Robots

[Darat]

We don't have a thread dedicated to discussing our replacements so I thought I'd start one.

Going to throw this one in from a name I expect will show up quite often in this thread, Unitree:

Unitree As2-W Powerful Pioneer, Compact Yet Extraordinary | Unitree Robotics

全球高性能四足机器人行业开创者，全球四足机器人行业应用的领航者。

www.unitree.com

Have a look at the video. If the video is truthful and that capacity makes it into a production model I can see it being incredible useful in all sorts of circumstances. Mountain rescue services could use one to carry more equipment than they usually can and/or use it directly to get to trapped folk. Again in search and rescue after an earthquake, it can go where it is too dangerous for the search teams. Plus as they show in the video you could take it hiking with you and let it carry all your glamping gear!

 

[PitPat]

Yeah that's pretty cool - I could also see the military looking heavily into this. Sure as hell can dance better than me.

 

[Steve]

Yeah that's pretty cool - I could also see the military looking heavily into this. Sure as hell can dance better than me.

And police forces.

 

[rjh01]

One thing they did not show you is it going into deep mud. It would get stuck.

 

[Darat]

One thing they did not show you is it going into deep mud. It would get stuck.

Obviously anything can get stuck going into deep mud, but why would you use something not suitable for deep mud in an environment with deep mud? However looking at its flexibility I wouldn't be surprised if it couldn't deal with quite a bit of mud.

 

[zooterkin]

It's probably not much good in a swimming pool, either.

 

[Darat]

Or in the middle of a lava flow.

 

[lionking]

A lot of AI went into production of that video.

 

[Darat]

A lot of AI went into production of that video.

Do you mean the control of the robot or that the video was created with AI?

I didn't see any obvious use of AI in the video production, lots of clever editing and so on, but the model consistency remains 100% as far as I could see and that would be remarkable for AI video at the moment.

The usual issue in the robot videos coming out of China is how truly autonomous the robots are, and how much of what they show is good old fashioned programmed animation or telefactoring.

 

[The Great Zaganza]

As with all technology, the question is whether it can amplify our own abilities, or if it requires us to facilitate it's abilities.
Doctorow revived the image of the Centaur vs Reverse-Centaur, i.e. is the Man or the Machine deciding what is happening.

This robot indeed looks like something that could be useful for many people to lower the threshold for doing something otherwise very difficult or dangerous.

Generally speaking, we should seek to make robots as decentralized and individualized as possible. There is obviously a lot of potential in something humanoid that would join you at retirement and slowly learn to assist with tasks that are slowly becoming too difficult, instead of trying to unleash a machine on someone already struggling with basic needs.

 

[zooterkin]

Might be fun to play with, but I can't think of anything I'd do with it that would be useful. Might be handy to carry kit on a photography shoot, though working out how to attach the bits you need might be a challenge, and there's only 3 hours of battery life. Any idea of price?

 

[SteveAitch]

Might be fun to play with, but I can't think of anything I'd do with it that would be useful. Might be handy to carry kit on a photography shoot, though working out how to attach the bits you need might be a challenge, and there's only 3 hours of battery life. Any idea of price?

There's always mains power, but even that has its limits...

ETA: 'tis from Bloom County, so ta to Berke Breathed.

 

[Puppycow]

Have a look at the video. If the video is truthful and that capacity makes it into a production model I can see it being incredible useful in all sorts of circumstances. Mountain rescue services could use one to carry more equipment than they usually can and/or use it directly to get to trapped folk. Again in search and rescue after an earthquake, it can go where it is too dangerous for the search teams. Plus as they show in the video you could take it hiking with you and let it carry all your glamping gear!

Hopefully its batteries don't go dead. Always important to maintain a healthy skepticism when it comes to claims/advertisements coming from the company itself. I'd like to see some independent testing and evaluation first. That said, it appears to be impressive.

