> In my country (Australia) all plumbers require a license and i would expect them all to be experts.
Good for you.
This didn't work out for the stevadores whose (human) licensing requirement in some ports meant their ports became obsolete and the majority of shipping moved to other ports.
> I dont think you appreciate that doing any skill such as plumbing is more than just watching a video. Nobody is going to become a brain surgeon or plumber just because they wear any sort of glasses or goggles or new age technical equipment.
I didn't say "just watching a video", I said "AR glasses", and I write that for a good reason: augmented reality, even the poor versions we have now, can draw things in your field of view and walk you through stuff.
A video from a language tutor will teach you a certain amount; someone using Google Translate's augmented reality mode which they copied from Word Lens after buying it 15 or so years ago, that's a very different experience: high chance you learn little to nothing by using it, and yet the user can hold up their phone to some Arabic or Chinese text they absolutely never studied and see those words written in their own language on their phone's screen.
Similar with taxis: a huge fraction of the population needed to learn to drive, so when the specialist skill of "remembering a city's whole map" was rendered obsolete by cheap GPS, Uber was quick to follow, and Uber (etc.) drivers are famously not well compensated.
For me, plumbing is an illustration rather than a prediction. Given some (future!) AI and AR, this is how it could turn from a high-skill role to a low-skill role. But hey, we might get those robots first, and then it's a "humans need not apply" role. I think general-purpose robotic AI is at least a decade off, possibly more. Workplace AI assistance is already being rolled out, the question is the scale and impact, which could be anywhere from "why did we bother?" to "well that whole profession no longer exists" and everywhere between; we will only be sure after the fact.
And also note: you're the one introducing brain surgery here, when I'm staying on plumbing. Surgery also needs dexterity/steady hands (which robots already have) and quite an unusual personality to be psychologically able to cut someone apart in the first place (yes for their own good, but even done right it will look like a first-person horror movie), so when this tech is good enough for that, it's a fully robotic surgeon not some guy letting themselves be a meat puppet for an AI.
> I think you fail to comprehend, that the market value of AI is over a trillion.
> THe only way they can even pay the interest on those investments is to charge EVERYONE thousands, and thats my point.
Lol no. 9% US GDP is about 3 trillion USD/year; I'm saying this only works out if the world is buying from the USA, at which point it's $3e12/8e9 = $375/person/year.
I don't think it gets distributed equally over all jobs, and even 9% is pushing optimistic due to 60% of world nominal GDP (about 50% by population) being in countries which are extremely suspicious about the US these days, but even then it's not "thousands".
For example: you can eliminate about 1.3 trillion USD/year worldwide just by displacing the role of "software developer", just wages and similar direct to employee compensation.
Reducing plumbing from current wages to minimum wages would be, IDK, perhaps $100e9/year? Something like that. But as I say, illustration. You want a sum, pick whichever jobs (or more precisely, tasks) you think are most vulnerable to AI (not just today's AI but what you think will come to exist over the time it takes for these investments to get paid off) and add them up.
(Also, market cap != investments. I absolutely expect the market caps to undergo a 'correction' and come down significantly; but this is a separate question to interest on loans etc.)