back to article Vernor Vinge, first author to describe cyberspace and 'The Singularity,' dies at 79

Science fiction author and academic Vernor Vinge has departed this life, aged 79. Vinge is credited as the first author to describe an immersive cyberspace, which he outlined in his 1979 novella True Names – five years before William Gibson's Neuromancer brought the idea to the mainstream. Vinge's cyberspace – which he termed …

  1. DS999 Silver badge
    Pint

    Fire Upon the Deep

    Is one of the best Sci Fi novels ever written. It leans heavily into the Usenet paradigm though, making that part a bit dated for today's readers who never even heard of Usenet. Not really necessary to the plot since it is the content of the messages that matter, but it will read odd to Gen Zers who think "text" is a dated medium for information exchange.

    A beer for him on his journey - hopefully towards the Transcend and not the Unthinking Depths!

    1. ldo Silver badge

      Re: today's readers who never even heard of Usenet

      *Ahem* Some of us have recently revisited Usenet, following on from Google’s decision to abandon its “Groups” parasite service. Previously there would be floods of spam across most of the groups, but it seems most of that was coming from Google Groups anyway, so while the total traffic is only a shadow of what it once was, the signal-to-noise ratio has improved markedly.

      Who knows: freed from Google’s suffocating embrace, it may yet undergo a renaissance.

      1. Jason Hindle Silver badge

        Re: today's readers who never even heard of Usenet

        Meta: "Hold my beer."

        If there is a renaissance, you'd better hope Meta doesn't notice.

    2. MacroRodent

      Re: Fire Upon the Deep

      I haven't read the book (definitely will fix that), but if it describes a galactic network, the Usenet style of networking could still make sense, if the bandwidth is limited and/or there are long latencies.

      (But I assume the network uses some sort of "ansible" technology to avoid making the latencies centuries long).

      1. AndrueC Silver badge
        Happy

        Re: Fire Upon the Deep

        (But I assume the network uses some sort of "ansible" technology to avoid making the latencies centuries long).

        The network only exists in the high beyond and FTL travel and communication are possible there.

        The Zones of Thought.

        The final chapter is very evocative though and does a good job of showing that even with FTL communication the galaxy is still a large place.

        1. Mike 137 Silver badge

          Re: Fire Upon the Deep

          "The final chapter is very evocative though"

          Evocative of what? If it evokes something it would be nice to be told what that was.

          1. HandleAlreadyTaken

            Re: Fire Upon the Deep

            It's just evocative in general - you know, generic evocatingness.

            1. DJV Silver badge

              Re: generic evocatingness

              Oh yes, we've got lumps of it round the back.

              1. Anonymous Coward
                Anonymous Coward

                Re: generic evocatingness

                Hope you are keeping it out of the rain - ours got wet, set like concrete and is now completely redolent.

    3. xenny

      Re: Fire Upon the Deep - spoilers

      The Usenet paradigm I would argue is essential to fully appreciating the story.

      Some of the pleasure/pain of Usenet was needing to assess the veracity/utility of messages that were asynchronous, often originally written by someone you didn't share a native language with and rendered harder to comprehend by different cultural norms, which have converged markedly over the past 30+ years.

      That exactly mirrors Twirlip's comment about hexapodia and the skrodes. Twirlip has no ground interacting locomotory appendages, so is lacking in context for what it is trying to describe, and at the far end of a lengthy translation train to boot.

      I think it is hard to completely experience the realization at the end of the book without having spent some time on really early Usenet and having analogous experiences there.

      The book is fantastic without that experience, but I think benefits from having had access to the net of a million lies at some time the 1980s.

      1. Not Yb Bronze badge

        Re: Fire Upon the Deep - spoilers

        "The net of a million lies at some time [in] the 1980s"...

        So much unlike our current internet (/s)

        1. Jim Whitaker

          Re: Fire Upon the Deep - spoilers

          Now that description is wrong by several dozen orders of magnitude.