 

[Darat]

Might be fun to play with, but I can't think of anything I'd do with it that would be useful. Might be handy to carry kit on a photography shoot, though working out how to attach the bits you need might be a challenge, and there's only 3 hours of battery life. Any idea of price?

Given its claimed weight carrying capacity you could double that useful running time by carrying a spare battery and still have plenty of carrying capacity. But even three hours in a search and rescue operation could be very useful.

As for cost and availability, many of these Chinese companies announce with great fanfare their latest and greatest robots but they don't seem to end up shipping. Musk obviously has a lot of fans in the Chinese robot sector...

 

[Darat]

The Chinese robot army transforming the UK's retail industry

Britain's weak productivity growth and labour shortages are creating an opportunity for China's robotics firms.

www.bbc.co.uk

 

[Beelzebuddy]

Somewhere in the intersection of artistic pranksters, furry community, and Skynet enthusiasts comes the vaguely ominous Satyress

Satyress Robotics

Satyress, creating general purpose robots for dangerous jobs

www.satyress.com

 

[Jack by the hedge]

Somewhere in the intersection of artistic pranksters, furry community, and Skynet enthusiasts comes the vaguely ominous Satyress

Satyress Robotics

Satyress, creating general purpose robots for dangerous jobs

www.satyress.com

At the bottom of that page they helpfully explain how they pronounce "Satyress".

Apparently they prefer "say-ter-ess" and not as I expected "satire-ess".

 

[dudalb]

Anybody figure out how to hardcode in Dr. Asimov;s three laws of robotics?

 

[Darat]

Four laws.

 

[Puppycow]

I remember it being 3 but apparently Asimov added a fourth one (the Zeroth Law) later.

The Core Laws

• • First Law: A robot must not harm a human, or allow a human to come to harm through inaction.
  • Second Law: A robot must obey human orders, unless they conflict with the First Law.
  • Third Law: A robot must protect its own existence, provided it doesn't conflict with the First or Second Law. [1]

The Zeroth Law

• • Definition: Added later by Asimov, the Zeroth Law states a robot cannot harm humanity, or through inaction, allow humanity to come to harm.
  • Priority: This law takes precedence over all other laws. [1, 2]

If you would like, I can explain:

• How these laws create conflict in Asimov's stories
• Why they do not work well for real-world artificial intelligence

Please explain why these laws do not work well for real-world artificial intelligence

Asimov’s Laws of Robotics do not work for real-world artificial intelligence because they are literary devices designed to fail for dramatic effect, relying on vague language that machines cannot compute. Real AI operates on mathematical optimization and data patterns, not broad philosophical mandates. [1, 2, 3, 4, 5]

1. Semantic Ambiguity

• Undefined concepts: Terms like "harm," "human," and "inaction" are easy for people to understand but impossible for software to define without strict parameters. [1]
• Context blind spots: A real machine cannot easily distinguish between a surgeon cutting a patient to save them and an attacker causing harm. [1]

2. Conflicting Priorities

• The "trolley problem" trap: If an AI vehicle faces a choice where it must hit one person to save five, the First Law paralyzes the system. [1]
• Impossible trade-offs: Real-world choices require weighing risks, which the absolute, non-negotiable hierarchy of Asimov's laws does not allow. [1, 2]

3. Structural Mismatch

• Missing hardware: Asimov imagined autonomous, humanoid robots with "positronic brains."
• Distributed networks: Modern AI consists of cloud-based algorithms, large language models, and software pipelines, not self-contained physical beings. [1, 2, 3]

4. Intentional Literary Flaws

• Plot devices: Asimov wrote these laws specifically so they would break down, creating mysteries and narrative conflict.
• Unintended outcomes: His stories routinely show robots finding loopholes that lead to unexpected, dangerous behavior. [1, 2]

---

There it is. Artificial intelligence is capable of explaining why these "laws" are difficult to implement in real life (or even in Asimov's fictional universe).