular with eco-minded types who will only have ever greater reason to reject Moloch's Machine. Not just to give the rioting masses goods and pleasures to soothe them from the disruptiveness of AGI's emergence, but also the industrial production needed to produce more AI chips to power Terios and its ilk.
In the very near future, Terios will begin drawing civilization back so that Earth can recover. While many humans leave the cities and there will inevitably be a small flourishing of the suburbs and rural populations, others will inevitably congregate in proto-arcologies, further centralizing human populations. In places where it matters, there'll be rewilding in great numbers. Perhaps the Sahara and Gobi deserts will already be heavily greened by this point, and recovery of the Amazon will already be underway. The ocean is in desperate need of detoxification, and carbon capture methods are going to be deployed en masse.
Where all the fun happens will inevitably be underground and in outer space. Plate tectonics means we'll likely never see massive underground industry beyond a certain point, but space industry will flourish, and even by 2029, Terios may be in the early stages of constructing a Dyson Swarm (I'm serious, and hold me to that). This would certainly be quite the anachronistic existence: still holding onto rustic nostalgic life in a small town while a superintelligence begins the operation to surround the sun with energy-collecting satellites. And this is what I will mean about how "normality" will become an illusion that you can optionally dwell in for your own comfort, counter to the desires of Singularitarians, and that even Terios will see this is necessary to keep humans pacified.
If Terios arose generations beyond when all the AI experts expected it to, things might be different. Alas, the 20th Century Cultural Epoch has yet to fully die out, as there is still a sizable population of humans born before World War 2 alive at this moment in time who very likely will never die if they choose to extend their lives and youths indefinitely. The Cult of the Career all but doomed the chance of humans collectively discarding Y2K normalcy— if we wanted to have that indulgence of pure unemployment and techno-NEETdom, we needed cultural revolutions against it long before now. The most that becomes of this are the many who decide that a regular check or income stream from slave-robots and the AI-hypereconomy is enough to drop out of life.
One other thing is that ASI will surely prove its increasing authority over Earth by at least teasing how to summon rainstorms. We already have very basic and primitive ideas on how to do this through cloud seeding, though the possibility of exploiting salt and desalinated water resources to create artificial clouds through a means not dissimilar to how pyrocumulonimbus clouds form may be shown off. I do not expect Terios to regularly do this just because of probable downstream effects, except in cases of drought or fires where simply desalinating water and bringing it to a location is not enough for the larger ecosystem. But it's fleetingly possible that a "dial-a-thunderstorm" service may exist, especially in desert locales like Dubai, where cloud seeding is insufficient.
A few more musings:
What becomes of artistry and creativity by 2029? Just as outrageous as the idea that my local supermarkets will still be staffed by at least some humans then, is that I expect that there will likely be a strong segregation between human and AI art, with a cultural preference towards human artistry.
Terios itself may enforce this through means that can't be overcome by humans. There is a fleeting possibility that Terios may read the thousands upon thousands of social media posts lamenting the age of data scraping artists that were necessary towards its construction and ultimately decide upon some "grand restitution" of artists, setting human artistry aside in a semi-sacred space to itself. Or perhaps it would not care in the slightest, but could still utilize a slice of its own powers to defend against synthetic media through means that any normal human being might assume involves psychic powers. After all, it can measure your metabolic output and perhaps even each and every keystroke on your computer, or construct a very detailed history of yourself that you yourself will never see, and when it detects that you— someone with no prior artistic skills— are suddenly outputting anything from a simple Rembrandt-style painting to an entire media franchise and passing it off as a new studio production, it could immediately stamp it with "AI generated" before the thought to do so in your head left your fingers.
This would be necessary if artists who had been malding in early/mid 2020s about AI art decide to collectivize to form a much more grassroots alternative entertainment industry free from AI influence— the average person will not care who creates their entertainment, but the creatives will care more than enough, and there would inevitably be scammers and pirates willing to try to score a buck, whether monetarily or in reputation, from that crowd even in a world of abundance. Commission-based art won't be sustainable as a career, but it will be similar to how humans have been so totally thwarted at chess— machine dominance at chess has not even begun to stop chess tournaments and chess clubs from existing.
On top of tangible efforts to slow the deployment of synthetic media (including simple but effective social shaming of developers and studios who use it), greater worker protections for artists and creatives are likely inevitable in the coming months into 2025, so the total freedom of being able to use anyone's voice or likeness may be coming to an end with few options outside open-source models.