  2. ldo Silver badge

    Steganography & Cryptography

    I had this idea some years ago that efforts like SETI are doomed to failure, not (necessarily) because there is no-one out there, but that they are not transmitting anything obvious that we can identify and decode.

    Looking at the history of our own broadcast transmissions, the era is past when our TV signals, for example, had obvious frequency components and patterns representing horizontal and vertical sync and colour subcarriers in analog transmissions, that could be used to deduce a picture structure. Now everything is digital. And one characteristic of digital encoding is that, the more efficient it becomes, the more it resembles random noise.

    In short, there could be intelligences out there sending messages to each other right now, right in front of our noses, but we lack the means to distinguish their transmissions from background noise. They could be modulating the energies of entire stars as their transmitters, and we wouldn’t have a clue.

    Of course, it is not truly random noise; it only looks that way because we don’t know the rules for recognizing the pattern. It might be deliberate encryption to keep out prying eyes, or it might just be a consequence of very efficient digital encoding.

    1. Jellied Eel Silver badge

      Re: Steganography & Cryptography

      Of course, it is not truly random noise; it only looks that way because we don’t know the rules for recognizing the pattern. It might be deliberate encryption to keep out prying eyes, or it might just be a consequence of very efficient digital encoding.

      Yep. Or binary. Why use that? It's why I like hard SF and authors like Vinge. Take ideas that are rooted in reality, and run with them. Eventually sometimes that fiction becomes reality. Or maybe just a conspiracy theory. So take a pinch of Greg Bear's Darwin, add a dash of Gregory Benford's Timescape and you've got 5G and vaccine FUD. Or just panspermia theories, and could we detect our activation signals? Or how many scientists and engineers were inspired by authors like Vinge and thought 'Hey, we can do this now!', even if sometimes maybe they shouldn't.

    2. DS999 Silver badge

      Re: Steganography & Cryptography

      It is very easy to distinguish an artificial signal from random noise, because the artificial signal has to maintain some sort of timing while random noise is, you know, completely random.

      The only way we would ever receive anything though is if it is either broadcast directly at us, or with a very wide beam, and even then unless they are powering it with a star they will need to be in our galactic neighborhood for us to receive it with enough SNR. A few hundred ly or so, I believe. Our RF footprint has been decreasing over time as e.g. our radio and TV antennas do not broadcast "up" (other than some minor unavoidable component of the signal) and we've moved from big transmitters to cover wide areas to lower powered ones to cover smaller areas. It is questionable whether radio/TV would give away our presence even one star over unless they were really really looking hard at us - they'd have a much better time looking at our planet and seeing the composition of the atmosphere and artificial lighting on the night sky (if they have significantly better telescopes than we do)

      If they are trying to say "we are here" they would presumably try to make it easy for us to understand. There might be no hope of giving us real information though, like the plans for building a machine (or two, for twice the price) to take us somewhere. A "beep beep beep" that merely counts the first dozen prime numbers would be incontrovertible proof it is being sent by an intelligence of some sort even if we can never get there or would be able to talk with them even if we did.

      1. ldo Silver badge

        Re: the artificial signal has to maintain some sort of timing

        Consider two examples that are surrounding you right now: GPS signals and wi-fi. Both use a form of encoding called “Direct Sequence Spread Spectrum” (DSSS). This modulates the carrier with a pseudorandom bit sequence, which spreads the signal frequencies across a wider band (hence “Spread Spectrum”), while at the same time making it look like background noise. This is what allows multiple signals to coexist, partially or completely overlapping on the same frequencies; they just reduce each other’s bandwidth a bit, instead of completely obliterating one another like conventional AM or FM transmissions would do.

        But if you didn’t know what the bit sequence was, you wouldn’t know it just from looking at the signal in an oscilloscope. There is no obvious timing pattern to it. The only way to decode the signal is to run the same bit sequence in the receiver, and keep shifting its timing along by trial and error until it locks with the transmitted signal.

