Hyperbolic Growth | Hacker News Reader

Hyperbolic Growth

65 pointscachehit51 comments

There is a famous quote from Douglas Adams:

Anything that is in the world when you’re born is normal and ordinary and is just a natural part of the way the world works. Anything that’s invented between when you’re fifteen and thirty-five is new and exciting and revolutionary and you can probably get a career in it. Anything invented after you’re thirty-five is against the natural order of things. Apply this list to movies, rock music, word processors and mobile phones to work out how old you are.

These feelings are just the cohort of which we are a part getting older.

I guess that is comforting in a way :-)

There is a famous quote from Kenneth Boulding “Anyone who believes in indefinite growth in anything physical, on a physically finite planet, is either mad or an economist."

In my opinion, if we want progress, there is no other way but to believe we can achieve infinite growth, because even if we can't, we are forced to innovate along the way. I sure don't want to live in an agrarian society just because we don't have infinite resources.

All we need to do is invent replicators like they have on Star Trek and we're home free. Just jigger atoms around and we're gold. Plus the transporter where we can just instantly transport anywhere we want to go in the world instantly.

It's all really simple when we finally get that going.

Mark Miller said something like, "Moore's law from the CPU's POV: humans are getting exponentially more expensive."

The fundamental shape of the future is clear: either you have some sort of legal claim on or equity in the primary producers and/or "rent extractors" or you don't.

As a programmer, the next iterations of e.g. GitHub Copilot will eventually eat my career.

Large corporations are effectively already AIs, and they have much more ability and capacity to leverage technology advances than the little people.

I don't really know what to do about any of it, even assuming I could...

I've thought about trying to start a kind of member-owned "general automation" corp that builds houses and neighborhoods and ecologically savvy farms, but the simple fact is, I'm not a people person.

I've also thought about buying one of those cheap desert plots, a couple of acres for a few thousand dollars, and going out there with a "kernel" or "seed" of, y'know, solar panels and chickens and a little earth-mover, etc. I could pretend it was like I was colonizing an alien planet and make a kind of performance art piece out of it, I dunno.

>What will be a viable career in 20 years? Will the concept of a “career” even make sense by then?

I think AIs have made some amazing incremental process, but I feel like the fears of being made obsolete are still premature. If you needed to pick though, learning the thing that's replacing you is a good start, so pick AI research or development. For that to be non-viable as a job or career, AIs would need to have the ability to build and improve themselves, and at that point, it's game over, man™.

None of these AI things are going to replace politicians, governments, or economies, and so you still have a vast amount of soft skill careers available as well.

Really I would think that politician would be one of the easiest roles to automate. It's a very simple algorithm that votes yes based on whoever gives it the most money.

Getting them to hand over their jobs is the hard part :)

Those jobs aren't being replaced by AI because the people in charge don't want them to be replaced. It's not the AI they should be worried about, but society shifting the paradigm and rendering half of all bureaucracy across the globe obsolete.

Endless bikeshedding is deemed very important amongst its practitioners, and is pretty lucrative to boot. Even when we never measured the benefits.

Every AI I have tested requires the user to become clever and proficient at getting it to generate good results.

Programming with AI is going to just elevate the level of the programmer cognitively just like compilers and high level languages did. It will be an even higher level of abstraction at which developers can operate.

These things are powerful tools but just tools. Same with the “art generators.” So far I see nothing capable of autonomous reasoning at a level that would actually replace the tool wielding operator. Someone still has to hold the hammer.

While I would agree that the fears are in many cases premature, 20 years is a long time in tech while being short enough to matter for careers.

By 2042 we’re likely to have single-atom transistors in consumer devices, while people mid-career now are likely to be retiring or getting radical life extension to confuse all this even further, and newborns today will be just about to graduate from universities.

Will we have AI as general-purpose and flexible as a human, even if only limited to an IQ of 85? Dunno, but even limited to that IQ an AGI would render about 15% of the population permanently unemployable, which would be enough to cause all sorts of social and political problems by itself.

But I think the limit on AGI right now isn’t compute resources, it’s that we don’t know the right algorithms to make the AI we have as data-efficient as humans (or even, AFAICT, dogs, cats, or mice).

>Will we have AI as general-purpose and flexible as a human, even if only limited to an IQ of 85?