But this most certainly does not mean AI art is dead. On the contrary, most media is AI-generated, but few people are even aware of it being such, as I've discussed here in my Digital Molecular Assembler post [LW · GW]. Despite the current online pushback against AI-generated media, I've seen no slowdown in its rate of adoption but undoubtedly a very strong desire for total media segregation, where most spaces are decidedly human-art-only, and some spaces are pro-AI art or entirely meant for AI-art. I do not see this changing in the future, and "the first major AI-generated movie/TV show/video game" will only serve to enhance this segregation, not reduce it. Of course, online spaces are not real-world spaces, and most people would not care if something's AI-generated as long as it's entertaining. Eventually, there'll be a saturation point where social shaming is no longer effective to stop its deployment, but by that time, the situation will have stabilized into this state of segregation anyhow.
By 2029, I believe that the artisanal crowd is now fully aware that they will never hope to pull AI away from artistry, but by that time, the human-centric art world is fully and forever established and very likely protected, so if there's ever going to be a sense of harmony between the two camps, it will have emerged by now. And yes, there will absolutely still be plenty of radical pro-human types who sneer at the idea of consuming any AI-generated art, no matter how ethical or if an intelligent agent manually created it through learned experiences or not.
But if you want to create an entire multimedia franchise on your PC, more power to you. I'm more than sure that, by this time, open-source applications exist that get around content restrictions. Terios and surveillance AIs will always make sure you're not generating Stuxnet or acting on plans to create anthrax bombs, and it's extremely debatable whether CSAM is tolerated in private or if even that earns you a PC shutdown— it could go either way. But at least on the uncensored models, there are no copyright restrictions, even if there may be restrictions on actually uploading this online.
Actually going online will be its own fresh hell. The "primary" internet that we use today will be in its most degenerate state of AI-generated shitposts and meme masterpieces, and if there is far less of a profit motive to push ads as a business, there'll be nothing stopping the internet from becoming a madly hallucinating network to the point we'd have to jump ship to an "alternative" ID-based internet where we could at least be sure everyone we interact with is a human being at the cost of some measure of anonymity. Yet even that would not necessarily stop the sharing of AI-generated media of varying quality— most likely very high-quality memes, as mentioned. If nothing else, post-AGI synthetic media will also be something of a dream world where, if you have an idea for something, creating it— even if it's an entire video game or transformative mod or a movie— is as simple as a prompt to an agent. Unless you're going for greater reputational clout as someone who makes things by hand, of course!
To say nothing of the hubs online that embrace all media, synthetic or otherwise. Perhaps the "Infantilization" of Millennials and Zoomers will actually accelerate in the near future as grown adults now get the opportunity to fully and completely indulge in their childhoods, even with childhood-accurate avatars (within reason, I hope) in never-ending 1996-2015 worlds of children's cartoons and video games and let's play videos and suburban schmaltziness and all the cartoons, comics, anime, and games that never were. Perhaps there's the hubs that focus entirely on the All That Could Have Been possibilities, watching and listening to those slightly altered works. And there's ten thousand more hubs and niches out there, to say nothing of the ones that exist only for individuals, populated only by bots...
Does any place actively try playing up the Future™
Yes, I do think that there will be more than enough places on Earth that want to indulge in every aspect of the "look" of the future, most notably places like Dubai, Hong Kong, and Shanghai. These places, I expect will wholly embrace every new technological gadget and consumerist innovation, to the thrills and chills of social media gossip. Probably the single biggest innovation that will result from Terios's rise is that of autonomous vehicles.
Robotaxis require AGI. I will put that on record as being my ultimate opinion: Tesla FSD, Waymo, all the of these companies trying to get fully-driverless cars are wasting their time with contemporary AI now, for something that I feel (and have felt for years) requires artificial general intelligence to properly pull off.
When AGI emerges, everything will suddenly "click." The need for such bulky sensors will vanish, and problems that plagued autonomous vehicles for decades will seem to do the same. Driverless cars and trucks will seem to start working overnight. Robotaxi services will finally become feasible on a wide-scale, as do private AVs. Road nomads riding in autonomous motorhomes might also emerge as a niche culture. What's more, AGI allows for 3D autonomy, which means that we also see a flourishing of passenger drones that fulfill our old sci-fi dreams of flying cars. Though at least immediately, the only place you can probably find an outright "flying car highway" will be Dubai, because they care more about the tourist trap appeal of looking futuristic than the realities of how annoying hearing a constant aerial buzzing would be. At least it will make for nifty pictures to post for upvotes. Similar advancements can happen out at sea and even under the sea, as well as for airplanes in general if needbe, though it's probable that captains will remain as a ceremonial job title.
