With AI coming, there is some general advice to go into skilled manual labor like the trades since a lot of knowledge work will be automated in the years ahead and worker productivity will go up to the point where white collar jobs can get the same amount of work done with less employees.
But I don’t feel the skilled trades are safe either.
The end goal is bipedal robots powered by AI, which will have the manual dexterity to perform skilled manual labor.
But another factor will be the widespread adoption of augmented reality. A situation where a person wears transparent glasses that overlays information about their environment onto the glasses. Companies like boeing already use augmented reality to help in their manufacturing and its improved speed and quality. Its common for people to go on youtube and look up info to fix things around their homes, which is like a rudimentary version of augmented reality.
I feel like thats just the start. In a few years we will probably be able to take someone with 1 year experience in a skilled trade and by giving them augmented reality glasses, have them perform at the same skill level as a 10+ year veteran. They could open a car hood and have step by step instructions given to them which they would be able to understand because they have the basic knowledge.
I just don’t see how the trades are a bastion of job security in the age of AI. In between augmented reality in the medium term and bipedal robots in the long term, those jobs will disappear too.
The only jobs I could see staying human are ones that are human due to regulation, or jobs that require in person human to human contact due to the fact that we are evolved as social creatures. Some forms of therapy will always require a human being, but even in those situation, the therapist will probably have feedback from AI about the emotional needs of their patient so the patient and therapist can communicate better.
Which to me is one of the underrated aspects of AI. I think it’ll bring people closer because it’ll help people improve their social skills. People will be more able to understand and express themselves, and better able to understand others, and better able to find effective social and emotional outlets due to AI.
Or most of the people will be emotionally stunted droids raised by machines.
ISTM a lot of what’s plaguing the COVID generation is too little live human contact and too much machine-mediated contact. Turning the entity at the other end of the wire from a real human to a fake human is IMO unlikely to improve the outcomes.
I feel like AI will give us guidance and advice for in person interaction. If you’re talking to a girl, AI can tell you if they like you or find you annoying based on things like her body language, tone of voice, etc. If you’re in therapy, AI could tell the therapist what kinds of traumas could be behind your dysfunctional behaviors. Stuff that the patient may not feel able to verbalize but the therapist can help coax it out of them. AI can help people write emotionally complex letters to people discussing emotionally complex situations so they can communicate their feelings.
Also AI could help us understand dysfunctional people to steer clear of them. People with untreated cluster B personality disorders can be very destructive. AI could give real world feedback that someone is giving subtle signs of something like ASPD or NPD, and let the person know to keep their guard up around them. For someone with an abusive childhood who is oblivious to red flags, AI could point out the red flags that they miss because they are unconsciously repeating childhood abuse.
Also if something traumatic happens, AI will be more likely to pick up on subtle signs of it. So hopefully things like rampant sexual abuse will become less common as AI helps the people around the victim understand something bad may have happened.
In real world usage, I’ve already met people who do things like put someone else’s text messages into AI programs and ask it if it sees signs of manipulation. Or people will write letters, then put it in AI and say ‘take out all of my anger and passive aggressiveness’ and the AI program will do that. So its already happening.
The only “safe” jobs will be CEOs and political leaders, since those are the people whose interests the AI will be designed to serve.
Unlikely. Any such AI will be carefully designed to turn people into submissive servants of the wealthy and powerful, not to help them. It’ll sabotage interpersonal relationships by design, since people who work together are a danger to the system. It’ll be designed to spread racial hatred, hatred between the genders, and hatred of anyone who isn’t wealthy for that purpose, as well as to suit the prejudices of the elite.
The result will be a populace who interacts almost entirely with their circle of machines and their boss, and who is taught to hate and fear everyone outside that circle.
And once the technology gets good enough that they are no longer needed, they won’t be able to provide meaningful resistance when the elite has the rest of the population exterminated.
