Project GR00T by Nvidia
twitter.com
twitter.com
But as a thought experiment, let's say this project or projects like it are a big success, costs come down enough to enable mass production, and within a few decades (sooner?), we have fully autonomous humanoid robots that can do useful tasks.
Realistically, what's next? I try not to get too caught up in fictional apocalyptic scenarios, but given what we already know about how the assimilation of technology into society is going, I can't help but feel pretty worried.
Isn't this a worrisome direction to be going? I love the potential use cases for assisting people who need assistance. But beyond this, it's hard not to feel deeply pessimistic about a wider deployment. They'll be expensive enough to be controlled by corporations, and you don't need some terminator or super-intelligence scenario for such bots to have a serious impact on society. I'm less worried about them getting out of control, and more concerned about the entities that control them.
What I would like to emphasize is that this is a problem of political economy rather than a problem of technology. I think most would agree that which future we end up with depends on if androids end up costing $100,000 per unit or $100 per unit. If the latter case ever arrives we will pass through the former case anyway.
Society may need to consider charging for that license and divvying up the proceeds. Not enough to discourage innovation, but enough to prevent society from devolving into the Eloi and the Morlocks.
The fact that these technologies seem to be exclusively the playground of well-resourced corps and governments does not give me too much hope for a positive outcome.
Regardless of the outcome, it should be a fun ride.
But i wouldn't give very great odds.
And even today the second transition is still not obvious to everybody. There are still a lot of people today who think jobs making stuff are "real jobs" and service sector jobs are lesser.
I guess that begs the question - how? The transitions you mentioned, as you said, eliminated a ton of jobs. This appears to be no different in the sense that computers and robots will do more both traditionally "skilled" and "unskilled" jobs alike. Perhaps the notion of what a job actually is, or how people view work. I don't know.
Once the population isn't growing, the only way to increase the level to which peoples' needs are met is through productivity increases.
Imagine how much labour would be required to meet the needs of every human to the level at which only the mega rich currently enjoy.
Now imagine that, as those needs are met, new needs and desires evolve.
The work is never ending.
Welllll Nvidia manufactures everything in Taiwan so I woudln't be so sure about China there.
Have you bothered to review what Waymo has accomplished? It's certainly not level 5 autonomy, but within its constrained service areas, it absolutely works.
In large parts of Phoenix, you can download an app and summon a driverless Waymo vehicle.
The future is here, just unevenly distributed.
The car drives to your destination without a driver, while avoiding the interstate.
It reminds you that your luggage is in the back, and won't let you say you're done until you take it out.
There's also a tablet in the front where you can watch it see different obstacles, people, and cars.
It's pretty cool, and it's very good at a lot of the harder technical aspects of driving, but it doesn't like bushes, and can get lost when it has to make turns/park close to bushes
While anecdotal, my experiences in Waymo have convinced me that Level 5 is possible in my lifetime. Tesla's latest update is also a huge step forward, and I no longer am assured that they won't "pull it off" - although I still have some doubts about them specficially. I still don't trust Tesla's current solution, but it is no longer in a dire state.
There is a long way to go with self-driving, but it took us a very long time to go a very short way. Recently , we have gone a very long way in a very short time.
There will be new jobs, that are going to be more engaging and make better use of our brains.
I am not worried about "robots", but about corporations wielding too much power.
We are probably at the last moment when oligopolies can be broken up and limits set so that corporations can't grow beyond certain size and they can't acquire other businesses.
Then we need to reform investment and pension funds, so that they collectively cannot have a controlling stake in any company.
If there's one thing that past 10 years or so has taught us, it's that at least in the US we are completely okay with oligarchy, and we have no viable mechanism in place to thwart it. At least not after "Citizens United."
We fear AIs. What if they don't do what we want? They are unpredictable. What if they turn on us, what if they rise up and kill us? The ruler fears the commoner in exactly the same way. That's why they have always been paranoid about maintaining control, monitoring and censoring communication, restricting access to information. My guess is that many of them would prefer AIs to humans, AIs might be more predictable and easier to control.
What do they need humans for? To protect them, fight their wars, grow their food, make their robots and AIs, and take care of all their needs. Once machines can do all that, a lot of them would feel safer and happier without people around shitting up the place, sharing their resources, spreading diseases, posting things on wikileaks, voting, protesting, revolting.