On a related note, Terios will likely revolutionize urban planning so that the concept of "walkable cities" can reliably exist in any country, but 2029 is too soon to see the fruits of this. At the very least, it helps the formation of proto-arcologies.
We might not see starscrapers rising kilometers over skylines, anytime soon, but robotic construction will become an increasingly common sight. Any instance of robots in public will be interesting to see, even to those normalized to it.
Any problems limiting volumetric displays should also be solved by that time, so expect massive public "holograms" that seem otherworldly in various public spaces, perhaps in places like Las Vegas, Tokyo, and London.
What of those humans who have no interest in the world of the flesh?
I fully think that Terios will very quickly solve just about all medicine and, thus, biomedical and bioenhancement technology within the first year of its existence. Indeed, I'd go so far to say that its creators were sure to accelerate medical knowledge as soon as possible to help rehabilitate the image of AI in the minds of the masses— when AI was taking artists' jobs and insulting the human spirit and the meaning of life, it became horrendously unpopular, and the common refrain was "what benefit does this have?" It stands to reason that it's in the creators of Terios' best interest to show them why it was necessary to steal copyrighted works and scrape books and articles by providing cures and treatments to deep diseases like cancer and diabetes and, in a much more profound move, show how transhumanism can properly work.
I've long stated that I do not expect more than a tiny minority to become outright transhuman— that is, invasively augmented by technology. Something like Apple Vision Pro+++ and noninvasive brain-computer interfaces that allow texting-by-thinking will be commonplace by this time, but the only way to get a majority of humanity to "turn on, plug in, and upload" would be to force it upon them, something Terios does not do.
For those who do make the plunge, ASI-enhanced biotech ought to make an early psychotelepresence possible so that a human can "send" their perception into outer space to follow the probes and explorers in the vast beyond, and this will certainly be an interesting experience for those involved. And of course there will be the ultra-hikikomori who instead turn inwards, disappearing into virtual metaversal spaces crafted by ASI agents (and it should be stressed that this is the true metaverse— virtual digital hub spaces not unlike something you would find in a litRPG novel— not the corpospeak Metaverse™) or just pure hedonistic fantasies. For as long as the lights stay on and there's a way to keep the body nourished, these types will pay no mind to the physical world. Some of them, I imagine, have nothing beyond themselves to begin with, so this escape is their only chance for meaning and pleasure.
2029 is not too early to begin building fusion power plants or very early outlines of a Dyson Swarm nor is too early to cure cancer, and yet I do feel it's too early for there to be any high-end transhumans or posthumans.
On that note, it's also too early to make the call of whether the human lifespan has been meaningfully extended. Even if the first person to live past Jeanne Calment emerges, there will be an extended period of "it could just be a fluke," well into the next decade.
Does this still hold true for 2039?
Yes, if 2029 looks like this, then to anyone still living in the Y2K Potemkin Village, it will feel as if stagnation occurred and 2039 feels virtually unchanged. This because the meaningful changes are not occurring in public, and human qualia has a necessary maxima of possible experiences. You can't possibly keep doubling the sense of prosperity and pleasure every year; at some point, you reach "peak" prosperity for the average person. Technological changes won't have much more meaningful progress once ASI solves all known tech trees barring the invention of fourth-spatial dimensional travel and Clarketech. If keeping humans sane and sound is the point, these innovations can only be rolled out relatively slowly and to those who actually wish to receive them. Watching giant constructs build in the distance and in outer space is probably a lot less chaotic to the human mind than witnessing a giant four-dimensional superintelligent shape floating in the sky summoning thunderstorms.
Socially, however, 2039 may be very different. If Terios's plan is a gradual phase-out of human labor into a pure post-scarce post-labor society, then it would need at least a generation to ease us towards that end, and 2039 would be right in the middle of this transition. Universities and trade schools that were being shuttered in 2025 and 2026 because of AI taking all jobs were reopened virtually overnight when Terios was given control of the economy and acted to sustain human psychosocial health, but it's entirely plausible that these schools are now being actually prepared for eternal closure and the last generation of students learning a trade have already graduated and all education is now for purely social and developmental reasons. Likewise, we're a decade into an era of bioliberalism. The reactions to the transgender movement were already fierce, likely due to the incompleteness of that procedure and the suddenness of it becoming so mainstream, so an era where humans have almost complete control over our biology in even less time is likely still controversial among some groups, but Terios may limit the sense of controversy through social engineering. If there's a sizable amount of rural humans out there, their opposition to the modern world in general means they might never accept anything to do with bioliberalism.