[quote=“Der_Trihs, post:5, topic:1020511, full:true”]
The only “safe” jobs will be CEOs and political leaders, since those are the people whose interests the AI will be designed to serve.[/quote]
This is true sadly. I’ve wondered how this will play out with stockholders, boards of directors and executives. The shareholders may want AI led companies since this will lead to better business outcomes, but the board of directors and executives will want to keep their jobs.
I’ve wondered about this too, but there will also be open source AI too. Also we may end up in a cold war scenario too where two nation states use AI for different purposes. In the cold war they had the capitalists vs the communists. But in reality any modern AI cold war will be between sociopathic, plutocratic fascists in the US and sociopathic, autocratic dictators in China.
Its kind of like saying that once cars come about, they will only be used to service the interests of the rich and powerful. But eventually the technology will be available to the masses too. However superintelligence is different, because an AI 3 years more advanced will make the older AI largely useless. And if the plutocrats, autocrats and fascists control the most advanced AI, then who knows if people can fight back.
I agree this is the AI world that the rich and powerful want. One where AI makes them rich and powerful, and keeps everyone else divided and oppressed. Then when the rich and powerful don’t need us who knows what they’ll do to us. But I don’t think the entire world will follow this path. I could see the US following this path, but I think other NATO nations would lead to some kind of post-scarcity economy. I feel like as far as wealthy democracies go, the US is an outlier in how we will handle AI. I think other wealthy democracies will transition to post scarcity societies, while the US moves towards fascism.
Also I wonder how blue states and urban areas will respond to AI. Even if the federal government wants AI to work for the rich and powerful, there will likely be pushback from blue states and urban areas to make AI work for everyone.
Yeah, it could take a few decades before we have affordable, highly functional bipedal robots. But by comparison, here are some robots from 2015
But in the medium term we will have augmented reality. So that’ll create a labor surplus in the skilled trades before bipedal robots come about.
Another thing we will have is robotic exoskeletons to help people perform blue collar jobs without doing as much damage to their muscles and joints.
I just feel like the mentality that turning to skilled labor is a good move isn’t true. In between more people joining looking for work, augmented reality, robotic exoskeletons, etc the supply will still outstrip demand and then in a decade or two we will have bipedal robots that are affordable anyway.
Due to a large skilled worker shortage, a similar situation already exists within the skilled trades. It’s common for a company to have, lets say 500 technicians of various skill levels. The company will then have 10 Technical Specialists. The specialists take calls from inexperienced technicians and talk them through jobs that would normally be done by experienced technicians. Facetime is pretty commonly used for this.
Sometimes the technician has never even seen the piece of equipment he will be working on before. It’s just a series of “Test this and tell me the value. Send me a photo of this. Do this and see if the situation improves, if not replace this part.”
I could see AI taking over this role. I just don’t think it will be replacing the technicians. It will be replacing the specialists.
My thought was that people would have AI glasses that can understand the machinery it is looking at and give step by step instructions to the worker. Doing this would allow blue collar workers with very little experience to perform more advanced work. This will make it far easier for people with 1-2 years of experience to perform at a master skill level. This will increase the labor supply of blue collar workers, driving down wages and job security.
Not necessarily. After all, everything you’ve said up to this point has also been true about the IT age in general. I agree that AR will be a boon compared to trying to search through some shitty Windows help file format bag of crap (can you guess if I’ve had to do this?), but it’s still just another step along that path. And for skilled blue collar workers of the kind you are describing, salaries have increased above inflation (depending on what country we’re talking about).
I’m not suggesting we shrug off such advances, just saying that the argument needs more than just “higher productivity therefore fewer, lower paid jobs”.
If you’ve got a leaky pipe, you need to have some pair of arms reaching underneath the sink to fix them. Eventually, there will robots that can do that. But they’re at least not going to come out of the current sort of research into AI.
Is it? If I’m running Jophiel Plumbing, I don’t need a bunch of C3P0 looking androids installing mainlines, I just need something with some appropriate pipe-manipulating appendages and the ability to do the job with the approximate maximum footprint of a person (though that can be flexible). Having it balance on a couple of legs and wear a hat is just a waste of resources. 50% of my time doing plumbing is spent cursing my human body with its large hands and need to lay upside down under a sink, trying to fit a piece into place or opening up a wall to access a valve. Life would be a lot easier if I just had spider limbs with nine points of articulation, rotating attachments and visual sensors on my wrists.
