I just don't get it. How do you go from writing the kinds of future visions he has to staring at the singularity practically hitting you in the face and calling it "the world's money-losingest technology"?
Is it because he isn't actually using the technology for work on a day-to-day basis like a lot of us?
I just cant get over this term, do you honestly believe in this? I use AI daily and while it is super useful I see too many limitations for it to “recursively improve and cause an intelligence explosion”.
1. Clearly, the people selling AI with this idea benefit greatly with this promise of infinite upside. Can you can trust them?
2. Singularity essentially requires to handwave away a lot of baked-in issues with LLMs or rely on unrealized innovation.
The best part of the Singularity is that, once everyone who wants to uploads their consciousness to a computer does so, we mortal humans can simply unplug it and go on about our lives.
I believe that AI is a singularity. Just like a Black hole, no information can escape and you're slowly (by subjective experience) dragged in and spaghettified.
I think having a technical argument is important, as we live in a time with lots of hype merchants who stand to benefit from record breaking IPOs. Propaganda can affect us all, how do you know you aren’t being sold to?
What about the tools makes you think we've hit the singularity? My experience with them is that they've memorized a lot of stuff, but can't make anything fundamentally new. Most of the useful things LLMs do amounts to semantic search.
They are new proofs and for sure useful but as far as I’ve understood mainly interpolative. E.g. an LLM can “create” a poem about a purple chicken as it has datapoints for “purple” and “chicken”, so it can create something plausible inbetween.
Similarly, in my mind it can interpolate proofs by interpolating between data points for technique A and technique B. This is novel and brute-forcing proofs this way is useful. It is analogus to how sometimes it can generate programs that pass unit tests, I think.
However, creating fundamentally new concepts outside of the interpolated datapoints is not something I am convinced of. Maybe it can extrapolate some things, if correct add it as a data point, continue. Essentially a search, and it would be amazing if this works and maybe we can get some recursive improvement this way. But the “ideas” it will use to conduct this search are a function of the input data points as well, and thus in my view fundamentally limited in novelty. I am not discounting the usefulness, but I am not convinced you can just keep doing this indefinitely scaling intelligence exponentially.
Of course nobody can know yet really and I am just speculating just like you. But I also think the “experts” Sam and Dario also don’t know, and given their incentives I am not really convinced by them.
They're finding new examples in well-known categories of stuff. Also proving is a question of search. They're finding some stuff we missed, but not because they're smart, just because they have lots of data and compute. Think of them as something closer to brute force.
>Think of them as something closer to brute force.
Well brute force is something I had waited for all my life so it could make a personal computing experience higher performance and more fulfilling.
By the time 1980 rolled around.
In some ways it's like a fundamental assumption that's always been there and nobody questions or really "thinks" about. It hurts the brain, sooner or later you want a computer to be able to do some thinking for you ;)
It's one of the easiest tech concepts to understand, especially for those who don't have very deep abilities.
Hardware is hardware, stronger is stronger, and quantity has its own qualities.
People in certain positions can go farther with nothing more than this firm a grasp of computer science or intelligence and it shows.
The only thing to be gained by continuing to grow the scale of "datacenters" from this point is the brutishness.
I expect it to pay off too, for those who can afford it, just not for everybody.
What astonishes me about singularity skepticism: every argument was, from when it was released, "but it can't do X", "it's impossible/too expensive". The benefits AI -> AGI are staggering, the competition is intense (both intra-industry and internationally), and the gains self-reinforcing.
I have fears of dystopian worst-case scenarios, and dread the rapid pace of change and what it will do to real lives lived in the world, but it's only clanker cope/wishcasting that believes that (AI+human ingenuity) won't produce real AI ingenuity.
Re-read your comment in four years, I'll take the bet you'll find it very naive.
So, what your saying is that there is a perhaps linear, perhaps exponential increase, and that you are projecting that increase forward indefinitely. Let me know if this is unfair.
Counter argument: does anything else work this way? E.g. Moores law had an end too right? I would argue that the core tech breakthrough (Transformer-based LLM) has been improved, but no fundamental further innovation seems to have been made. The current architecture fundamentally hallucinates, even Fabel even on trivial problems. I.e. as number tokens increase error likelihood goes to infinity. How then, can this scale recursively to infinity?
The singularity usually refers to the idea that although the rate of relative improvement of a technology is approximately constant at a constant (human) intelligence level ([dy/dt]/y ≈ c), producing exponential growth, if said constant were actually proportional to intelligence and the technology under improvement being intelligence itself ([dy/dt]/y ≈ cy), then solving the differential equation leads to a blow-up to infinity in finite time, a mathematical singularity.
