If so it kind of fails at it. Flight is well defined, intelligence isn’t. Nobody doubts that an airplane flies, but lots of people—including myself—would never call alpha-go an intelligent machine.
A decade or even a year before AlphaGo, most people would have said an AI capable of playing Go at superhuman levels was intelligent.
Intelligence is a mystery. As soon as we can build a rational system that demonstrates a particular kind of intelligence, the mystery and perception of intelligence is lost.
That's because we use the question "Who can solve this task?" to estimate difficulty, and we assume a priori that computer solvable tasks don't require much intelligence.
Per Moravec, it's tended to be the opposite. The tendency to assume that excellence in chess - a game machines play by lookup table - is a sign of extreme intelligence because people aren't particularly good calculators and most of them haven't memorised many openings, but walking around interacting with an environment and forming goals as a result is something any dumb animal can do.
> A decade or even a year before AlphaGo, most people would have said an AI capable of playing Go at superhuman levels was intelligent.
Citation needed. We had machines playing other games at better-than-grandmaster level by that point. People talked about AI's struggles to play another perfect information game with very simple rules because it couldn't brute force parse every single permutation as an indication of how limited it was, not an indication that true intelligence was pruning search trees.
I disagree on a philosophical level. Intelligence has always been—and always will be—a moving target. As soon as we discover something that we previously thought was unique to humans, we redefine what intelligence is to exclude that trait.
Previously using tools was a sign of intelligence until we discovered that other animals (even insects) do that. Then it was required that you had to make tools, until Jane Goodall showed us that other apes do that. Until AlphaGo we had this idea of particular games, always upping the size of the decision tree. And now it looks like we are in a crisis and just have this vaguely defined general intelligence, a term with quite a racist history, that nobody actually knows what means.
Me personally am of the opinion that we need to ditch the word intelligence in science and technology (it can remain in philosophy). It has all the same problems of a grand unifying theory in cosmology, and a whole host of more (see Stephen Jay Gould to discover those).
Calling it a mystery is being optimistic, rather, I would call it a misdirection.
Not as funny as reality though, where AI is suddenly good at art and writing, but still bad at flipping burgers, or simply navigating floor of a house.
> Moravec's paradox is the observation by artificial intelligence and robotics researchers that, contrary to traditional assumptions, reasoning requires very little computation, but sensorimotor and perception skills require enormous computational resources.
Moravec's paradox is just an observation made on 80's state of AI. It doesn't hold true - high-level reasoning is absolutely not cheaper than sensorimotor stuff, and most importantly there might be no difference between them.
> Moravec wrote in 1988, "it is comparatively easy to make computers exhibit adult level performance on intelligence tests or playing checkers, and difficult or impossible to give them the skills of a one-year-old when it comes to perception and mobility".
That is, few compute resources are needed for a program that can beat human checkers players, as compared to what’s required to solve perception problems.
I wonder if you have a different thought of what “reasoning” was being discussed, as you switched the term to “high level reasoning”?
I forget who recently observed this: AI is doing well at art and fiction writing because those don't have to be "correct" but it's less useful for writing code or non-fiction.
I feel like OP is almost claiming the opposite. Art isn't clearly defined - there's no "correct art". Therefore whether you prefer art generated by an AI to art generated by a human artist is, within some margin in a blind test, random.
If I have a human and an AI go illustrate a manual for assembling my new sofa, or produce the advertising materials for the grand opening of a local restaurant, it's much easier for the AI to clearly produce an incorrect illustration in a way a human generally wouldn't.
Most of the "art" created by AI is on the level of assembling a sofa. And sure, lots of people have poor taste and criteria for understanding art.
If nothing is being communicated, there is no art, in my opinion. Until AI is conscious, that will continue to be the case.
Sure, but if you see a piece of art with no context for its creation, you have no way to know what the artist intended to communicate or if they intended to communicate anything at all, or if there was even a human artist involved. I'm aware that Mondrian was a real human who made art, but if you told me Mondrian was an AI, would that make my enjoyment of their work less valid?
In short, I mostly agree that good art is communicative, but also that some of that is because of the context where someone shows you something and tells you it's communicative and maybe what it's intended to communicate. There's plenty of famous works of art that don't look that different than programmatically generated NFT works, let alone things Stable Diffusion does.
People are really good at generating meaning and patterns from actually literally nothing. See conspiracy theories and the gambler's fallacy. If you know it came from an AI that intends nothing, it's easy to see the seams that tell you it's all pieces of meaningless nothing. If you're told that it's from some artist with a long history and passion for expressing some message, it's easy to read communication into anything.
There is no discussion possible with that line of reasoning. Invoking God is terminus for any discussion independent of your beliefs.
In your frame of reference consider as though God is building it through the human.
It still make no difference, whether we could still do it through conscious design than just evolutionary pressures which was what this discussion is really about.