There's going to be interesting times ahead. Whatever happens it certainly won't involve the have-nots with unlimited leasure time, courtesy of those who control the machines.
Trump does seem like an old school egomaniac who wants to be loved and rule people and have statues of himself. But his kind isn't so common in the west and he is not well liked by the more typical ones who take the cold, scientific approach to wealth extraction and don't seem to care in the slightest how the commoners view them.
If your question is "what's the point of anything when a robot could do it," your real question is "what's the point of life," which is an exercise left to the reader.
Maybe in factories? However, the main reason to have a humanoid robot is to have it fit in places designed for humans. Factories can be designed around robots, so the human-shaped advantage isn't there. The robots we currently see in factories are only arms with tools, I don't see that changing much.
I do think smarter robots will take off, but I don't think this form factor will be as important as it seems right now. Maybe just for novelty or entertainment.
I will want one in my home when I'm old and weak to the point that the alternative is either institutionalization or letting strangers do its job.
People said similar things about those early NNs generating images, code, music.
Surely we will solve the fine motor skills soon. It just takes one incremental step at a time, there are no fundamental roadblocks on the way.
> No one will want one in their homes.
I would kill for a robot that keeps my house clean, makes my favorite food, fetches me things, etc.
While I could pay a human to do that, it just seems more convenient if that wasn't a real person. Because with a real person I can't e.g. walk naked around them and behave generally weird, because I need to respect their feelings, which I don't have to do with a robot.
> People said similar things about those early NNs generating images, code, music.
Yes and… they still are expensive, bulky, and creepy. The expense and bulk is externalised to the stock market (but maybe improving), the creepy is endemic and not getting any better.
An expensive/bulky GPU doesn't have to be built in the robot body itself — it could be a box mounted somewhere in the house with the wireless connection to the robot body.
I am sure at some point there will be hardware that would protect model vendors from extracting/copying the model weights, like on-chip encryption when doing inference or something.
Or maybe there will be GPUs with homomorphic encryption¹, so everything will be end-to-end encrypted, and it will be safe to compute in the cloud without compromising user privacy.
This is an incredibly bold claim, and mostly just sounds like your personal opinion towards humanoid robots. Why do you expect economies of scale to not apply here?
> No one will want one in their homes.
I am literally prepared to pay a car's worth of money for a robot, humanoid or otherwise, that can meaningfully take care of the drudgery of household chores.
I think that's the key. I like my Roomba, and I didn't have to take out a loan to get it. Humanoid robots will always be more expensive than a Roomba. Short of heavy subsidies, no amount of scale will change that.
It seems like a human-like form the best way forward for robots that will assist in the home -- pretty much everything in the home (including humans themselves) is designed around the human form factor.
This always irked me about sci-fi and robotics in general. It’s been demonstrated by Boston Dynamics and others that different forms are better like forms based on dogs and insects, and also no need for artificial lighting with other types of sensors available.
For example, if a factory was staffed with robots it’s only illuminated for the humans and maybe cameras.
I think when you're young you might not see the whole picture, but you're still fresh and untainted by greed and way the world works.
We're already seeing software wrote software, we're kind of seeing autonomous machines and by the time I'll die we'll have machines growing our food.
But the monetary crisis will need to be solved. The accumulation of wealth being the main goal of everyone will destroy us in the end.
We could already feed the planet, but we aren't because of money and what it means. It means power over other people. Power to change things, power to live a comfortable life where others need to work.
If we can get machines to work for us, at we won't need to subjugate other humans, which might lead to a better world. But the way there, if those with money don't want to let go of their control of other humans and additionally use machines to extend their control even further, this will be a bloody page im human history.
It could go either way.
Figure 01 robot demos its OpenAI integration - https://news.ycombinator.com/item?id=39691886 5 days ago, 112 comments)
Because the robot in the video looks very similar to the figure 01, but maybe this is just a coincidence.
More seriously, gun toting autonomous platforms are already in use - no human form is required. Sure there is a human in the loop - for now. But even on things like our last generation of cluster munitions from 20 years ago they could select their own targets on the way down (no human in the loop after the thing was dropped).
I doubt it, you can already fuck torso-sized pieces of silicon, why would adding bones and verbalization add to that experience?
No way would I ever fuck silicon, no matter how desperate I was