If instead additional intelligence does little to speed up AI development (due to the need for other inputs like caputal and time), you could get a world where AI becomes better than humans at AI development and begins a cycle of recursive self-improvement without explosive growth leading to a singularity.
That's a pretty good concept and you could say there is a fairly wide gap between your everyday singularity and the ultimate singularity.
Which seems to me likely the result of an unforeseen variable or variables, and that's got to have outsized, uncharacterized, and unexpected importance to have such a strong effect.
When the overwhelming consensus is that wonderful things are waiting just around the corner, it still could turn out to be just the opposite and you'll never know until you actually turn the corner.
It used to be that society always took some time (read: decades or generations) to absorb new technology. To percolate into various niches, people understanding how it works & how to use, its limits/drawbacks, getting used to its effects, etc. Even mass-production machinery takes time to build.
Modern AI has taken those brakes off. Humans are taken out of the loop. IT infrastructure that modern society leans so heavily on, could be overhauled overnight into a very different beast. We wouldn't know what's happening before it hit us.
If anything, society should be reluctant to give AI powered systems 'hands & feet'. Use systems offline where possible. But we don't. Maybe some "Morris worm of the AI age" will get us to pay attention.
Whatever is coming, if it's bad we probably deserve it.
There’s a short story out there about a device that can actually do this and then let you stay in contact with your alternate self for a while to see how things turn out. Fascinating stuff.
There's another short story out there which assumes that when you die in one branched universe, you survive in another. So in your subjective experience, you always survive, no matter how low the chances. But eventually your survival requires stranger and stranger events to occur. After a thousand years, your subjective experience becomes utterly implausible, and yet there you are. But you may not like the universe in which you survive that long. Which is too bad, because you can't subjectively die.
I'll probably remember the title tomorrow, if nobody beats me to it. It was a fairly disturbing story.
Ah, thank you! That is it. Just as biting and dark as I remember it.
Greg Egan also has a number of stories about people who rebelled against a "many worlds" universe by porting human cognition to run on "Quantum Single Processors," which would carefully isolate themselves from the universe until they had made a single decision across every possible timeline. This didn't prevent them from being "branched" by outside factors, but it at least gave them an approximation of free will.
Greg Egan writes true sci-fi. Exploring what-ifs further than anyone has explored them, and putting them up in story format so you don't have to be Einstein.
I think MWI is appealing to a lot of people as part of the quiet, but ever-present undercurrent of "how can we make any sort of God totally unnecessary to the universe?" that science has, but if you think about it deeply enough, it becomes clear that MWI, if true, is unbelievably horrible. We'd all better hope it's not the correct interpretation.
Edit: Thanks for the link to the story. I had not read it or seen it before. Same principles for sure. I don't think it's a crazy extrapolation of MWI, I think it's the only logical outcome. I can put that opinion into more firm mathematical language but it's more than I can put into an HN comment, and I haven't typed it out anywhere else either, and it really is just that opinion, in more mathematical language.
Having a slew of physical theories with profound existential ramifications to choose from, why do you think the "many-worlds interpretation" is particularly connected to secularization?
As far as I’m concerned, believing in MWI is the realization that causality is not a function, but a relation (in the mathematical sense), and that actually feels more natural and inuitive, all things considered.
I think there was also a concept for scientifically testing the MWI, but that only works subjectively. You essentially play Russian roulette, preferably based on the outcomes of a quantum measurement. Repeat until you either die or are satisfied that the world you are in is so implausible as to be impossible unless indeed every outcome is realised.
Say, fire an electron at a double slit with a detector in one of the slits, and kill yourself if the particle is not detected. Run the experiment 10,000 times, or 100,000,000 times - one copy of you will eventually be satisfied that in any probabilistic interpretation of QM this is not plausible, and all the other copies will be dead.
This works because the MWI predicts that any outcome that has non-0 probability according to the Born rule will be guaranteed to happen (it just "happens less" by some hard to define metric).
Personally I believe the entire notion is absurd, and that this type of thought experiment makes it clear, but still some like to be contrary.
I arrived at this philosophy independently, so I tend to subscribe to it. Along the lines of The Secret and manifestation, I've noticed that whatever we think about tends to happen in reality (as above, so below).
So the main difference between someone like a Buddhist monk and a former US president with a taste for gold is one of choice. The monk acknowledges that all routes to living one's best life are possible so abstains from attaching to outcomes too strongly, while the former president asserts his ego to maximize a certain dimension like personal wealth at the expense of all the others. Too much choice and we risk being ungrounded, too little choice and we end up caught in a web of our own design.