That's the funny things about the article. Though it mocks the idea, in the end we can't create a machine that has the same functionality as birds. Of course, humans only use a sliver of that functionality for their own ends, but even then we still raise billions of birds for our own ends. Most people still find cats to be the best way to hunt vermin, insect the best way to pollinate plants, seeing eye dogs better guides than any robot, snails better at auto-cleaning fish tanks than any device, dogs better at tracking scents in public than any machine, etc.
Dijkstra said the same thing. Why bother trying to make a computer think like a human? Why not make it do something better? I tend to agree.
Or to borrow from Chomsky, why would we try to make a machine deadlift a box like a powerlifter when we can just make a fork-lift truck / front-end loader?
Well, function implies form in this universe to some degree. The more functions we want to put into a machine the more it's form is constrained. If we want something that does all what birds do .. we need to build a bird.
That said we already have drones. We can add a charging station, an artificial beak, a speaker and it can do bird stuff pretty well :)
What we suck at is building self-replicators.
Cats are considered better vermin hunters than simple machines, like rat traps, because they have a lot of extra benefits of being household pets. (Which rat traps don't do well. Also usually dogs are used to hunt vermin. [I wanted to find that video where dogs hunt on the field as a tractor plows it, but only found this about minks :o https://www.youtube.com/watch?v=MB1I7UixASY ].)
Fish tank cleaning is also a very specific problem. It's basically a niche. If we look at it as an engineering challenge it's a hard problem. Small space, hard to reach spaces, mostly mechanical work is needed (as fish tank life is very sensitive to water quality), etc. And there's also cost. It's very possible that if we were to spend a few billion dollars on the problem we could come up with a fish tank cleaner, but ... snails are provided by evolution for free.
I'm guessing the stinger evolved in response to most insects, not humans. Here's a video explaining how they don't necessarily always lose their stinger, even when they sting humans.
I think the difference is though, that nature doesn’t intend to make something. A friend of mine once described evolution as not running toward something, but rather running away from something. And there are millions of years of wrong-turns, and vestigial remnants of things that didn’t work out.
Nature takes a long time to build and specialize things though. We humans for example took some time to get going. But now in the space of a few generations, we've gone from soldering together some transistors and other electronics to building software systems that can do a lot of things. Some of the people that did the soldering and early tinkering with software are still alive even.
So, we're a bit impatient but overall things are moving really quickly. I'm only 48 years old, I expect to see some amazing things in the next few decades.
This article’s main point is that AI technology has economic potential. It goes about this by parodying common arguments against whether such AI can develop into goal-driven beings.
One example of such criticism is this five year old piece: https://www.wired.com/2017/04/the-myth-of-a-superhuman-ai/ Few are skeptical about AI technologies’ economic impact. In that sense, the article misses the ball slightly, but it’s so funny that I don’t mind.
To state my point in the vein of the article: The bulldozer did change our economy. It does most, if not all, of its productive work in cooperation with a human. Turning a bulldozer on and letting it run by itself is usually a waste, and may be dangerous.
> This article’s main point is that AI technology has economic potential.
Thanks for a great, entertaining ditty. (I assume you're the
author). I paused at the end, thinking it through, without immediately
jumping to that conclusion. I'd say I'm a quite careful and
sophisticated reader, so I love the way you left interpretation open.
Fully agreed. I first misunderstood the article as mocking AI alarmism, but it really does the opposite: It's parodying articles that underestimate the potential of the tech, because applications are not a 1:1 drop-in replacement to what we have today.
Hence references to such future-proof technologies as delivering messages or cargo(!) via carrier pigeon.
Doesn't bode well for artists if they are essentially compared with carrier pigeons here...
Of course, if people had questioned the potential of flight because most of the people egging its potential was arguing that within a few years we'd no longer need cities because we'd all live wherever we wanted and whistle for the nearest mechanical bird to give us a lift, they wouldn't have been wrong to question that scenario. Even more so if discussion of powered flight was dominated by speculation about mechanical birds throwing us off cliffs if we didn't solve their alignment problems. Back in the world of AI, a spreadsheet can't even play chess, but it's a remarkably useful thing to have, has given people that used to be employed as clerks more tasks rather than less, and the person that notes how this changed everything but not at all like the futurists said it would isn't necessarily pessimistic about spreadsheets.
Powered flight is enormously useful in a lot of domains, and yet for all its obvious superiority in many dimensions it hasn't even made infantry or shipping obsolete. And if I felt like hunting, I'd probably prefer a hawk...
I'd say that people question the potential of any new technology. "There is no reason for any individual to have a computer in their home" quote attributed to Ken Olsen of DEC is one of my favourite examples. The general aversion to change leads to deprciation/ignoring the adavantages and over-exposing of risks - just like you perfectly captured it in this passage "mechanical birds throwing us off cliffs".
> I'd say that people question the potential of any new technology.
Sure, but exaggerating the potential is much more common. Look at where people in the 50's thought we would be in terms of space or personal flying devices. This stems from both overestimating the technological capabilities, overestimating the usefulness if we were able to succeed, and failing to see the non-technological limitations.