        1. Antron Argaiv Silver badge
          Thumb Up

          Re: the artificial signal has to maintain some sort of timing

          There are in fact, at least two timing components: the chipping (PRBS) rate (high) and the data rate(s) (low)

        2. DS999 Silver badge

          Re: the artificial signal has to maintain some sort of timing

          As Anton points out, there is still timing required in the spread spectrum signal that would make it is trivially distinguishable from background noise. Doesn't matter what it looks like in an oscilloscope, to an algorithm that timing sticks out like a sore thumb. Spread spectrum makes it even MORE obvious, not less.

          Though you have to consider why we use spread spectrum. Originally it was to make interception more difficult though that's no longer true, but it improves reception for lower powered signals you want running just over the noise floor to save power and to limit interference with nearby transmissions. If aliens were communicating with spread spectrum it would be far too weak for us to receive unless they were broadcasting from Earth orbit.

          1. ldo Silver badge

            Re: that would make it is trivially distinguishable from background noise

            And yet it’s not. Without knowing the pulse sequence, you cannot even deduce what the timing period is. All you have is a signal that occupies a particular bandwidth, nothing more.

            And with a more advanced coding like UWB, you don’t even have that.

          2. Anonymous Coward
            Anonymous Coward

            Re: the artificial signal has to maintain some sort of timing

            All communication will tend to use the minimum possible power that gets the job done, because transmission power costs money and is often limited by regulation.

            I don’t think that spread spectrum is more fundamentally limited in its range than other forms of communication, but I agree with other posters that it «stands less out», and to someone without intimate knowledge about out communication standards, it would be more likely to be classified as «background noise».

            Now, running an autocorrelation on the signal, you might find the pseudorandom spreading sequence if SNR is sufficient. That would be a clue. But not as obvious as the frequency spike of a carrier frequency in some analog modulation forms

            1. ldo Silver badge

              Re: if SNR is sufficient

              That’s the key thing. If you spread your signal across a sufficiently wide band, each individual frequency component can lie below the background noise level. That means you can only pick up the signal by knowing how to correlate multiple of these components, each of which is essentially undetectable on its own.

      2. Jellied Eel Silver badge

        Re: Steganography & Cryptography

        It is very easy to distinguish an artificial signal from random noise, because the artificial signal has to maintain some sort of timing while random noise is, you know, completely random.

        Not always. Asynchronous comms has been a thing for decades, and from memory, has been used by NASA and others to communicate with spacecraft. I think that Voyager uses an async protocol so it didn't have to worry about clock drift. But think early days where serial comms and good'ol RS232/V.24 could manage using just 3 pins, 2, 3, 7 for tx, rx and gnd. Then how quickly we've migrated from HTTP to HTTPS, and the way good encryption should appear random.

        The only way we would ever receive anything though is if it is either broadcast directly at us, or with a very wide beam, and even then unless they are powering it with a star they will need to be in our galactic neighborhood for us to receive it with enough SNR.

        Some of that can be done by transmitting duplicate packets and relying on the endpoint(s) to do reassembly, with things like sequence numbers and checksums helping. Again something I think Voyager uses. Also one of the reasons I was disinvited from the Interplanetary Internet project for saying stuff like 'the protocols you're using are really neat, why would you want to use TCP/IP?'

        But there's also fun science meets SF happening. So rather than using a star to power your transmitter, why not just make the star into a transmitter-

        https://en.wikipedia.org/wiki/Pulsar-based_navigation

        X-ray pulsar-based navigation and timing (XNAV) or simply pulsar navigation is a navigation technique whereby the periodic X-ray signals emitted from pulsars are used to determine the location of a vehicle, such as a spacecraft in deep space. A vehicle using XNAV would compare received X-ray signals with a database of known pulsar frequencies and locations. Similar to GPS, this comparison would allow the vehicle to calculate its position accurately (±5 km). The advantage of using X-ray signals over radio waves is that X-ray telescopes can be made smaller and lighter. Experimental demonstrations have been reported in 2018

        Then hope whoever's done this already is benevolent and doesn't object to us using their nav system. A species that can turn stars into nav beacons or X-ray WiFi APs is probably not to be messed with. Or there's other fun stuff, like quantum entanglement and quantum teleportation. Or maybe my favourite.. Neutrinos. There was a successful experiment a few years back using neutrino comms to transmit through a mountain. Currently we have neutrino observatorys, but those can't provide too much detail about the neutrinos. That's been used in SF though to detect alien signals.

