Okay, putting a robot through an industrial cleaning process after each job is a lot easier than doing it with a prostitute, though.
No, it isn't. You wouldn't put your iPad through such a process. Imagine how much more sophisticated a robot prostitute would have to be.
You're making a category error here. You're seeing the features of durability, crude specialisation, disease resistance, and cheapness to be features of machines. They're not, they're the features of relatively simple constructions, however arrived at. A simple biological tube/rob would be no more or less durable, washable, disease resistant or cheap than a machine equivalent with the same characteristics. All that makes something durable and easy to clean is it being simple. The 'machiness' or 'robotness' is irrelevant.
What I'm saying is that any machine that has the same or similar features to a human is going to share the same issues as the human, because those issues arise not from being biological, or being human, but from being complicated.
They're also self-repairing on the cellular level. If you get the technology to create a self-repairing robot, you probably have the technology for an infinitely upgradeable human.
We already have the technology to create self-repairing robots.
No, we have robots that can replace parts of themselves given a pile of conveniently located spare parts.
Only if the robot modular, which implies only a loose connection between a mental and physical function. That has quality implications.
Not really. There's a different between a robot that is truly modular in all its functions, and one where you can simply slap another chassis on top. The latter isn't really modular, and it's no different than putting a different chassis on a car.
Given that automotive chassis technology has progressed to the point that slapping a new chassis on an existing car has vast performance implications, I'm inclined to agree with you.
How do you know? Human downtime is healing and reprogramming time. A robot as sophisticated will have similar requirements.
...and how would *you* know? There's absolutely no reason to think that a robot that sophisticated would have similar requirements.
Yeah, there is. It's because those requirements aren't arbitrary, they are a function of the performance characteristics involved. If you have a sophisticated self-repair mechanism that relies on onboard power and can conduct even cosmetic repairs, you're going to need some downtime to make best use of it. If you have a sophisticated neural social modeller, like the human brain, then it's going to need some time to analyse inputs and construct neuronal changes. The requirements are driven by the performance characteristics.
Giving a robot the impression of having hobbies and interest may be harder than simply giving it time off to the develop them.
But we already have AI's that can do that pretty well.
No, they can do it pretty well for AIs. They're still fairly dull conversationalists. It's like a dog riding a bike - very impressive, but not because the dog is going to outrace anyone.
How do you know? Have you built one?
No, but others have. You do realize robots used for sex already exist, right?
Not to anything like the sophistication of a prostitute.
Because they are mirroring the functions and designs of a human. You're airily assuming that wouldn't come with downsides like it does in humans, but I can't see any principle that you're following to reach that conclusion. It takes people 18 years to reach the stage of acting like adults. On what basis do you assume that a robot designed to achieve the same functions as a human wouldn't take the same time?
...However, let's for the sake of argument assume that it still takes 18 years. Okay, well then it just takes 18 years *once*. After that, you can make as many copies of the AI as you want.
You could, but having one universal sex-bot isn't the same as having a unique and individual prostitute. Again, what robotics does do is allow you to cut corners. If you think that your line of bots would do better without real personalities, reliant on replacement parts when they get damaged, needing regular deep industrial cleaning, and all looking and acting exactly the same, then you can make substantial savings. But plenty of people will claim that your bots aren't as good as the real thing, and on purely objective grounds, they'd have a point.
But biology is just very complicated construction, nothing more. What makes you think that your equally complicated construction will be better? 'Made of plastic' isn't a superpower.
Because again, we can improve upon nature's flawed designs. We can already create artificial muscles far stronger than human muscles.
This, I've not heard of? With the same size, weight and performance characteristics? Do you have a cite?
We can create artificial skin far stronger. We can create materials that regenerate themselves more efficiently than human tissue can.
Again, not heard of this from my contacts in neurophysiology and AI research. Citation please?
For short periods of time. But they are still watched, maintained and supervised by humans who can reprogram them as necessary. Unless you have an example otherwise?
That's moving the goalpost.
No, it's really not. If you want robots replacing humans, you need examples of robots replacing humans. If you want to claim that robots are cheap, and durable, then the robots that actually replace the humans need to be cheap and durable. Having a demonstration of a robot under very carefully controlled conditions doing a subsection of the tasks a human could under very close supervision and a vast support infrastructure doesn't really come anywhere close.
Humans are largely limited by the laws of physics, and their performance is far higher than cheap durable materials will take you.
That's a big assumption that is already not true. The performance of the material we're made of is NOT 'far higher' than many cheap durable materials we have today (to say nothing of those we'll have in the future). Human performance is only superior right now as a total package. Once you break us down to our individual components there are very few parts of us that don't have a technological equivalent that has superior performance.
Well, sure, I'm not claiming that humans are better at holding up buildings than a steel bar. But the total package matters. My PC is, in many ways, strictly inferior to an abacus at adding up small columns of numbers, when it comes to cheapness, durability, virus resistance, environmental survival, electricity dependence and speed. In fact, there's not much a PC is actually best at it, but we keep them because they can be used for a wide range of tasks.
If that's the distinction you're relying on, then robots will only ever replace humans in narrowly defined specialised tasks. Which is a lot. But to replace humans entirely we still need some reason to believe that capabilities that humans have, in the proportions and trade-offs that humans have, will somehow be better done by robots. And that's where I'm expressing scepticism.
I don't know that we're a million miles apart here. You're saying that a great many tasks will end up being performed by robots. I'm agreeing with you. I just don't think that they all will.
We've been here before. We have a vast swathe of local government run sports and leisure facilities in this country, because planners back in the 1970s were concerned that automation would lead to us only working 3-4 hours a day, and that we'd need a way to fill in the extra time. All the same arguments you're making now were made then. But it didn't work out that way.