Flight is actually a good example of the last one. For instance, the problem with Edisons prediction that people would fly to another city for a couple of hours after breakfast and be back by lunch isn't that technology isn't advanced enough. But the costs are prohibitive, the airports are often far from the city center, and even when you get to the airport it takes a couple of hours to actually get into the air.
Planes were hot, so Edison thought travelers "will largely scorn such earth crawling." But earth crawling, even long distance, is still very useful, hence the desire to invest in things like high speed rail.
It is very hard to predict the path of an S-curve. If you don't estimate when it takes off and when it slows down just right you will be way off on your view of what will happen.
Crafting arguments of the form "predicting the future is hard, as shown by this past attempt failing" can be challenging. Understanding the past, guessing at the contingency of development paths, etc.
I recall such a talk by someone respected which, IIRC, included dissing the Ladies's Home Journal 1901 "What May Happen in the Next Hundred Years"'s[1] prediction "The horse will have become practically extinct." with a dismissive "The aren't!". Seeming underappreciating the time's massive horse footprint and diversity, now vanished and diminished.
> But the costs are prohibitive, [...] even when you get to the airport it takes a couple of hours to actually get into the air.
I wonder how predictions might deal with the diversity of timelines? And thus include the fork two decades back, where 9/11 didn't happen, the airline industry remained competitive, and I can still do a Boston-to-NYC short sprint from curb, across a narrow terminal, down the ramp, and squeeze through a closing airliner door, with no staff interaction beyond their shouting "run faster!". And everyone pays cash to stewardesses. And no ids (after all, showing papers to travel is an un-American Soviet thing). There's a lot of timeline diversity.
> [it] isn't that technology isn't advanced enough [...] the airports are often far from the city center
I wonder if there's a timeline where Robert Moses loved rail, smashed a straight network of lines across Long Island, and now NYC-anywhere to JFK is minutes. In contrast, there's the timeline where the massive and competitive subway construction of Edison's life, stalled out shortly after his death, and remained dead for a century. Perhaps the medieval warrens of European cities, persisting but for revolutions until aerial mass bombardment, might have been a hint. But new york was still doing massive "tear down and build new". Perhaps there's a timeline where a cold war nuclear strike on nyc cleared the way for greater development, as with the city fires of old. It does emphasize the contingency of history and thus progress, but it doesn't seem unreasonable that Edison didn't expect the trip from still-existing 1910-ish apartments in Manhattan, out to a "opened a generation after his death" Idlewild airport, to remain largely unchanged for a century. If you travel around Xmass, the train you catch might have a just-after-his-death 1932 vintage car. Current US journalism about China repeatedly reminds me of European writing from warrens about Edison's America.
Perhaps we need some expanded conceptual vocabulary for sketching possible futures?
FWIW, the eagle footage was edited to the point of disbelief, though it is sort of half real. The eagle was trained, and though wild Golden Eagles are known to attack prey up to 100lbs., they do not throw goats off cliffs nor can they carry them back to their nest. It's a shame that David Attenborough used this footage because it is bullshit.
I think you're confusing a couple of different things. The eagle goat footage from Frozen Planet was shot by Attenborough's own team. There was nothing faked about it.
There is some older footage of a similar thing but it's from a film and is all fake. The Frozen Planet footage is the first time it's been recorded for real.
There is also some Frozen Planet footage showing an eagle's view of flying over some mountains. That footage did use a trained eagle mainly because it's difficult to strap a camera to the back of a wild eagle and still have a face left.
> The eagle goat footage from Frozen Planet was shot by Attenborough's own team. There was nothing faked about it.
Attenborough does not have his own production team. Most of the footage in Frozen Planet was produced by the BBC, but not all of it. Over 30 shoots were produced locally.
The Frozen Planet golden eagle footage is mixed and contrived, and it is easy to tell by the editing. The BBC shot all in ultra high definition, and some of the footage of golden eagles is in 8K and shot with a really decent professional camera. The drone footage is of a trained eagle, otherwise it would have taken down the drone. And the footage of the eagle taking down the kid and flying with the kid (for some reason we don't see it lifting the catch, probably because it is a stuffed prop) and dropping the kid is in 1080p shot with a shitty semi-professional camera with shitty lenses, upsampled to 8K and downsampled in post, as is evident by the grain. The eagle in the nest feeding the baby is not the same eagle, and it is not feeding the baby the kid.
A falconer named Jordi Isern[1] that shoots in the Andes shot the footage of his own pet golden eagle that he trained to hunt chamois (the goats in the footage).[2] That is Isern's footage that the BBC remotely directed and later edited to mix with their own footage. Wild golden eagles do not hunt chamois, and for good reason.[3] Chamois are incredibly resilient and tough creatures, while the bones of golden eagles are hollow. The diet of golden eagles is rabbits, hares, ground squirrels, prairie dogs, marmots, fish and carrion, all of which is plentiful, so there is no reason for the grave risk of taking such large and dangerous prey.