There's a great scene on the show Vikings where Ragnar says:
Power is only given to those who are prepared to lower themselves to pick it up.
Really everything is possible, and we can use our will to sidestep into other realities. But from a framework of reincarnation and the multiverse, our choices can impose on the freedoms of others, so we should be mindful of the impacts of our decisions, because others are aspects of ourselves in another life.
I feel rather strongly that most of the world's problems like wealth inequality and war stem from overexertion of the ego. People constrain themselves into corners and then project their anxieties onto others to the point where it seems like nobody gets to live their best life.
I feel like this is the popular sentiment these days but I don't fully agree.
I've seen so many companies spend themselves to death and its easy to have it happen much faster than anybody in charge realizes.
I've also seen companies grow expenses as things go well and suddenly there's a massive downturn in the market and they can't recover in time. Meanwhile, more frugal companies have the runway to weather unexpected events.
So I still think there's more value in frugality than is generally accepted.
"Based on the Starlink website, that all satellites will be deployed below 600km. In this case, satellites may be visible during twilight but not after nightfall, greatly reducing the potential impact to astronomy."
It's the physics of the arrangement of satellite, sun, and Earth. The satellites are only visible due to reflected sunlight, and they're flying quite low for satellites, i.e. near to Earth. The nearer an orbital object is to Earth, the larger percentage of the arc of its orbit is occluded by the Earth's shadow.
Given their close proximity to Earth, they can reflect sunlight at dawn and dusk (when the sun is shining "along" the surface relative to the observer) but when an observer is in nighttime (i.e. within Earth's shadow), Earth blocks the path of light to reflect off the satellites and they won't shine. Hypothetically, they could still pass in front of distant objects and occlude them, but I'm assuming based on the way the problem is described that flooding a telescope with reflected light is more of a problem than occluding a distant object for a few moments.
There's nothing wrong with crying or in general showing or having emotions, "negative" or otherwise.
Some people just have that reaction to difficulty and frustration or whatever else, it doesn't have to mean that an environment is terrible or immoral if people cry about it sometimes.
Sometimes it’s going to be unavoidable. I used to work in help desk and at least once a week an employee would either cry, or take their phone out of the queue and storm out of the office, something to that effect. I remember two occasions where an employee punched a hole in the wall. I took a day off work because one of the callers threatened to come to the office with his gun if I couldn’t solve his problem.
Sometimes there are jobs that have to be done but that make it really hard to control your emotions. And some people have a really hard time controlling their emotions anyway. It’s better to have a safe place to let that out.
I'm sorry you had to go through that, but unless you work at a suicide hotline or one of the departments that tracks down child-porn creators or something along those lines, people breaking down at least once a week should not be "unavoidable." I know help desk is always going to be stressful, but employees need to be empowered to speak sternly to, and hang up on, abusive customers. If they're not, that's a management failure, not just an unavoidable reality.
Worked with a developer who had to leave the office temporarily one day to go outside and cry in his car.
Was in a team where he could have spoken to any of us and we all would have listened and helped, but I guess sometimes, some people, just need to be alone and cry.
I tell you it was fairly destabilizing working in a small office in a highrise in a city center. 90 minutes at best from home on two trains, open office plan in a small office space, it was like being in a cage, being so far from a space which was "mine". The only escapes were spending a long time commuting, walking around outside on busy city streets, or sitting in a bathroom stall.
I think the psychology has something to do with growing up on a few hundred acres.
Boy could I have used a crying room. Not to actually cry but just to escape for a while now and then.
I moved from the country to a city. Its been four years I still feel like this!
Out in the country I was five minutes walk from lochs and l forests. I used to be able to leave the house and walk for twenty plus miles and maybe see a handful of people. Just me, the outside and hours of alone time.
You dont realise how much you miss that! Now in a city with no car and its so hard to get away from everything. Jusy to get away and clear my head.
Yeah, I live alone but sitting inside my house unable to get out and away from life can feel so trapping.
Trusted Insight | Full-Stack Web Developer | San Francisco or REMOTE (USA) | https://www.thetrustedinsight.com/ | Full Time or Part Time
Trusted Insight is a well-funded startup building Fund Studio - a "CRM" for the Venture Capital and Investment industries helpings LPs and GPs manage and evaluate their investments.
We are looking for Full Stack developers to join our growing team. We're built on Python 3 and Django 2 on the backend and Angular 6 on the frontend. Machine Learning experience is also a plus.
Soooooo... if people use their standard password on this you'll be able to login as them as soon as you get any personally identifiable info - like email or FB account, right?
Is it because he isn't actually using the technology for work on a day-to-day basis like a lot of us?