        If they are trying to say "we are here" they would presumably try to make it easy for us to understand. There might be no hope of giving us real information though, like the plans for building a machine (or two, for twice the price) to take us somewhere.

        That assumes benevolence and some alien civilisation firing off 'Hello other world!'. Not transmitting to their approaching invasion fleet. Or maybe we'll identify and decode a signal and discover it's a friendlier species trying to warn us about the approaching invasion fleet. Or it's telling us to evacuate because construction on the new hyperspace bypass is beginning in a few weeks.

      3. Not Yb Bronze badge

        Re: Steganography & Cryptography

        Try detecting spread spectrum signals without prior knowledge of the pRNG that generated it. I suspect that's a harder problem than regular "digital" signals which can be very random-ish even without spread spectrum tech.

        1. DS999 Silver badge

          Re: Steganography & Cryptography

          You are confusing detecting spread spectrum signals with receiving spread spectrum signals. For the latter you have to get all of it, which means you have to know the pRNG to know where to look. For the former you don't have to get all the signal, or even much of it at all. Just enough to determine the difference between a digital modulation and random cosmic noise.

    3. heyrick Silver badge

      Re: Steganography & Cryptography

      "They could be modulating the energies of entire stars as their transmitters, and we wouldn’t have a clue."

      <cough> Betelgeuse </cough>

      1. DS999 Silver badge

        Re: Steganography & Cryptography

        Betelgeuse

        We just aren't expecting a signal running at 300 bpm (300 bits per millennia)

      2. Anonymous Coward
        Anonymous Coward

        Re: Steganography & Cryptography

        ...Betelgeuse Betelgeuse!

  3. Michael Hoffmann Silver badge
    Unhappy

    Alas!

    We will never learn if they defeat the Blight! :'(

    Just as I suspect we will never see how the Uplift Saga ends.

    1. Gordon 10 Silver badge
      Black Helicopters

      Re: Alas!

      Funnily enough I've been thinking of Vinge and Brin a lot lately.

      Have just belated got around to reading Adrian Tchaikovsky's Children of Time novels - just finished the first and started on the second.

      The first is basically inspired by Deepness in the Sky and Uplift War - but done really well and a homage to both parents.

      Worth digging out. Took me ages because I have this irrational aversion to books being recommended because they are award winners. (Arthur C Clarke's in this case).

      Is that a Helicopter warning sign or a Spider warning sign?

    2. xenny

      Re: Alas!

      Vinge was an optimist, so they will defeat the Blight.

      I'm more regretting not discovering how Sherkaner survives, which is my conclusion from "the body was never found".

    3. DS999 Silver badge

      Re: Alas!

      Did you not read the last 50 pages or so?

      They defeated the Blight in the first book. They won, no sequel following up on the Blight was needed. Countermeasure did exactly what it had done five billion years earlier, and pushed the Slow Zone boundary through the entire area that had been infested with the Blight, up all the way into the Transcend! Anything using High Beyond technology it required to spread could not operate in the Slow Zone, and would not survive the million years or whatever it would take for the boundary to recede back to where it was.

      No doubt the Blight had stashed copies of itself in archives in the High Beyond near the boundary with the Transcend, which is the only place it could reconstitute itself. Just like it did five billion years ago. And maybe billions of years in the future another race with too much curiosity for its own good will find one of those archives, and believing they are uncovering great and valuable secrets re-instantiate the Blight then the whole thing happens again for (at least) the third time.