While I agree to a certain extent on this premise, there will still be some areas that AI won’t replace because humans also have an emotional value component.

Consider if you’d rather have a massage from a robut or person. Or, if you are infirm and would prefer a caretaker who is human or machine. It’s not just a matter of processing data when humans are emotional creatures.

If anything, AGI may free up humans to focus on human relational jobs which we’re much more evolved to be good at and value.

> Consider if you’d rather have a massage from a robut or person. Or, if you are infirm and would prefer a caretaker who is human or machine.

Sure, I’m also implicitly assuming robotics will improve. At some point — don’t know when — we can’t tell the difference any more, and at that point, the humans indistinguishable from machines are likely to be made redundant by the machines.

I’m saying there’s something innately different emotionally about interacting with a human.

So if you know it’s a machine, it will inherently change the subjective interaction. If you’re saying that it wont matter because we will get to a point where humans and machines are indistinguishable, that’s implying robots will be emotional and no longer just processing data in the way we currently think of AGI. Essentially it means they will be conscious with subjective experience of their own or just philosophical zombies. Both are essentially fictitious thought experiments at this point.

That's what it seems like they're saying. At some point it's likely they will be indistinguishable from us on many levels, including emotions. There's a demand for it and a solution/peoduct will eventually be created to address that demand, once the means of doing so are technically and financially feasible/scalable.

Currently, yes, there's an innate difference, but it's only fictitious that there won't be a difference until it's not. It's plausible speculation. Writing it off as fictitious as of now feels like an "ostrich in the sand" means of molding around a static perception of reality.

Things will change, we're not special, labor as we know it will be replaced, machines will at some point acquire the same abilities as us, and we'll have to strive to figure out as a whole if we need to change the means required to carry on and/or individually adapt/overcome such sea changes in labor/income.

It will probably cause many issues within our societies, but many will also see opportunities. It's going to be a lot of grey, likely more so than now.

I agree that there’s hubris in thinking we’re somehow special. But pretending we will get there “someday” isn’t much different than saying one day we’ll time travel. Even if plausible, it’s so far from where we’re currently at it belongs more to philosophical thought experiments than serious technology inquiry

I'd agree that they would either be conscious or p-zombies, but I think humans are sufficiently easy to fool that the latter may well happen. I offer panpsychism and celebrity fandoms/parasocial relationships as examples of people's emotional sense of connection being misleading.

I don't think anyone has a sufficiently concrete definition of consciousness to even tell either how far we are from that now, nor how to get there from here.

https://kitsunesoftware.wordpress.com/2022/06/18/lamda-turin...

I do think humans are easy to fool but so far only in extremely constrained scenarios. E.g., limited to chat bots. But that’s not what I’m talking about here. To fool people in the manner that’s being discussed, robuts would need to fool people with buddy language and on a subconscious level. So much more of how we communicate is beyond reading words on a screen. Consider the uncanny valley that is still prevalent with deep fakes. Getting better for sure, but that’s still constrained to premise video. To be able to do that in person in real time is a much larger problem.

The other examples appear to me as magnifications of mental illness instead of generalizable scenarios, possibly including the Google employee when you consider his background.

Hm. To me, reaching uncanny valley means they're probably quite close in form and behaviour. But it may well be the case that the remaining 1% (or whatever the percentage is) of the details may be harder than the first 99%.

I can see your point. But to me, the uncanny valley isn’t evidence that we’re super close, it’s evidence that we don’t know how to fully define the problem because there are elements we know are not wore right but can’t say why. Because we can’t fully define it means we aren’t close to solving it.

I certainly agree that we don't know what we're missing. My best guess timeline for this is very vague for this exact reason.

Do you see/feel this "something innately different emotionally" when you do your banking? Most human bank tellers have been replaced by ATMs and websites.

Neither of those examples are a good illustration of your point as many and probably most will prefer the machine in both scenarios. I don't get a massage for an emotional experience with the masseuse; I'd prefer the likely much cheaper and higher quality and consistent machine massage. Ditto with a permanent caretaker as it'll allow you to feel far more independant; most of us don't want to feel like a burden on other people. However, you do have a valid point, just not good examples imho.

AGI will make the vast majority unemployable, it'll be a disaster and those jobs won't simply be replaced by new ones: people who think this haven't thought through it very deeply at all.