Trained eagles will hunt anything they're trained to hunt.[4] Isern is not the first to train a golden eagle to do this. There are other falconers in California (the first to do this as a proof of concept, shot on actual film before the advent of HD video, that I can not locate, lost in Google's shitty algorithm and the noise of the Internet perhaps forever, but it exists) and Spain[5] that have done the same thing. It's such a fantastic idea that the BBC and Attenborough couldn't resist, with Attenborough rationalizing that maybe kids were prey, but not adult chamois, thus Isern was directed to show the adolescent struck rather than an adult chamois being pulled off a mountain by its horns. Wild golden eagles don't fuck around wrestling with their prey. They always kill on impact.
Those falconers suck, and it sucks the BBC and Attenborough are involved. But it's definitely bullshit.
I don't quite understand the "art" comment. I've seen it several times so this is not aimed at you.
The point of art art is that it is created by a human or group of humans. Artists (performance or visual) will be out of a job when runners are all replaced by race-cars, and pro golfers, tennis players, football players, and basketball players by robots.
There is a difference between artists who create for their own sake, and commercial or professional artists who make a living off art. The latter is what many people worry about with AI. Of course, it's more an indictment on a lack of UBI than on AI itself; to blame it on AI is a Neo-Luddite-esque argument that confuses the root cause the root cause with the more proximal cause. It's akin to people being mad at immigrants taking their jobs than the forces of globalism.
Quite the opposite. Art is a form of communication. Illustration however, is not always. Or the message is simple enough that non-human level AI can do it, as is the case with much commercial illustration.
> The point of art art is that it is created by a human or group of humans.
Not for me as a consumer of art.
If I read a fictional story, or admire a graphic I do not really care about author. Except as something that hints at quality. Maybe in extreme cases I would want to avoid some due to unrelated issues[1].
It is different for nonfiction, but I would read engaging story that was entirely automatically generated.
Yes, it is different for nonfiction and author can add relevant context that can increase or dimmish the art, but in vast majority of cases of fiction/music/graphics I know nothing about author and I am not really interested in them.
----
[1] I would not hang in my room landscape painted by certain painter who become the leader of Germany. Even if landscape would be pretty.
> > The point of art art is that it is created by a human or group of humans.
> Not for me as a consumer of art.
There's a sense in which you don't really get to say that.
Artists are the people who define art, much as scientists get to
define what science is.
So if someone said, "The point of science is that it is conducted
according to the principles of reason and scientific method", and you
responded; "Not for me as a consumer of science", then you'd be in a
sticky position.
You could certainly believe that amethyst rays will cure your gout.
But really you'll be a consumer of something else, something that's
not science even if you insist on calling it that.
Then artists may discover that art defined in this way is only part of what is considered as art nowadays. It is quite curious - is it half of it? 1%? 75%? 0.1%?
Also:
> Artists are the people who define art, much as scientists get to define what science is.
Not really, that is not how language works. As far as language goes, it is defined by more wider community.
See poor cryptologists and meaning of word "crypto", one more victim of other bunch of events.
And in that sense, there is no threat. People will still be able to make any of the things they do now as a form of self expression.
But much of art is created as a commercial work, and it is that market that AI may threaten.
This is true of any field where workers are at risk of being replaced. People aren't going to be prevented from engaging in that activity, but it will become harder or impossible to do so commercially.
That may be true of the kind of art that people go to galleries for, or hang on their walls and admire. But most professional artists make a living off of commercial work. If AI image generation advances far enough, many of those people could be out of a job. The purity of art won't save them.
Yeah, I don't really get it either. If the artist didn't matter, there would be a huge market for people who can simply copy the most popular artist's style and yet, from my perspective from the anime industry (where AI generation has been especially controversial), that just isn't the case. Companies still pay large amounts of money to get stuff done from the most popular artists despite there being plenty of people who can imitate the style.
Similarly, fans tend to be able to recognize the artist and follow their work too. AI isn't going to fill that hole because the artist does matter.
AI won't replace artist jobs, although it'll make new kinds of art possible.
I made a post [0] about trying to displace myself --not even at professional level but at hobbyist level-- and the results were underwhelming.
I would say it could be revolutionary if software did not already revolutionize the 2D-pipeline 20 years ago.
For now I think its impact is limited to an assistive role in some subtasks.
I've looked at Naoki Saitou's take on the subject [1] and he has a similar POV.
the flip side of this is that every airline company has, over the long term, been a net loss for investors.
also the airplane manufacturers require huge government contracts to even exist, as well as special treatment (like Boeing getting away with killing 300+ people) and periodic bailouts.
There's also bats who have much more flight control than birds. Since they have a hand in their wing which is covered with skin and muscle, they have a large amount of additional control over birds. Here's a great video of a study.
Are you "standing still" if you're sprinting at full pace, out of breath, heart pounding, but feet slipping on an icy slope so you have ground speed zero?
saying something can hover usually implies that it can have a ground speed of 0 at an air speed of 0. If you leave out the air speed requirement then anything that can fly can hover
That doesn’t make sense. If something only hovers at air speed zero along with ground speed zero, then as soon as any wind whatsoever picks up, then that something is no longer hovering by your definition, even if it maintains the same place in the air. Hovering means to stay in the same place in the air. It doesn’t say anything about air speed, which is the delta between wind speed and ground speed.