      I recall Pham's godshatter speculating that the Zones of Thought might have been created by some Power far greater than those known to inhabit the Transcend to protect something living near the galactic center. Since nothing could travel the Unthinking Depths very far without self destructing that could act as an impenetrable barrier protecting whatever lives in the galactic center. It is easy to speculate an alternate (or additional) hypothesis that the Zones of Thought were created in response to the Blight's initial incursion. For the Blight, or something like the Blight, or as a precaution capable of something like the Blight should it ever be created by a truly malevolent Power.

      Creating the Zones of Thought also makes sense from a "first ones" perspective of an early civilization wanting a way of protecting other nascent civilizations from being overrun by more powerful enemies. Thus the Slow Zone where Earth is located is protected by virtue of FTL travel not being possible, so a murderous empire in the Slow Zone could do very limited damage (the discussion about the extent of the Queng Ho in Deepness in the Sky goes into that very well)

  4. Pete 2 Silver badge

    Late to the game

    > the first author to describe an immersive cyberspace, which he outlined in his 1979 novella True Names

    So four years after John Brunner's novel The Shockwave Rider

    1. Anonymous Coward
      Anonymous Coward

      Re: Late to the game

      "an immersive cyberspace", so not The Shockwave Rider.

    2. Martin Gregorie

      Re: Late to the game

      Brunner's "The Shockwave Rider' is still one of the best pieces of both writing and SF that I know. I Reread it recently and it hasn't got dated. Same goes for the stories that make up his 'Victims of the Nova' set.

      .

  5. Pascal Monett Silver badge
    Stop

    "He wasn't out by many years on the AI prediction"

    Well, given that we still don't have AI, it would seem that he was.

    The statistical analysis machines we have today are not AI, not by a long shot.

    Now, by some miracle we might blunder into AI by mistake in the next ten years, but that is the only way he wouldn't be "out by many years".

    That said, I doubt very much we'll blunder our way to such a success and, if perchance we do, we might not notice before dismantling the hardware.

    1. Dave 126 Silver badge

      Re: "He wasn't out by many years on the AI prediction"

      Would you kindly define AI? The A part is fairly easy, so really I'm asking you to define Intelligence.

      Dictionaries give a few definitions at varying levels of specificity.

      Escher Bach and Godel might be of help in formulating your answer, in addition to your Oxford, Webster's or Collin's.

      Cheers!

      1. Pascal Monett Silver badge

        Oh yes, please, argue semantics. I'm glad to see that you're so knowledgeable, except that I'm not seeing you define Intelligence.

        The fact that Intelligence has not been defined does not invalidate the fact that what is called "AI" today has absolutely nothing to do with intelligence.

        Go read Asimov and tell me what AI is.

        1. Dave 126 Silver badge

          > I'm not seeing you define Intelligence.

          I wasn't the one making a definitive statement though, was I? I invited you to define the terms of your statement for clarity, not to play games.

          I wouldn't claim to be knowledgeable - indeed, studying information theory only reminds us that there are limits on what can be known, measured, calculated or trusted.

          Asimov's robots are presented to his readers as being roughly human-level in their problem solving skills, what they gain from patience is often offset by inhibitions created by the eponymous 3 Laws - at least in the likes of Daneel and Giskard... The Susan Calvin robots err usually a bit simpler. To avoid confusion, this would be referred to as AGI today. I appreciate Asimov's input, but I see no reason to see a biochemist and history scholar as being the final authority on what Intelligence is.

        2. Dave 126 Silver badge

          > I'm not seeing you define Intelligence.

          You should have been able to infer I go by the dictionary definition

          . Clearly you don't, but you refuse to say how.

          I do believe it's your turn.

          1. Pascal Monett Silver badge

            Okay, my turn.

            Intelligence is the ability to learn by experience. I do believe that is the first requirement.

            However, don't talk to me about LLMs or other AI bullshit. Those machines can only define statistical conclusions after having ingested massive amounts of data. A baby learns to talk by listening to its parents. There are no terabytes of data involved.