Maybe. My original example that came to mind was a caretaker for someone terminally ill, I.e., hospice. Having volunteered in that area, people want human contact, even if it’s not performing any pragmatic function. I tend to think automation would just make somebody feel more isolated. People yearn for connection, not automation.

I believe the research shows there is a strong psychological component to massage that goes beyond the mechanical manipulation. I guess it would depend on whether somebody uses it for stress relieve or merely recovery.

Oh, terminal, yea that's one where I agree, they want company more than anything; but even that will eventually fall prey to AGI. People will be best friends and family with AGI's; grow up with them, presuming we don't go down the Terminator route; people are very quick to anthropomorphize anything remotely reminiscent of human behavior.

I think I'd rather have a $20 hug from a real person than a $5 hug from a heated robot. Even if the robot had better AI conversational skills than the person.

Hugs would have been a great example, human contact is the point; not true with massage, not unless you're being creepy. They're there for your muscles, not to turn you on and fill some emotional need.

Many people use massage for stress reduction which is an emotional need. Emotional does not necessarily mean romantic or sexual emotion.

This is my issue with discussions on tech. People tend to oversimply the problem as if a human is just another rational data input. Massage is more than just mechanically manipulation of the body. Autonomous driving is more than just perception.

Oversimplification leads to wildly optimistic ideas and timelines. I’m not saying they’re impossible, just that the problems are much more complicated than we give them credit.

Advanced auto-complete and sort-of right concept art is not going to disrupt the world nor the software industry. We have had algorithms trained on machine learning do IMO way more socially affecting things like decide what news and videos we watch, who gets approved for loans, what politicians say - but this gets way less press or thought.

You will never replace software engineers just as you will never replace writers and artists no matter how “good” the machine generated stuff gets. There is no soul there and it only appears interesting at a superficial, “that’s sort of neat” level that doesn’t hold up. No one is buying master works of AI, they spend their weekend reading real books and blogs from real people.

It’s 2022 - where is my self driving car? Where is the tens of millions of job losses from AI predicted that necessitated UBI (see the entire presidential thesis of Andrew Yang)?

It does get a lot of hype and raises a lot of VC money though!

https://bounded-regret.ghost.io/ai-forecasting-one-year-in/

May I ask if you have tried using Copilot? Because this sounds exactly like the kind of response I would have had half a year ago. It makes many mistakes and can by no means write code by itself. But that doesn't change the fact that it's orders of magnitude better at it than I thought possible.

I'm not the parent post and I have not tried autopilot, but I do have a question.

Does autopilot make you orders of magnitude more productive?

No, it does not.

One issue with copilot is that debugging broken code is far harder than writing it in the first place. We've seen similar transitions before though. When good optimizing compilers started appearing you had plenty of programmers like Mel [1] that said 'there's no art in it, a compiler couldn't do the kind of optimizations I do'. And yeah, it's true, but compilers got good enough that 99% of people stopped caring and hardcore optimization became a niche skill few people have. What does the industry look like when all that most programmers know is debugging generated code and what kinds of tools need to exist to support that?

[1] http://catb.org/jargon/html/story-of-mel.html

I reluctantly moved to an architecture position because developer salaries have stagnated and remote work normalization will put even more pressure on western developer wages.

Coding isn't nearly as much fun anymore. Agile, CI/CD, Pull Requests, code reviews, constant software updates, locked down laptops put so many barriers in the way of my creative flow.

I still code on my own equipment building apps for myself and I still love it but im done doing it professionally

> remote work normalization will put even more pressure on western developer wages

I'm not sure.. 1) The issues with putting developers in 'low-cost' countries are the same as before.. 2) it's true though that there's/will be increased competition with all the devolopers in your country..

Being made obsolete by technology is nothing new (see the Luddites). But regardless of how replaceable any individual is, we're still in a human society. Maximizing profits by laying off humans can work for a while, and systems can rebalance, but if prevailing dynamics collectively lead to a large percentage of the population being "unemployable", then the political situation will become very explosive. People need to feel some purpose in life. Once things destabilize to a certain point, the global productivity that allows for the enormous compute power behind modern tech and AI will come crashing down pretty fast. Sure, there are authoritarian scenarios to worry about, but I have a hard time convincing myself that tech enables authoritarian control beyond what was possible throughout history.