I don’t understand the weird arguments going on here. Hovering means to stay in one place in the air. It says nothing about air speed. Ground is the only reference needed for “in place”.
I feel we're talking too much about the article and AI and not enough about eagles throwing goats off mountains. Say what you will about drone warfare, at least you don't get much time to think about it.
Holy f**! I assumed the description of the video was a joke before I watched it. Now that's an amazingly specialized, if somewhat disturbing, hunting strategy.
Also, if they really want to succeed in the labor market, birds should teach more of their offspring to be like cuckoos. Everybody can be a cuckoo, if they work hard enough.
Right, so obviously tasks oriented around biological survival and production are different than tasks oriented around competitiveness in a modern economy. But the question does make me wonder:
What if we got AGI, failed to solve the alignment problem, and were just... fine? Because the AIs' goals were compatible enough with ours? We can't make a system like a bird, but if we did there's no real way it could spell our doom, considering how we're fine with all the existing bird-like systems. (Not the first time I've been skeptical of AI alarmism, but a different perspective on it.)
That is one potential outcome, but it’s really scary to bet our future on it being correct.
I think economic incentives will lead us to making that experiment, but there is a risk that it fails. There’s obviously vehement disagreement on the probabilities involved.
It’ll be like «well, maybe nuclear weapons start a self-sustaining fusion reaction in our atmosphere» again, but with lower reassurance this time.
UNIVAC predicts an Eisenhower landslide and people disbelieving it think some mistake occurred and then tried to make it be more reasonable. Starts about 1:35 picks up again around 11:35
The 1952 election is probably the most interesting one I know about. Truman attempted a 3rd term but then quickly backed out after Iowa, Estes Kefauver who was investigating crime bosses almost won the nomination but party bosses were friends with crime bosses and so a dark horse Stevenson came forth. There was talk of Eisenhower running on both parties ticket, a computer predicted the results, got it right but then everyone thought they had just made a mistake... There's so much there
143 comments
[ 3.6 ms ] story [ 213 ms ] thread[1] - https://www.youtube.com/watch?v=JK1dXuMEpT0
Intelligence is a mystery. As soon as we can build a rational system that demonstrates a particular kind of intelligence, the mystery and perception of intelligence is lost.
Citation needed. We had machines playing other games at better-than-grandmaster level by that point. People talked about AI's struggles to play another perfect information game with very simple rules because it couldn't brute force parse every single permutation as an indication of how limited it was, not an indication that true intelligence was pruning search trees.
Previously using tools was a sign of intelligence until we discovered that other animals (even insects) do that. Then it was required that you had to make tools, until Jane Goodall showed us that other apes do that. Until AlphaGo we had this idea of particular games, always upping the size of the decision tree. And now it looks like we are in a crisis and just have this vaguely defined general intelligence, a term with quite a racist history, that nobody actually knows what means.
Me personally am of the opinion that we need to ditch the word intelligence in science and technology (it can remain in philosophy). It has all the same problems of a grand unifying theory in cosmology, and a whole host of more (see Stephen Jay Gould to discover those).
Calling it a mystery is being optimistic, rather, I would call it a misdirection.
https://arxiv.org/abs/1703.10987
Not as funny as reality though, where AI is suddenly good at art and writing, but still bad at flipping burgers, or simply navigating floor of a house.
https://en.wikipedia.org/wiki/Moravec%27s_paradox
> Moravec wrote in 1988, "it is comparatively easy to make computers exhibit adult level performance on intelligence tests or playing checkers, and difficult or impossible to give them the skills of a one-year-old when it comes to perception and mobility".
That is, few compute resources are needed for a program that can beat human checkers players, as compared to what’s required to solve perception problems.
I wonder if you have a different thought of what “reasoning” was being discussed, as you switched the term to “high level reasoning”?
If I have a human and an AI go illustrate a manual for assembling my new sofa, or produce the advertising materials for the grand opening of a local restaurant, it's much easier for the AI to clearly produce an incorrect illustration in a way a human generally wouldn't.
In short, I mostly agree that good art is communicative, but also that some of that is because of the context where someone shows you something and tells you it's communicative and maybe what it's intended to communicate. There's plenty of famous works of art that don't look that different than programmatically generated NFT works, let alone things Stable Diffusion does.
People are really good at generating meaning and patterns from actually literally nothing. See conspiracy theories and the gambler's fallacy. If you know it came from an AI that intends nothing, it's easy to see the seams that tell you it's all pieces of meaningless nothing. If you're told that it's from some artist with a long history and passion for expressing some message, it's easy to read communication into anything.
Compare this bee stinger to a needle or the wings of a butterfly to the iridescent materials we can create today.
https://i.imgur.com/a7wMjlu.png
In your frame of reference consider as though God is building it through the human.