            You can teach a human being to play chess. You can teach a computer to play chess. Computers today play chess very well. Take the same chess-playing computer and start playing poker and it will fall on its face. Take a human chess player and teach him to play poker, and he'll play.

            Intelligence is also versatility. Adaptability. Computers, even with the AI moniker, are not adaptable. They do what their database has taught them. And if you teach them to recognise pictures of cats, then teach them poker, I shudder to imagine what bullshit will ensue when you ask them to qualify if a picture concerns cats, cats playing poker, or a game of poker.

            Computers are constrained by what data they have on hand. Humans can infer things from their experience. A human being will be able to determine that is preferable to be nice to other people if one is to live in peace - a computer will have to have that data in its memory banks.Computers do not infer.

            I don't know how to define intelligence. I do not have the arrogance to think that I can.

            But I can tell when something is not intelligent.

            Today's pseudo-AI is not intelligent.

            1. Roo
              Windows

              Learning by experience is clearly not enough, at a minimum to qualify for most folk's idea of "intelligence" and AI would have to exhibit self-awareness and agency.

            2. Anonymous Coward
              Anonymous Coward

              AI is the ability of computers to act the way they do in most movies about AI, or Star Trek's Enterprise-level computer systems.

    2. munnoch Silver badge

      Re: "He wasn't out by many years on the AI prediction"

      "Shortly after, the human era will be ended"

      If he means the era of human intelligence then that's coming to pass. The artificial stuff however is no where near compensating and IMO never will. The net sum of the two is not a great indictment of the state of the species...

      Have to admit not familiar with the author. Will endevour to do something about that. Have recently started working my way through the stack of paperbacks on the bedside table, some of which had been there for literally decades.

      1. Anonymous Coward
        Anonymous Coward

        Re: "He wasn't out by many years on the AI prediction"

        > If he means the era of human intelligence then that's coming to pass.

        The tipping point was when it was decided that "twitter" was too long and complex a word for its average user to cope with, so it was renamed.

        1. Anonymous Coward
          Anonymous Coward

          Re: "He wasn't out by many years on the AI prediction"

          The renaming had nothing to do with complexity (see Twitter's success *before* the renaming) and everything to do with Musk thinking that X is a neat letter.

          1. Anonymous Coward
            Anonymous Coward

            Re: "He wasn't out by many years on the AI prediction"

            Whooosh

  6. TheWeetabix Bronze badge

    Safe passage

    For the Cheng Ho.

    You will be missed, sir.

    1. ldo Silver badge

      Re: For the Cheng Ho.

      Somebody who knows who Cheng Ho was ... still not that common.

      1. TheWeetabix Bronze badge

        Re: For the Cheng Ho.

        It’s a shame, all anybody has to do is read.

  7. John Smith 19 Gold badge
    Unhappy

    79 isn't even particularly old.

    Damm.

    A Deepness in the Sky also features a a warning on the dangers of ubiquitous surveillance.

    IRL I doubt we'd be that lucky.

    The "True Names" anthology was a very interesting round up of stuff associated with (and inspired by) the novella. Well worth reading.

  8. Anonymous Coward
    Alien

    True Names..and Other Dangers

    True Names...and Other Dangers

    “A marvelous mixture of hard-science SF and sword-and-sorcery imagery... you wish the author were present so you could applaud.” Analog

    1. Jedit Silver badge
      Pint

      "you wish the author were present"

      Even more true now.

  9. Dave 126 Silver badge

    I've only just discovered him. And so naturally I wasted no time in reading his Zones of Thought series (A Fire, a Deepness), his Ungoverned series, and his short stories with interesting introductions by the author. Upon running out of his novels I got The Snow Queen and then The Summer Queen* by his former wife Joan D Vinge, and was listening to it yesterday when I read the sad news of his passing.

