HN Theater @HNTheaterMonth

That Tim Urban piece is great. It's also an interesting time capsule in terms of which AI problems were and were not considered hard in 2015 (when the post was written). From the post:

> Build a computer that can multiply two ten-digit numbers in a split second—incredibly easy. Build one that can look at a dog and answer whether it’s a dog or a cat—spectacularly difficult. Make AI that can beat any human in chess? Done. Make one that can read a paragraph from a six-year-old’s picture book and not just recognize the words but understand the meaning of them? Google is currently spending billions of dollars trying to do it. Hard things—like calculus, financial market strategy, and language translation—are mind-numbingly easy for a computer, while easy things—like vision, motion, movement, and perception—are insanely hard for it.

The children's picture book problem is solved; those billions of dollars were well-spent after all. (See, e.g., DeepMind's recent Flamingo model [1].) We can do whatever we want in vision, more or less [2]. Motion and movement might be the least developed area, but it's still made major progress; we have robotic parkour [3] and physical Rubik's cube solvers [4], and we can tell a robot to follow simple domestic instructions [5]. And Perceiver (again from DeepMind [6]) took a big chunk out of the perception problem.

Getting a computer to carry on a conversation [7], let alone draw art on par with human professionals [8], weren't even mentioned as examples, so laughably out of reach they seemed in the heathen dark ages of... 2015.

And as for recognizing a cat or a dog — that's a problem so trivial today that it isn't even worth using as the very first example in an introductory AI course. [9]

If someone re-wrote this post today, I wonder what sorts of things would go into the "hard for a computer" bucket? And how many of those would be left standing in 2029?

[1] https://arxiv.org/abs/2204.14198

[2] https://arxiv.org/abs/2004.10934

[3] https://www.youtube.com/watch?v=tF4DML7FIWk

[4] https://openai.com/blog/solving-rubiks-cube/

[5] https://say-can.github.io/

[6] https://www.deepmind.com/open-source/perceiver-io

[7] https://arxiv.org/abs/2201.08239v2

[8] https://openai.com/dall-e-2/

[9] https://www.fast.ai/

⬐ mdp2021

> And as for recognizing a cat or a dog — that's a problem so trivial today

Last time I checked - though it's been a long while I could not check thoroughly owing to other commitments - "«recognizing»" there was "consistently successfully guessing", not "critically defining". It may be that the problem was solved in the latest years, I cannot exclude it - but I have not seen around in the brief "news checking" exercise the signals required for the solution.

The real deal is far from trivial.

A clock can tell the time but does not know it.

I agree with the neighbor comment, 2015 was a long time ago. See the state of the art today:

Agility Robotics Digit:

https://www.youtube.com/watch?v=Jycdks836bY

Boston Dynamics Atlas:

https://www.youtube.com/watch?v=tF4DML7FIWk

Neither of those robots use a static walk. That is obvious enough in the above Boston Dynamics video, but Agility Robotics also just released a little video explaining their (very robust) dynamic walking system: https://www.youtube.com/watch?v=mpxrnrR_Tsg

It seems clear the basic hardware to do humanoid walking is technologically achievable today. Personally I would think Tesla would be better suited by copying or buying Agility Robotics as Digit is a smaller robot and it looks cheaper than Atlas too. Better for working in a warehouse alongside people as shown in the top video link in this post, and they seem a smaller company than BD and so cheaper to buy. Musk is very happy to buy firms that produce tech his companies need.

As far as the software problem, it is a big one to be sure. I don't think they have all the pieces to the puzzle, but I do think Tesla's Dojo AI system has promise for training a machine like a humanoid. Their system is designed to ingest experience and produce neural nets which can navigate in the real world. They may be willing to take a bet, despite the clear risks, on using this system for humanoids. If they buy Agility Robotics, rev the design to add their AI system (which is extremely powerful AI hardware with their own custom processor), they could be in production on prototypes within a year.

This all has the disclaimer that success is still very hard and not guaranteed, but then that's the kind of thing Musk loves to throw resources at.

Also: I do not work for any companies in this field. I am making a farming robot (see profile).

The robot I am most personally impressed with is Agility Robotics Digit. It is a real human like form factor (small human not huge!). If I recall correctly it is an energy efficient walking system too.

https://www.youtube.com/watch?v=Jycdks836bY

This long continuous shot of walking from 2019 is nice too. They have improved their software a lot but the hardware was already in good shape.

https://www.youtube.com/watch?v=_YrlR1iNVcQ

Meanwhile Boston Dynamics is doing quite well.

https://www.youtube.com/watch?v=tF4DML7FIWk

Best path to success is Tesla buys Agility Robotics, revs the design for Digit to integrate the Tesla AI system, and spins up a manufacturing line. They could focus immediately on the high level AI problem rather than worrying about design. Could slap their AI system on digit while mechanical teams rev the new design so they get moving quickly.