It still make no difference, whether we could still do it through conscious design than just evolutionary pressures which was what this discussion is really about.
It's both nature, but "evolution" took millions of years to build the ant, and "humans" could probably do it I a fraction of the time.
The picture is of a bee stinger compared to a sewing needle.
Someone in the comment section posted another comparison https://i.imgur.com/m2m6SV2.jpg
Or to borrow from Chomsky, why would we try to make a machine deadlift a box like a powerlifter when we can just make a fork-lift truck / front-end loader?
That said we already have drones. We can add a charging station, an artificial beak, a speaker and it can do bird stuff pretty well :)
What we suck at is building self-replicators.
Cats are considered better vermin hunters than simple machines, like rat traps, because they have a lot of extra benefits of being household pets. (Which rat traps don't do well. Also usually dogs are used to hunt vermin. [I wanted to find that video where dogs hunt on the field as a tractor plows it, but only found this about minks :o https://www.youtube.com/watch?v=MB1I7UixASY ].)
Fish tank cleaning is also a very specific problem. It's basically a niche. If we look at it as an engineering challenge it's a hard problem. Small space, hard to reach spaces, mostly mechanical work is needed (as fish tank life is very sensitive to water quality), etc. And there's also cost. It's very possible that if we were to spend a few billion dollars on the problem we could come up with a fish tank cleaner, but ... snails are provided by evolution for free.
On the other hand, the risk of ripping your entrails out while using a sewing needle is quite low.
Nature didn't anticipate the hairless apes to get so numerous and infiltrate every environment where bees exist.
https://www.youtube.com/watch?v=G-C77ujnLZo
In the sense that any natural development is intended, it's intended for the stinger to leave the bee. Compare wasps, which have reusable stingers.
Sticking it into the elastic material that covers mammals is not it's intended use case.
On swarm level it doesn't matter. But psychological effect on an attacker or even group of such is achieved after one event, sometimes even inherited.
Perhaps not economically at mass scale though.
But then again neither can bees.
So, we're a bit impatient but overall things are moving really quickly. I'm only 48 years old, I expect to see some amazing things in the next few decades.
One example of such criticism is this five year old piece: https://www.wired.com/2017/04/the-myth-of-a-superhuman-ai/ Few are skeptical about AI technologies’ economic impact. In that sense, the article misses the ball slightly, but it’s so funny that I don’t mind.
To state my point in the vein of the article: The bulldozer did change our economy. It does most, if not all, of its productive work in cooperation with a human. Turning a bulldozer on and letting it run by itself is usually a waste, and may be dangerous.
Thanks for a great, entertaining ditty. (I assume you're the author). I paused at the end, thinking it through, without immediately jumping to that conclusion. I'd say I'm a quite careful and sophisticated reader, so I love the way you left interpretation open.
Hence references to such future-proof technologies as delivering messages or cargo(!) via carrier pigeon.
Doesn't bode well for artists if they are essentially compared with carrier pigeons here...
Powered flight is enormously useful in a lot of domains, and yet for all its obvious superiority in many dimensions it hasn't even made infantry or shipping obsolete. And if I felt like hunting, I'd probably prefer a hawk...
Sure, but exaggerating the potential is much more common. Look at where people in the 50's thought we would be in terms of space or personal flying devices. This stems from both overestimating the technological capabilities, overestimating the usefulness if we were able to succeed, and failing to see the non-technological limitations.
Flight is actually a good example of the last one. For instance, the problem with Edisons prediction that people would fly to another city for a couple of hours after breakfast and be back by lunch isn't that technology isn't advanced enough. But the costs are prohibitive, the airports are often far from the city center, and even when you get to the airport it takes a couple of hours to actually get into the air.
Planes were hot, so Edison thought travelers "will largely scorn such earth crawling." But earth crawling, even long distance, is still very useful, hence the desire to invest in things like high speed rail.
[1] https://www.cbsnews.com/news/thomas-edisons-predictions-spot...
I recall such a talk by someone respected which, IIRC, included dissing the Ladies's Home Journal 1901 "What May Happen in the Next Hundred Years"'s[1] prediction "The horse will have become practically extinct." with a dismissive "The aren't!". Seeming underappreciating the time's massive horse footprint and diversity, now vanished and diminished.
> But the costs are prohibitive, [...] even when you get to the airport it takes a couple of hours to actually get into the air.
I wonder how predictions might deal with the diversity of timelines? And thus include the fork two decades back, where 9/11 didn't happen, the airline industry remained competitive, and I can still do a Boston-to-NYC short sprint from curb, across a narrow terminal, down the ramp, and squeeze through a closing airliner door, with no staff interaction beyond their shouting "run faster!". And everyone pays cash to stewardesses. And no ids (after all, showing papers to travel is an un-American Soviet thing). There's a lot of timeline diversity.