I would imagine that, at some point in the future, you would just have your plumbing, electric, HVAC, etc installed with monitoring systems that then communicate to a service robot when something needs physical replacement or repair. But, again, there’s no good reason for it to be human formed and plenty of good reasons to NOT be a bipedal android thing.
The built environment, as completed, is designed for bipeds about 5.0 to 6.5 feet tall. Robots working in that environment will be most successful if shaped the same.
But …
The process of building that built environment has lots of places and lots of processes where the human form is a misfit. There’s a reason construction work involves a lot of ladders, pliers, kneepads, and awkward body positioning. There is nothing about drywalling a ceiling that is ergonomically suited to the human form.
Punchline:
The robots best suited to building a house, or maintaining its systems will look very different than those best suited to cleaning its interior, cooking in it, or tending to the humans in it.
They are more subtle than that. They won’t exterminate them, just force them into unemployment, not use the money made from AI to do some sort of payments to them. Because, they don’t have jobs because they are lazy.
The blue states will tax the profits on AI, and pay for medical care and food with it. The red states will let the people starve or drive 200 miles to a hospital. It’s starting already, and with the Medicaid cuts I suspect a lot of hospitals in red states are going to close.
While researching a chapter on ambulances I found that there is thought about putting VR googles in ambulances so that doctors in emergency rooms can direct the EMTs to do emergency treatment - and even perhaps have gloves that will direct the hands of the EMTs. I don’t think these things exist yet. I can see them being proved in on cars and plumbing.
First, once they have their own fully automated economy and army, they don’t need to be subtle anymore. Second, extermination is extermination whether you do it with guns or mass starvation. And third, letting 99% of humanity starve to death gives that 99% time enough to organize, arm themselves and kill off the 1% instead; they can’t afford to let that happen. They are outnumbered hundreds of millions to one, they’ll have to overwhelm and kill the rest of humanity before that happens.
I think this assumption undervalues the benefit of direct experience as well as assuming that the ‘expert system’ is sufficiently reliable to match up with a “10 year veteran”. In fact, with physical skills the psychomotor aspect of being able to perform complex manipulations is as important as analytically diagnosing problems and understanding the procedure for fixing or manipulating a system. I can explain to you how to pick a lock, how the pins or wafers work and how you progressively bind them to line up the shear line but until you pick up a pick and tension wrench and spend the time to practice you will not be able to spin that core regardless of how much guidance you get from AI or augmented reality. The same holds for many physically complex or delicate skills.
There is a tendency by advocates of autonomous robots to discounts the experience, judgment, and manipulative abilities of skilled and semi-skilled labor but while there have been substantial improvements in the dynamic mobility abilities of bipedal and quadrupedal robots, doing actual complex manipulations is still a challenge, not so much in being able to emulate movements but having fine tactile feedback and control, and despite the ability to entrain complex knowledge into transformer-based models, operating as open-ended autonomous agents requires a high degree of confidence in the reliability of a system. A lot of skilled trades are going to be relatIvely safe from replacement by robots although semi-autonomous aid systems may enhance the safety, quality, or productivity of human labor.
As others have noted, while there is a lot of focus on bipedal robots because of the supposedly easy integration into human labor roles, it’s not the ideal form for a lot of physical work in construction, fabrication, maintenance, et cetera. Our ‘built environment’ is shaped about bipeds but much of the liminal spaces and underlying infrastructure is not really designed for easy access by people, and anyone who has worked on modern engines can attest if would often be advantageous to have four hands ending in pseudopodia or plier jaws. A general purpose welder-bot that could get into awkward locations that a human would have poor reach or work in overhead positions would be much more efficient and safer. One can certainly come up with many examples of work for which a non-bipedal form that would be advantageous, even possibly for jobs like a nurse or agricultural field worker.