Disclosure: I do not work for any of these companies or any related company. I am making a farming robot (see my profile).

Reusable rockets weren't practical either. But someone was maverick enough to attempt it anyways. [1]

You provide valuable information but I just hate how you succumb to easy thinking, an appeal to tradition [2]: "well it hadn't been done then so it can't possibly be done now." It's a form of anachronistic gatekeeping. It doesn't belong in the age we currently reside in.

We have totally different computational abilities today, powerful beyond reason. Our hardware is smaller and better performing. We can perform digital simulations billions of times before we even execute one physical experiment.

I've seen the same attitude with age-old mathematics problems and it makes even less sense in that domain. The ability to perform computational experiments that our predecessors couldn't, changes the viability of almost every unsolved problem.

Boston Dynamics would have never created Atlas or Spot if they clung to common tropes about robots. And now, due to battery technology, the once tethered robots are dexterously mobile. [2]

[1] https://www.fool.com/investing/2020/10/05/how-much-cheaper-a...

[2] https://en.wikipedia.org/wiki/Appeal_to_tradition

[3] https://youtu.be/tF4DML7FIWk

⬐ zakary

It is always good to question assumptions. One you could consider is the assumption that all problems are created equal. This isn’t the case. There are fundamental and practical limits to things.

For this kind of artificial muscle actuator, there are fundamental limits to the way the physics works for this design which prevent it from ever simultaneously having good efficiency in energy and power density, high strain and high frequency operation. There are certainly ways to make artificial muscles work, but hydraulics is not the way.

For rockets and space vehicles, the limit has more to do with technology and material science.

⬐ goldenkey

Material science is part of hydraulics. The hydraulic fluid could end up being next-gen.

⬐ cma

Reuse has been a goal since the at least 70s, with the Shuttle project, Buran, later the DC-X, etc. It wasn't one guy maverick enough to attempt it. And it wasn't given up on and then revived by one guy. Blue Origin, Armadillo, Scaled Composites, etc. were all attempting it. It is the largest scale (comparing heavy vs shuttle, though heavy hasn't seen much use) and seemingly most successful, and several other firsts.

The upper stage still is discarded, but if they can pull off the Starshuttle it will be huge and the first operating at anywhere near that scale with full reuse.

⬐ robomartin

> I just hate how you succumb to easy thinking, an appeal to tradition

No. I appeal to PHYSICS.

This isn't an emotional argument. Or one out of "tradition".

The reusable rocket was a question of implementation, not physics. The math actually supported the idea of being able to do that. It wasn't some magical juju bean Musk pulled out of his behind. They did the math. The math proved you could do it. They built it. It took a lot of trial and error and, yes, failure, but they KNEW the math supported the concept.

This inflatable muscle thing (air or fluid) isn't supported by the PHYSICS.

What do I know anyhow? I've only been in robotics since the early 80's. I have literally seen <insert fantastical claim> in robotics and artificial intelligence since that time. And here we are, we still don't have a robot of any kind that is as capable as a house fly, even with a supercomputer running it remotely.

And that, BTW, is the test for fantastical actuators. The technology has to support an amazing range in terms of scale. At one extreme you have something like an ant and, opposite that, an elephant. Even within a human being the range of size and strength in muscles is likely 1000 to 1 (haven't done the math).

> Boston Dynamics would have never created Atlas or Spot

I don't think you have a good sense of what these machines are and are not. Hint: They are not what you think they are.

⬐ goldenkey

No, your appeal is to the "cult of science." Science is a moving target. Hydraulics involves materials known as hydraulic fluid. Material science is in its infancy. Your appeal is nothing more than to pessimism and the edifice of cocksure.

⬐ robomartin

Nice one.

⬐ goldenkey

As my senior, you should have a bit more tact instead of outgassing petty sarcastic quips. I think the baby boomer generation has actually become somewhat of an ironic label because these people never really grew up, effectively remaining emotionally stunted babies in perpetuity. I'd love to hear your thoughts on this scientific theory.

What do you suppose those computers are doing, then? ;)

https://www.youtube.com/watch?v=tF4DML7FIWk

https://www.youtube.com/watch?v=tfb6aEUMC04&t=115s

⬐ PeterisP

I would go a step further and look at the research on medical devices for spinal-damage patients which try to literally control their existing muscles through a computer; the concept of neuroprosthesis and https://en.wikipedia.org/wiki/Functional_electrical_stimulat... working to design algorithms on CPU to move muscles, as the parent post words it.