> [it] isn't that technology isn't advanced enough [...] the airports are often far from the city center
I wonder if there's a timeline where Robert Moses loved rail, smashed a straight network of lines across Long Island, and now NYC-anywhere to JFK is minutes. In contrast, there's the timeline where the massive and competitive subway construction of Edison's life, stalled out shortly after his death, and remained dead for a century. Perhaps the medieval warrens of European cities, persisting but for revolutions until aerial mass bombardment, might have been a hint. But new york was still doing massive "tear down and build new". Perhaps there's a timeline where a cold war nuclear strike on nyc cleared the way for greater development, as with the city fires of old. It does emphasize the contingency of history and thus progress, but it doesn't seem unreasonable that Edison didn't expect the trip from still-existing 1910-ish apartments in Manhattan, out to a "opened a generation after his death" Idlewild airport, to remain largely unchanged for a century. If you travel around Xmass, the train you catch might have a just-after-his-death 1932 vintage car. Current US journalism about China repeatedly reminds me of European writing from warrens about Edison's America.
Perhaps we need some expanded conceptual vocabulary for sketching possible futures?
[1] https://www.personal.psu.edu/staff/t/w/twa101/whatmayhappen.... https://www.mentalfloss.com/article/58851/fact-check-26-ladi...
There is some older footage of a similar thing but it's from a film and is all fake. The Frozen Planet footage is the first time it's been recorded for real.
There is also some Frozen Planet footage showing an eagle's view of flying over some mountains. That footage did use a trained eagle mainly because it's difficult to strap a camera to the back of a wild eagle and still have a face left.
Attenborough does not have his own production team. Most of the footage in Frozen Planet was produced by the BBC, but not all of it. Over 30 shoots were produced locally.
The Frozen Planet golden eagle footage is mixed and contrived, and it is easy to tell by the editing. The BBC shot all in ultra high definition, and some of the footage of golden eagles is in 8K and shot with a really decent professional camera. The drone footage is of a trained eagle, otherwise it would have taken down the drone. And the footage of the eagle taking down the kid and flying with the kid (for some reason we don't see it lifting the catch, probably because it is a stuffed prop) and dropping the kid is in 1080p shot with a shitty semi-professional camera with shitty lenses, upsampled to 8K and downsampled in post, as is evident by the grain. The eagle in the nest feeding the baby is not the same eagle, and it is not feeding the baby the kid.
A falconer named Jordi Isern[1] that shoots in the Andes shot the footage of his own pet golden eagle that he trained to hunt chamois (the goats in the footage).[2] That is Isern's footage that the BBC remotely directed and later edited to mix with their own footage. Wild golden eagles do not hunt chamois, and for good reason.[3] Chamois are incredibly resilient and tough creatures, while the bones of golden eagles are hollow. The diet of golden eagles is rabbits, hares, ground squirrels, prairie dogs, marmots, fish and carrion, all of which is plentiful, so there is no reason for the grave risk of taking such large and dangerous prey.
Trained eagles will hunt anything they're trained to hunt.[4] Isern is not the first to train a golden eagle to do this. There are other falconers in California (the first to do this as a proof of concept, shot on actual film before the advent of HD video, that I can not locate, lost in Google's shitty algorithm and the noise of the Internet perhaps forever, but it exists) and Spain[5] that have done the same thing. It's such a fantastic idea that the BBC and Attenborough couldn't resist, with Attenborough rationalizing that maybe kids were prey, but not adult chamois, thus Isern was directed to show the adolescent struck rather than an adult chamois being pulled off a mountain by its horns. Wild golden eagles don't fuck around wrestling with their prey. They always kill on impact.
Those falconers suck, and it sucks the BBC and Attenborough are involved. But it's definitely bullshit.
[1] https://www.facebook.com/groups/eaglesfalconry/posts/3802235...
[2] https://www.youtube.com/watch?v=XRLpk46OAA0&t=8m56s
[3] https://tinyurl.com/2nn7crcu
[4] https://www.youtube.com/watch?v=DPXAK2lJmEQ (LOL! fake.)
[5] https://www.youtube.com/watch?v=Yz7FFlFy8eM&t=4m52s
The point of art art is that it is created by a human or group of humans. Artists (performance or visual) will be out of a job when runners are all replaced by race-cars, and pro golfers, tennis players, football players, and basketball players by robots.
Not for me as a consumer of art.
If I read a fictional story, or admire a graphic I do not really care about author. Except as something that hints at quality. Maybe in extreme cases I would want to avoid some due to unrelated issues[1].
It is different for nonfiction, but I would read engaging story that was entirely automatically generated.
I admired fractals despite that human effort that went into https://en.wikipedia.org/wiki/Mandelbrot_set#/media/File:Man... was only part of outcome. If AI would find other interesting parts of fractal - it also would be enjoyable.
----------
Yes, it is different for nonfiction and author can add relevant context that can increase or dimmish the art, but in vast majority of cases of fiction/music/graphics I know nothing about author and I am not really interested in them.
----
[1] I would not hang in my room landscape painted by certain painter who become the leader of Germany. Even if landscape would be pretty.
> Not for me as a consumer of art.
There's a sense in which you don't really get to say that.