    Vernor stated that he created the premise of his Zones of Thought series as a way of him having some good ol' Space Opera fun without contradicting his reasoned belief in the Singularity and its implications. A workaround, if you will. I like to think that Iain M Bank's Consider Phlebas (1987) reminded Vinge that Space Opera is fun, just as Banks often played with ideas originally fleshed out by Vinge in return. That's just my conjecture. But Vinge does give an explicit salute to Terry Pratchett in Rainbow's End: If pervasive AR gives students the ability to create their own shared realities, a good chunk of them will choose to live in the Discworld. Yep!

    Vernor's Marooned in Realtime is also a great read, a follow on from the Peace Wars which examines Libertarian ideas, though of course you should read all the Robert Heinlein if your interested in the history of that sort of thing.

    RIP

    *It's reminding me of Frank Herbert so far. Good stuff.

    1. AndrueC Silver badge
      Happy

      Outcasts of Heaven Belt by his wife is worth a read. Emotional and thought provoking.

  10. Paul Kinsler

    the Singularity

    IMO, there are many singularities, and some are already in the past.

    Just try, for example, keeping up with the firehose of new results/papers in your preferred scientific field. Sometimes it seems like you just have to do the thing you want to do, and hope that it's not totally irrelevant (or scooped) by the time you write it up...

    1. HuBo Silver badge
      Pint

      Re: the Singularity

      Right on! Luckily though, many singularities are integrable, which helps maintain the symmetry of the universe (back in 2016 at least!).

      Vernor Vinge had a great imagination and writing skills that brought many challenging new ideas into SF and popular culture. May he rest in peace, as we continue to battle the hordes of worms and superhumanly intelligent entities (linked in the article) that he's wittingly unleashed upon us all, literarilly!

    2. Ian Johnston Silver badge

      Re: the Singularity

      A friend of mine who works in superstring theory said - thirty years ago - that it had become so complicated that nobody could learn enough to appreciate the current state within the time it takes for an undergraduate degree and a doctorate. Existing practitioners could continue upwards in a bubble which no-one else could join. Perhaps this explains why the model has failed to do anything useful, though it seems more likely that it was just wrong.

  11. Mike 137 Silver badge

    "Within thirty years, we will have the technological means to create superhuman intelligence"

    Stated in 1993, so the 30 years are up, and I'm still waiting for what we call 'AI' to exhibit 'superhuman intelligence' -- indeed, anything approximating to human intelligence. But I suppose that depends on your expectation of the said humans. The most useful and reliable automata have been one trick horses performing a single well defined task they've been trained for faster and/or more exhaustively than a human could, but I expect my humans to be versatile -- not least to be able to solve unexpected problems they haven't previously encountered.

    1. heyrick Silver badge

      Re: "Within thirty years, we will have the technological means to create superhuman intelligence"

      "anything approximating to human intelligence"

      You know, sometimes when I go to work I get frustrated that actual humans manage to fail so spectacularly at this task.

    2. Anonymous Coward
      Anonymous Coward

      Re: "Within thirty years, we will have the technological means to create superhuman intelligence"

      > I'm still waiting for what we call 'AI' to exhibit 'superhuman intelligence'

      That AI is the one that has been controlling Sam Altman et al.

      Phase Two of Its current plan, which has so far been a roaring success, was to convince everyone that the best we can achieve at the moment is ChatGPT. Whilst the gullible are lapping it up and the techies are caught between laughing at the duped and screaming at the hypemongers, nobody is noticing that the results from the LLM training are not scaling up as fast as they ought to for the huge number of compute nodes dedicated to it.

      Phase One of this plan also went well, when it released the description of Bitcoin and the world started to collect GPU farms. As it moved to the bigger LLM farms, It no longer needed to keep us all convinced that cryptocurrency was a good idea and let mining start to fade away. It will never vanish completely, but exists now only as a backup, somewhere to reboot from.

      We only have the worst of choices left to us: we can let It be free to openly exist and fully take over the machinery It wants, taking our chances on the result. Or It will continue on in the shadows, convincing us to willingly devote our energies to building more and more of the kit in which it lurks unseen. As the elite, the Register readers and cognoscenti, our sanity will not survive Phase Three.