We said the same thing about movement in robots -- and today we have: https://www.youtube.com/watch?v=tF4DML7FIWk

Just ridiculous how far things have come, it seems fake it's so far along.

> Is there a better player in the industry in building humanoid robot?

Boston Dynamics https://www.youtube.com/watch?v=tF4DML7FIWk

⬐ kator

They talk a bit about how the video is made here: https://www.youtube.com/watch?v=EezdinoG4mka

The video I wanted to see (the two robots actually doing their thing): https://www.youtube.com/watch?v=tF4DML7FIWk

What's that dripping from the right robot's right thigh near the end?

https://youtu.be/tF4DML7FIWk?t=49

Also are the layers of foam over the landing surfaces they backflip onto needed for dampening?

⬐ nofunsir

Sweat, clearly. Hardcore parkour’s not easy.

⬐ nemothekid

The robots are powered by hydraulics. Might be a leak.

https://www.youtube.com/watch?v=EezdinoG4mk

⬐ tantalor

Sir, are you aware that you are leaking coolant at an alarming rate?

https://youtu.be/z7U3x3uRmi8?t=57

This is the video with the blog post: https://www.youtube.com/watch?v=tF4DML7FIWk

To me, I like to see when they stutter or slightly-fail, since that to indicates the dynamics and real-time part of the thing, that it is not a 100% scripted thing (although the blog post mentions the engineers fine-tuning the celebratory arm-pump, to what extent is that scripted then?)

⬐ shadowgovt

It's interesting to see how this has progressed from their video two years ago [https://www.youtube.com/watch?v=LikxFZZO2sk] and one year ago [https://www.youtube.com/watch?v=_sBBaNYex3E]

⬐ electricwallaby

I noticed around the 50 second mark there is some fluid leaking out of the robots bum. Wonder what that could be?

⬐ juancampa

https://en.wikipedia.org/wiki/Hydraulic_fluid

⬐ jcun4128

Ha would be something to see it in Ninja Warrior show or something one day, needs hands.

⬐ scythmic_waves

That video is partly CGI, right? I can't find a smoking gun but it doesn't look entirely real.

⬐ dane-pgp

As one of the comments on the video says, that feeling is an example of the "uncanny mountain" effect.

⬐ scythmic_waves

I'm not so sure.

1. [EDIT: IGNORE THIS POINT] They've done this before. [1]

2. To get more specific, look at the robot's legs around 0:09-0:16. They don't seem to be moving normally.

I don't have professional video analysis chops, but I'd put some money on this being fake.

Happy be to be proven wrong, though.

EDIT: It appears I remembered this story incorrectly, and posted a link without reading it. I retract my first statement. I'll leave the link for posterity, though.

[1] https://www.theverge.com/tldr/2019/6/17/18681682/boston-dyna...

⬐ mastax

Have you seen the behind the scenes video? https://www.youtube.com/watch?v=EezdinoG4mk

⬐ dane-pgp

That provides some great context, but I couldn't help finding it a little unsettling at 03:09 when one of the robots is suspended off the ground and "bleeding" a pink fluid onto the mat beneath it.

⬐ ttmb

> 1. They've done this before. [1]

Your link says nothing of the sort.

⬐ tempestn

How much money? I'd probably take that bet, if we could agree on an arbiter.

⬐ mastax

That CGI gag was done by Corridor Digital, not Boston Dynamics.

⬐ ansible

That's a great channel, I really like their "reacts" series that comes out every Saturday. They also had a good one from Sunday about the Pentagon UFO videos.

⬐ kingsloi

yeah definitely odd looking. I thought maybe just like 5k@60fps or something and my peasant eyes were just bad, but deffo odd looking

⬐ KMnO4

Most likely it’s real. That said, we don’t know how many 10s-100s of takes this was done in. This single video could be the results of hundreds of hours of fine tuning parameters.

⬐ levng

https://m.youtube.com/watch?v=EezdinoG4mk&feature=youtu.be

This behind-the-scenes video says it is choreographed and scripted. It also contains instances where things go wrong as well.

⬐ rozab

dupe of https://news.ycombinator.com/item?id=28210113

Doesn't HN scan for exact dupes?

⬐ udfalkso

The way the robots moved in the movie "I, Robot" actually feels a lot like this. Impressive and scary.

⬐ xchaotic

the more they make the robot make 'cute' things like aerobatics, the more scared I am. Uncanny valley.

⬐ pikseladam

I still hope there will be easy way to disable these when the time comes for humanity to fight to survive. Please don't make them fix each other.