Why? Tech enables a small group of people to have extreme power/leverage. In the case of military/control you don’t even need the labor class to serve in the military in large numbers anymore for global supremacy.

For all the leverage the US has militarily, what control did it actually exert in Afghanistan? What about Russia in Ukraine? War and occupation is one thing, but I'm speaking more of the consent of the governed. Do you think the billionaire class can keep hoovering up all the economic gains as jobs are automated based solely on their self-interested narrative of capital ownership? What happens when unemployment hits 20%, 30%, 50%? Are they going to send drones against the populace?

Even in China, the CCP's policies only work as long as they deliver a certain measure of prosperity to the people. Propaganda and police brutality can only take you so far before the people have had enough.

Yeah, I’m fairly certain things will be fine. Everyone’s consuming curated propaganda on their phones every single day. The little people won’t be doing much.

It feels like part of this may be some needed mental shift from a scarcity mindset to a prosperity mindset.

My mind always goes to the very optimistic, Utopian world in Star Trek, where humans have identified that we can afford for people to stop working for the purpose of survival (the bottom of Maslow's hierarchy of needs) and instead start working just for the personal challenge and growth.

Not saying we're there yet, but with so many jobs rapidly becoming obsolete by machines, it seems to me to be the direction we're going. Rather than meaninglessness, joblessness, poverty, homelessness, I hope society can shift into this paradigm that automation could mean the exact opposite: empowerment to more freedom.

This is probably a very scary perspective for many capitalists! The majority of people actually do want to work, they just don't want to work for the sake of work itself. A huge benefit of capital is that it means we don't really have to trust each other at a high level, because fundamentally our incentives all align around survival which money is a tool to survive.

One day maybe we can make a world where there is more trust in each other; that just about everyone wants to do the best that we can with the exception of a small number of malignant personality disorders that is the minority.

This new trust will enable some amazing doors to open up.

You should read Trekonimics if your into this way of thinking.

interesting! thank you, I'll check it out

One would think people would have more freedom and will explore life far more than today and get into philosophical stuff. But the reality is messy. Imagine a single individual who leads quite a busy life and yearns to have some free time to be more "free" and do things that matter. You know what happens when that person gets whole lot of free time? They are not sure about what to do and it takes a lot to start doing something new. If this happens on a large scale, I'm really not sure that it'll make the world a better place to live. Maybe I'm pessimistic, but people would have ample lot of time to do more useless political stuff like manipulate, influence others for their own gain.

There's this anime called Log Horizon which kind of plays this scenario out. Everyone gets stuck in this metaverse sort of game where people generally come for fun. Once stuck, people start killing each other and doing random things because they don't know what to do. There's no purpose in their "life".

>Creativity is a big part of what I feel defines me, and if that loses all value then it will surely affect me.

Isn't that the whole problem with these tools? That they can't capture creativity and context, and therefore only work as augments to competent users rather than replacing users entirely?

>What will be a viable career in 20 years? Will the concept of a “career” even make sense by then? I have no idea, and I am uncomfortable with that.

Which is why society should answer this ASAP. It's evident that putting the means of production in the hands of a few becomes a problem for the many, once the many are no longer necessary. This has been going on since pre-industrial. The premise has always been "we will make more jobs than we remove". What if that premise is no longer true? What if that premise wasn't true to begin with?

Good time to mention the story "Manna – Two Views of Humanity’s Future"

https://marshallbrain.com/manna1

It essentially offers two different futures:

1. A dystopia view of the current trend of having more and more things automated

2. A utopian view

#1 is decidedly more Black Mirror and #2 is more Star Trek: Next Generation

The real world always seems to be a mix between the two. Technology does transform the world and make some things better, but it’s a double a edged sword and has unintended consequences that make some things worse, or creates new problems.

Why the word hyperbolic. It doesn't seem to make sense. Exponential automation seems better.

Hyperbolic growth has a singularity in finite time. I.e. there is a "crucial" point where things truly go haywire. Not a very important distinction in a philosophical discussion like this, but that was the reasoning.

This seems like a novel definition I have never heard before and is very different from the dictionary definition that does not have any reference to any singularity

https://www.merriam-webster.com/dictionary/hyperbolic

Reply on news.ycombinator.com