Artists are the people who define art, much as scientists get to define what science is.
So if someone said, "The point of science is that it is conducted according to the principles of reason and scientific method", and you responded; "Not for me as a consumer of science", then you'd be in a sticky position.
You could certainly believe that amethyst rays will cure your gout. But really you'll be a consumer of something else, something that's not science even if you insist on calling it that.
Also:
> Artists are the people who define art, much as scientists get to define what science is.
Not really, that is not how language works. As far as language goes, it is defined by more wider community.
See poor cryptologists and meaning of word "crypto", one more victim of other bunch of events.
If you threw away all the art and literature and waited 250 years, what you ended up with wouldn’t be a regeneration of that art.
If you threw away all the science knowledge and waited 250 years, most of the same science would regenerate.
But much of art is created as a commercial work, and it is that market that AI may threaten.
This is true of any field where workers are at risk of being replaced. People aren't going to be prevented from engaging in that activity, but it will become harder or impossible to do so commercially.
Not only that! People creating as a hobby typically want an audience.
And unlike weightlifting it may be really hard to have people interested in process rather than in outcome.
Similarly, fans tend to be able to recognize the artist and follow their work too. AI isn't going to fill that hole because the artist does matter.
AI won't replace artist jobs, although it'll make new kinds of art possible.
https://www.youtube.com/watch?v=Kj96nxtHdTU
I made a post [0] about trying to displace myself --not even at professional level but at hobbyist level-- and the results were underwhelming. I would say it could be revolutionary if software did not already revolutionize the 2D-pipeline 20 years ago. For now I think its impact is limited to an assistive role in some subtasks.
I've looked at Naoki Saitou's take on the subject [1] and he has a similar POV.
[0] https://woolion.art/2022/11/16/SDDB.html
[1] Will the AI Take Over the Job of Illustrators? https://www.youtube.com/watch?v=C1Un6D_POk
Is there any doubt that human “intelligence” will soon be surpassed by machines. It doesn’t even need to be super intelligence.
We seem to be oblivious to how unintelligent we are.
For example, “Science advances one funeral at a time” because of our limitations.
also the airplane manufacturers require huge government contracts to even exist, as well as special treatment (like Boeing getting away with killing 300+ people) and periodic bailouts.
https://youtu.be/BNNAxCuaYoc
And there's kestrels hovering. I have also seen hawks do this.
https://youtu.be/7j6OsP7zL6w
https://youtu.be/mDRcLAkRZ50
Edit: Now I've gotten on a YouTube binge and landed on New Zealand keas, which I didn't know about before. And damn are those birds intelligent.
They're also utterly fearless, to the point of being disturbing.
I'd never had a wild bird come up and approach to less than arm-length, and not flinch or worry about what I was doing with my camera/tripod/etc.
I backed up, repeatedly, and it just kept getting closer.
...and whether they taste any good. Yes, I'd heard all the stories about Kea destroying cars.
https://www.youtube.com/watch?v=c6vY0s83NLg
> Goat was just minding his own business, singing to himself, "Love lift us up where we belong. Where eagles cry on a mountain high."
I think the question answers itself.
Maybe it can happen, but is it possible that this video is mostly faked?
*****
What if we got AGI, failed to solve the alignment problem, and were just... fine? Because the AIs' goals were compatible enough with ours? We can't make a system like a bird, but if we did there's no real way it could spell our doom, considering how we're fine with all the existing bird-like systems. (Not the first time I've been skeptical of AI alarmism, but a different perspective on it.)
I think economic incentives will lead us to making that experiment, but there is a risk that it fails. There’s obviously vehement disagreement on the probabilities involved.
It’ll be like «well, maybe nuclear weapons start a self-sustaining fusion reaction in our atmosphere» again, but with lower reassurance this time.
Sure: https://www.wearethemighty.com/popular/pigeons-for-sea-rescu...
Example: https://m.youtube.com/watch?v=5vjD0d8D9Ec
UNIVAC predicts an Eisenhower landslide and people disbelieving it think some mistake occurred and then tried to make it be more reasonable. Starts about 1:35 picks up again around 11:35
Background https://m.youtube.com/watch?v=nHov1Atrjzk
The 1952 election is probably the most interesting one I know about. Truman attempted a 3rd term but then quickly backed out after Iowa, Estes Kefauver who was investigating crime bosses almost won the nomination but party bosses were friends with crime bosses and so a dark horse Stevenson came forth. There was talk of Eisenhower running on both parties ticket, a computer predicted the results, got it right but then everyone thought they had just made a mistake... There's so much there
Fisher: https://en.wikipedia.org/wiki/Cormorant_fishing
Hunter: https://en.wikipedia.org/wiki/Falconry
Racer: https://en.wikipedia.org/wiki/Pigeon_racing
Mount: https://en.wikipedia.org/wiki/Common_ostrich#Racing
Fortune teller: https://en.wikipedia.org/wiki/Ornithomancy
Fighter: https://en.wikipedia.org/wiki/Cockfight