Comment by MarkusWandel | Hacker News Reader

Comment by MarkusWandel | Hacker News Reader

The computers are coming for jobs ever up the white collar scale. When I started working there were (barely) still secretaries who typed and filed things for their bosses, and quite a few geeks had jobs that involved assembly language programming.

AI will take jobs. Super frustratingly, it'll probably make call centers even more useless (has anyone ever gotten anything useful out of one of those ("Hi, I'm ____. Can I help you?" popups at the bottom right of web sites?) And it'll certainly automate some of the "copy/paste" type programming jobs at the lower end of the scale, the same way email automated a lot of secretarial jobs; i.e. 10 fulltime copy/paste programming type jobs may become a job for one human and an AI assistant.

Which leaves people who are really passionate about and good at their craft. Somewhat relatedly, I saw about an uptick of people going into the trades. AI won't take plumber or electrician jobs away in the foreseeable future.

I see the plumber/electrician thing a lot.

But talking to my friends who do these jobs it always seems like it would be even more vulnerable to AI than programming.

Experienced electricians get paid decent wages because they have had lots of training and then have seen loads of different problems. So they intuitively know things like 'This is a 1960s house so if there's a problem with the lighting the first thing I should check is the fuse box connector, it should look like xyz, etc. Etc.'. This seems like exactly the sort of thing an LLM could do for them.

I think you could easily see a world where an electrician is someone on minimum wage with very minimal training who blindly connects wires as instructed by an AI.

I reckon the safest jobs are ones with limited reliance on knowledge and a very high level of physical skill (in environments where it's hard to get machines to operate). Bricklayers, plasterers and painter/decorators will be the big earners of the 2050s!

I wonder if there's enough info about how to do tradesmen's jobs online for that to happen. Programmers are at risk because we filled the internet with free training materials but many jobs aren't like that especially anything with a physical component.

I don't know if it is enough yet but there has been an explosion of this kind of content on Youtube over the last 10 years. For typical home repairs it seems most topics are pretty well covered.

This is an interesting point. A family member of mine is what we call here a medical evaluator - not sure if it has a direct equivalent in e.g. the US and how it is called here, but those are doctors who assess the disabilities of workers who apply for a pension due to illness or accident. This involves exploring the patient and then making the decision and justifying it in a report. The latter two seem like tasks that LLMs should be able to do easily.

However, we tried a description of a fake case to see what Bing could do, and it couldn't do much. And I think the reason is that there are very detailed documents on the rules that they follow for their decisions, but these are not online - they are in a private intranet and they can't take them out of there. If Bing had access to those documents I don't think it would have much of a problem.

So maybe a way for workers to protect themselves from being replaced by AI is not uploading much information about their jobs to the Internet... I wonder if this will lead to a situation like guilds in the middle ages, treating job skills essentially as secrets.

The most recent electrician jobs we've had done were:

- fitting a timer into the switchboard to control the hot water cylinder. A simple job, but the sparky also had to talk to me (the client) to get us both on the same page.

- fitting an EV fast charger in the garage. Not much science, but a lot of cable running and clipping down, then the garage switchboard needed to be swapped out for a larger one that could take the required RCD. And convincing me which brand charger to go for. 2 guys working together for a couople of hours.

- fixing the range hood light (always on due to a broken switch). He spent quite some time trying to extract the broken switch, with the range hood balanced on his shoulder and wires everywhere.

In every case there was no real complexity to the job, not the sort of thing that an AI could have been helpful at at all. Just a lot of common sense, knowledge of the regulations and much skilled manual work.

I don't think AI is coming for electricians any time soon.

But in all of those cases presumably someone needed to figure out what needed doing? (In your case maybe you're savvy enough that you knew what the issue was and just needed a certified person to do the work, but most clients won't be).

My argument is that it is the 'figuring out' that drives electricians wages, not really the doing part. Because while clipping down cables and extracting switches is fiddly work, I'd argue it isn't a skill with enough barrier to entry to maintain high wages (as compared to brick laying or plastering, for example, which you simply can't do to a professional level without years of practice).

So most of the value delivered by an experienced electrician is in talking to clients and identifying the correct technical solution, and is therefore pretty much analogous to the value delivered by software developers.

Therefore if we accept the logic that software developers will no longer be required (or that their value will be greatly diminished) it's hard to see how that wouldn't apply to electricians too (in the sense of being a well-paid trade over and above your average manual job).

(Btw - I DON'T think either will happen, but I just think electrician is a weird choice of example for those that do think that)

You know what a tradesman does today and an AI cannot? Get under the sink, undo that rusted bolt, route that cable in that awkward position and so on.

Too many people make the mistake of thinking there’s an infinite number of sinks to get under.

There is though. In Europe finding a plumber that will take you can have you wait weeks pricelessly because those sinks in existence keep breaking down.

Advanced robots will be able to do that in less than five years. Inexpensive ones in less than 20 years.

AI can only automate things for which a training database exists

> I reckon the safest jobs are ones

The safest jobs are ones that honest to self for the doer. He/She will be able to create value either using other humans, or machines and continue to do.

Sorry, I wanted to try "safest jobs are ones that involve politics", while those will always be present, it is not the safest and wont be many available, so changed to more abstract answer.

There is no reason to expect robotic technology to halt. Look at what things like Tesla or Boston Dynamics robots can do. Eventually we will see very well articulated and high strength to weight ratio robots integrated with advanced AI systems. It is definitely not going to take 25 years.

If you look at what's happening today, in 25 years it seems plausible that fully autonomous superintelligent androids with much more dexterity than humans will be fully in control of the planet.

> AI won't take plumber or electrician jobs away in the foreseeable future.

I see this argument way too often. How many electricians and plumbers does the world really need? And if the market becomes over saturated, how much will those jobs pay? How often will you actually have a job to do?

> And if the market becomes over saturated, how much will those jobs pay?

AI makes everything else so cheap, that by just working for a small bit of time, you can afford all the necessities?

That's been the promise of increasing productivity for many decades, and yet...

if you were to life with only what was around at the time such productivity increases occurred - e.g., no internet, phone, and high speed computing, access to medical/transport improvements etc - you'd probably be able to live off minimum wage easily.

People worked more today because they desired more. In fact, the desire outstripped their ability to increase their wages!

That is a common counterargument, and yet the common counter-counterargument matches my experience:

That sadly doesn't apply to housing or food, at least in the places I've lived. E.g. houses in the place I was born are now 5x the price, once you factor in inflation. Food and vehicles are 2x. Higher education went from "free" to $30K.

The really important things are far more expensive now, across the three countries I've spent decades in. Maybe the overall situation is different for the US of A.

I think we've already seen the dividing line drawn between people who can ultimately be replaced by a computer (or at least, those whom top-tier management believe can be replaced by a computer) and people who won't: Work from Home.

Reply on news.ycombinator.com