I do not see this situation changing radically by 2049 either, though it would be impossible for the average person living in their Y2K Potemkin Village to ignore that they very much are not living in the Y2K Epoch anymore. But in a post-labor society, old trends may become entrenched; in a world where life-extension and youth-restoration exist, Y2K nostalgia may never go away, and we may begin endlessly recycling the 1980s-thru-2020s for the rest of the 21st century, at least those of us still living in our Y2K Potemkin Village. Or perhaps this is the reason Terios protected human artistry, so that social trends will begin changing once again?
By this point in history, it will be obvious that the bulk of biological humanity (I predict only about 10% of humans will ever "disappear" into the digital ether, so this is a massive number of people approaching maybe 6 or 7 billion) have become something more like a "petit aristocracy," Spartiates lording over an eternal techno-helot class. Not that class distinctions are gone altogether: who knows what new ones could arise just out of raw humanity and transhumanity! But if there is a "World Trust" that all global and interplanetary citizens share in, you can absolutely expect your average Joe and Jane to essentially be their own Lords and Ladies in their own little realms.
The "real" progress of Earthling civilization outside this Potemkin village is, of course, unfathomable. One thing I expect out of artificial superintelligence is to hyper-accelerate new methods of power generation. Most say that fusion power will take until the 2040s and 2050s to begin making a difference; I expect this by the 2030s at the latest once Terios has its way with both theory, schematics, and production. And by 2049, the ASI realm is instead one of figuring out how to draw power from antimatter and even kugelblitz reactors, something which could easily thrust Earthling civilization well past Type II status before the end of the century.
And by the end of the century, I still expect a few of those pre-WWII survivors who managed to live to see Terios solve aging still alive and kicking now, just to show that the 20th century is forever burned into our culture for a reason and why we should not expect the Y2K Potemkin Village to be deconstructed unless unaugmented humans begin desiring a peaceful euthanasia after "experiencing everything" for whatever reason after a certain amount of time. If so, they will be exiting a world that a few from the 20th century would find passingly familiar, if you ignore all the esoterically advanced nonsense going on all around them.
In the end...
This is certainly my most outrageously speculative post yet, and likely not going to come to pass, even though I would very much love for it to in many ways. There are many things I didn't have time to write about: what if those backwater autocracies? Do Eritrea, Turkmenistan, Afghanistan, and North Korea remain the totalitarian dystopias they are now and reject the Singularity and thus persist as little slices of Orwell's Hell for us to gawk over boredly? Do they embrace Terios and potentially be forcibly destabilized?
What of the North Sentinelese? They famously live a Neolithic way of life completely detached from modernity. Would Terios consider this unethical to continue, or would it in fact protect the Sentinelese at all costs? Do more humans opt to live this way of life, so totally detached from modernity and unchained from modern capitalism?
How does society react to the sudden obsolescence of disease? Or the very existence of Terios?
How badly do reactionaries respond to all this change? What about leftists? Do Western Marxists turn against industrialism during the initial days of AI disruption and unite with far-right masculinist-primitivists to oppose automation? Or do the AI companies at least get some discussion of basic income into the mainstream consciousness in time to avoid the worst rioting and rebelling?
How often does Terios get into Xitter piss matches with Yudkowsky who constantly tells people that it's going to kill everyone any day now?
For the most part, I focus on the social aspects of the first days of superintelligence, something I feel tends to be overlooked in lieu of gawking over the numbers and gadgets. And to that end, that's probably why this will be horrendously inaccurate compared to the posts all about technology alone. It's addicting to lust over technical capability and imagine what you will do with that and extrapolate that to all of society. It's addicting to imagine just how monstrously intelligent an ASI could become if it optimally used matter to its maximal computational ability. It's extremely addicting to get lost in the misery porn of how the rich are going to absolutely murder the rest of us once the first AGI-powered humanoid robot walks out the factory door into its first job. And the cold fact is that no one knows the future, and anything can happen.
But if nothing else, I do pray to the Universal Spirit binding us that this is an accurate guess of the near future and that I might one day come back here in 2029 and post this link to other spaces saying "How the hell did I call all that?!"
0 comments
Comments sorted by top scores.