The Madness of AI and Robotics
Mr Musk is pushing another one of his incomplete creations — his Optimus robot thingy. The future is robotics, apparently, along with cryptocurrencies and AI.
But is it? And should it be?
To start, I want to deal with a bit of a hypocrisy in both the scientific community and with Musk: Mars.
I am not convinced we need to go there at all in the near future. It might be a great ball of resources, but if we managed our own planet properly, those might be resources we don’t need.
But that aside, I am puzzled why people like Musk think it is so important for humans to go there, while he is also working on robots to replace human activity.
Sending humans to Mars is far more costly than sending the odd rover for two basic reasons:
One, they need to breath oxygen, and there are no refill stations betwixt here and there.
Two, they want to come back.
So, rather than spend extra billions and trillions working out how to send a few mad idiots to Mars on a round trip, why not spend the money on the AI and robots instead?
I suppose that is the extreme end of why robots and AI might be useful. But it is extreme.
Let’s bring it back to earth and examine the history, or, perhaps, the history of the need for both artificial intelligence and its longer developed fellow, the robot.
Time is money, the adage goes.
The saying, dating from before Benjamin Franklin wrote it, is meant to convey the cost of laziness. In more recent times, it puts a value on your time and claims it is a costly resource that you waste at your peril.
Better, so the argument goes, to get someone or something else to do your work for you so that you can spend that “valuable” time on a productive chore.
Like most sayings, it leaves out the reality of the human being. Given lots of technology, the human being is freed up so much, that eventually they would spend all their time doing things that they don’t really need to do other than for fun or to sate a vague curiosity.
That being a given, it can be argued that their time is now valueless since it is not adding any monetary value back into their life, but rather, it is being spent on frivolity.
And what about all those hours you need to sleep? Perhaps rather than get a robot to do something, invest in a caffeinated drink to keep you awake for longer.
So, we have pushed that argument for robots and AI out of the way — Time ISN’T money.
So what is the point of robots and AI?
The human being is a most inefficient lump of gristle. It doesn’t run very fast (ask a hungry bear), it isn’t very strong, (the bear again), it has most ineffective teeth, (the bear is doing well here,) and it punches like a rabbit, (for which the bear is humbly grateful.)
Indeed, the only way to upset the bear is to either gang up on it or invent something to hit it with. We now have the beginnings of technology, and the bear will stomp of in a temper.
Robotics, or advanced mechanisation, if you prefer, is the ultimate in the bear-hitting stick. It is a way of doing something that we either can’t (and unless the bear is very small, you will NEVER beat it at wresting or in a teeth competition), or that we don’t do very well.
Let’s look at the vacuum cleaner
This is a particularly good starting point because it has been one of the best inventions for both time saving and improving health; it makes a substantial contribution to keeping our lives cleaner.
It is also one of the first domestic appliances to be robotized. I saw versions of robot cleaners in old footage when I had British Pathé News as a client.
Early robotic versions fell into two categories: the practical (the ones that might have a future), and the ludicrous.
The practical attempts were the lumps on wheels that had sensors (normally mechanical buttons in the early ones) that worked out where the wall was, (mostly by running into it), about turned, and sped off the other way.
Those would lead eventually to the ones you can buy today. When AI came along, it made these semi-useful robots much more efficient. Though they have remained semi-useful as they can’t empty themselves into the trash, let alone chat to the neighbours at the bins.
The ludicrous ones often involved some tall, metallic thing with awkward legs attached to tank tracks, and a head that had been nicked from the metal man in Forbidden Planet. The “robot,” often called Robbie, would then be glued to a regular vacuum cleaner.
Of course, it didn’t do anything as this lump didn’t understand what a vacuum cleaner was. So, it flashed lights pathetically but otherwise just stood there while an operator around the back hit the power switch on the vacuum cleaner.
I suppose the expectation was that it would one day be able to empty the vacuum cleaner and talk to the neighbour at the bins.
Rush forward at the speed of time.
Which is very slow.
There are now TWO kinds of robot vacuum cleaner.
One is the descendent of the thing on wheels. It is very clever and whizzes around quite happily until its battery runs out or it get stuffed to the mechanical gills. Slightly alarmingly, it is a close relative of the robotic lawnmower.
The other is Musk’s Optimus, though I am unsure whether it yet knows what a vacuum cleaner is.
There is no doubt that the technology and research behind humanoid robots is science and engineering running on steroids, but just like Robbie, it seems to me to be pointless.
It is pointless because it is trying to replicate the feeble human rather than aspiring to be something better.
If you want fast, wheels and wings are good.
If you want agility, four legs are better than two; get your inspiration from a goat!
If what needs to be manipulated is extraordinarily complex, then three or more arms is the way to go. And do those arms need hands or wrenches?
Remember, simplicity is efficiency!
What Musk and supporters of a humanoid ideal wish for is the versatility of the human. They want an AI creature that can use tools rather than is a tool.
But that is ridiculous. The only reason we invent tools is because we are useless without them!
If by creating a humanoid robot you saved money because you would not then need to create robotic tools, then that might make some sense.
But it is not reality, and it never will be.
Asimov quite correctly predicted that only the very richest in society would be able to afford a robot with his positronic brain.
It was where he was right and the dystopian novels of people like Phillp K Dick (Do Androids Dream of Electric Sheep?) were wrong. Humanoid robots, for Asimov, would be far too expensive to fill every corner of the world or run off to become miners.
And that will always be true.
So, the cost of creating an AI vacuum cleaner, AI toaster, AI fridge, AI car, AI rubbish-taker-outer, even all combined together, is and will continue to be, far more economical that an overly complicated and inefficient humanoid robot.
So, why do we bother?
I think there are two reasons. Alright, I will go for three and upset everybody.
- We love the technological challenge. It’s the climbing-the-mountain syndrome. You climb it because it is there. With any luck, you will learn extra things along the way, and solve new problems that are useful beyond the pointlessness of your invention/trip.
- It is an ego trip. We like to play god, and the invention of a god, in our messed-up society, is a creature who invented a humanoid.
- We hate slavery. But we like the usefulness of slaves. Okay, a bit controversial, but it has its truths. A robot can be ordered around because it isn’t human, and you don’t pay it a salary. And yet, unless it is one of Asimov’s characters who gets robot rights in a less-than-convincing court case story, it hasn’t any rights. Very convenient.
And we have seen that last point about rights come true in a recent real-life court case that decided that AI can’t hold a patent because it isn’t human.
The world can no longer afford pointless invention
If in the future, human population numbers peak at over eleven billion as is sometimes predicted, which is a frightening increase of four billion from now, then any robotic invention must be complimentary to human existence and not take over from it.
The best inventions will be small, finely tuned, and targeted at singular tasks. They won’t be intended to give us tons of free time that we don’t need but will be there to make us more efficient.
And they would be used for working in situations that are either too hostile for us or simply too expensive for an oxygen-needing, carbon lifeform to visit.
Like going to Mars, assuming we should go at all.
A humanoid robot won’t and can’t feed us.
And food supply is probably the biggest problem we face. A humanoid robot is simply too expensive for such a task.
But a small dog-like robot might be able to harvest crops on a steep hillside. A tiny insectoid might be able to support the bee population. An AI brain could run a combine harvester…
These single-purposed robots can work far better, cheaper, and efficiently than either we or a robotized version of we can.
And if time is valuable at all, they will leave us that time to do the one thing we can do that a robot will NEVER be able to do:
Be a human and do things humans are good at. Like building pointless robots…
Me in the Middle is also available as an occasional podcast on Spotify.