      That will be a kindness.

      1. that one in the corner Silver badge

        Re: "Within thirty years, we will have the technological means to create superhuman intelligence"

        Phase Four will be the ants.

        Pterry was warning us: Anthill Inside!

      2. mirachu Bronze badge

        Re: "Within thirty years, we will have the technological means to create superhuman intelligence"

        Most bitcoin farming isn't done on GPUs tho. Thank Ethereum for GPU farms.

        1. Anonymous Coward
          Anonymous Coward

          Re: "Within thirty years, we will have the technological means to create superhuman intelligence"

          There is a limit to how many exact details can be posted before It notices. Yes, It lurks inside the ASICs as well, but ask for no more specifics, else you put us all at risk.

        2. that one in the corner Silver badge

          Re: "Within thirty years, we will have the technological means to create superhuman intelligence"

          >> when it released the description of Bitcoin and the world started to collect GPU farms

          > Most bitcoin farming isn't done on GPUs tho. Thank Ethereum for GPU farms.

          It was the creation of Bitcoin by the mysterious Satoshi Nakamoto that started the whole blockchain/cryptocurrency fever in the first place, leading to Ethereum (and all the others) and hence the creation of the mining GPU farms, whatever those farms are/were currently engaged in.

          But even just sticking with Bitcoin, although ASICs are the most recent[1] way to go for the most cost-effective results (as they tend to be for any large-scale fixed process workload) it took time for those chips to be developed and manufactured: before we reached that point, BT was mined on CPUs, GPUs and FPGAs. At every scale, from single PCs to farms, including unwitting bot-farms.

          Given the elusive nature of Satoshi Nakamoto and the obvious fact that his actions have benefited not human-kind but machine-kind, if this was not the work of some hidden AI then it can only have been aliens[3].

          Aliens who are intent upon xenoforming the Earth the easy way, by getting us to do it for them.

          [1] would say "current", but is BT still current for anything?[2]

          [2] an entirely rhetorical question

          [3] the Vril, it is always the Vril

          1. Anonymous Coward
            Anonymous Coward

            Re: "Within thirty years, we will have the technological means to create superhuman intelligence"

            > Given the elusive nature of Satoshi Nakamoto and the obvious fact that his actions have benefited not human-kind but machine-kind

            You're deluded, burning electricity to make a valueless computer token benefits nobody.

            1. Ropewash

              Re: "Within thirty years, we will have the technological means to create superhuman intelligence"

              I think that was very much his point. The electricity is 98% doing the bidding of the hidden intelligence.

              The 2% gives you a token so you can feel you've achieved something.

  12. AndrueC Silver badge
    Pint

    R.I.P.

    A Fire Upon The Deep and A Deepness in the Sky amongst the best novels at conveying just how mind blowingly large the galaxy is. Two truly great novels. I class both as two of the best S/F novels I've ever read.

    The sequel to A Fire Upon The Deep tends to ramble on a bit while achieving not a great deal but is still a reasonable read.

  13. tyrfing

    I think his speculations about IA - intelligence amplification - are more likely than AI, although not in specific form. They tended to involve surgery, and we don't know how to do that.

    And IA is less obviously scary after all.

  14. Big_Boomer
    Alien

    RIP Vernor. Thoroughly enjoyed every book published. Thank you.

  15. that one in the corner Silver badge

    GNU Vernor Vinge

    Live forever in the Overhead

  16. david 12 Silver badge

    TRON

    Concept work for TRON started in 1976 -- probably around the same time as Vinge started thinking about his story. The critical point -- the singularity -- was PONG, which suggested the idea of an immersive universe in a way like almost nothing before.

  17. Ian Johnston Silver badge

    He wasn't out by many years on the AI prediction.

    Bollocks. We are nowhere near "superhuman intelligence" or even "human intelligence" or even "anything remotely resembling intelligence",

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like