I still think that all jobs are as a general safe for the foreseeable future. But we will he expected to use AI tools and just produce more and more, so that a few people will gain more and more resources and power.
E.g. as engineers we will do less and less actual planning, but we will run AIs like it were a team of engineer slaves.
And I think this will be similar for other branches. A music composer will run AIs to compose parts of a song, adjust it, readjust other parts, till the song is good. I mean, afaik this is already how much of it works.
Am I supposed to read this as simultaneous (those jobs are currently safe… for now, the others are not) or progressive (all these jobs are human/skilled and halfway they get replaced by robots)…?
I suppose either way it’s commenting that you can’t take your position for granted. AI isn’t coming to replace you, but it is going to evolve your field, and workers that don’t adapt will be supplanted by those that do.
Drivers were on the edge for a long time. Lawyers are on the edge for the past 2-3 years. Cooks are probably the closest ones to be on the edge too.
How drivers were on the edge?
The rich will always have money to pay better people to make beautiful things for them
Just be useful to the rich and you’ll survive
Just like they planned it
I’d rather make them fertilizer
I just watched a movie (Geostorm) where these obviously super wealthy people were in a skyscraper and the movies like “oh no, they might die if no one stops this!”
Good? I’m more concerned about all the people below them getting swept away. These rich fucks should finally feel fear for fucking once.
Zero argument here
Everything can be automated, just with lower quality, speed, and a high up front and ongoing cost.
But for a large segment of jobs, no one cares about quality. Speed can be increased by increasing the number of parallel automatons, thus cost. If you really want to get rid of all work, raise the minimum wage to $100/hour for one year. Don’t tell anyone that it will only be a year. By the end of the year, almost every job will be automated.
How safe a profession is depends on how much more expensive replacing robots are than replacing people
I ve seen robot in exbibition failing just because of working all day, never forget maintainace also
When I see these kinds of posts I just look over at the vibe coders and just laugh harder than any joke about ai taking our jobs
Except Vibe-Coders are kicking back & sipping margaritas & your job is still gone
I was extremely skeptical so I looked into it and it absolutely does not work. There was also a guy on YouTube who basically tried to make a Minecraft clone with Vibe coding and it just fell apart almost instantly.
All I was trying to do was get it to set up a basic scene in UE5 with some lighting effects and import a model of the building from the assets library. Nope, did not work. I didn’t even bother trying to implement game logic as it was so clearly a waste of time. The amount of time I spent trying to get it to do basic stuff, stuff that you would be able to do in UE5 after half an hour of training, I could have made significant progress on a gray box by then.
Personal is a career?
Probably a hallucination of the AI that generated this
I assumed it was supposed to be Personal Assistant, but the text got cut off.
I wanted robots to do my menial unpleasant chores for me so I’d have more time to do art, writing, and analytics. I didn’t want robots to do all the art, writing, and analytics so I had more time for chores & menial tasks 😭
People under Capitalism: Oh no, our jobs are being automated. 😱😭
People under Socialism: Finally! Now that our jobs are being automated, I can chill and watch TV, maybe go on a vacation. 😎🏖🍺🎉🎊🎇🎆
(Btw, USSR/Russia and PRC are not socialist, don’t get confused)
But you’re living in capitalism. Unless government forces billionaires to fund social programs, they will just keep getting richer, just like it’s happening right now (if we ignore the crashing markets, but you get the idea)
That’s why we used to tax the morbidly rich at a 90% rate in the 50s
Oh man is translation not possible with AI. You have no idea how little languages have in common. A lot of terms don’t mean a thing, but combine concepts you don’t have or associate to point at a thing.
My dad said, about learning a new language, ‘‘cat means cat, not gato, don’t translate’’ and I think that holds up pretty well from my experience.
You can’t be serious, buddy. I’m translating an entire episode with ai and it’s turning out better than the Netflix translation!
I mean given that “AI” are language models built on context and relations between words I’d argue that that’s one of the more applicable jobs compared to what’s listed in OP. With none of them is it capable of doing well, but I just wouldn’t argue that translation is outside that realm of what’s listed above
The problem is that the AI doesn’t understand cultural context. I dunno where you’re from so pardon me for assuming you’re likely an English speaker.
A good translation isn’t just to translate what the text says but to communicate the same idea to the reader or viewer within their cultural context. A good example is Disney’s Aladdin where Robin Williams improvised A LOT during the recording sessions and most of his jokes are full of contemporary American cultural context. I’m Danish and most Danish kids didn’t understand these American jokes so our translators decided to switch out some jokes with other jokes that conveyed similar points but within a Danish cultural context.
An AI cannot do that. It will translate what is written and it will be fucking nonsense to the receiver because they don’t understand the context or the references.
AI is only good at translating as long as what is written can be translated 1:1. And even then I sometimes wonder. Because as a Dane I have noticed how terrible Word is at Danish when it comes to corrections. It follows English language context and will underline correct words in red and suggest alternative that aren’t real Danish. For example, Danish words are slammed together while in English they are separated = skolelærer - school teacher. Word could very well decide to red line skolelærer and suggest to you that you should separate the word and make it two = skole lærer. But in Danish that would nullify the meaning. Now it is no longer a school teacher but a school and a teacher.
And I have seen on streaming services like Netflix and on steam how they lazily threw descriptions into a translator and it is just the most broken Danish I have ever read. It is so fucked because the newer generations of Danes who use these services are being influenced by them to learn incorrect Danish.
I have very limited trust in AI to do a better job at it since it isn’t Danish people that have trained it and it doesn’t understand our culture, our history nor how we communicate with one another. Everything that comes out of digital text based platforms from the US is our language filtered and massacred through US context. It is very very bad in my opinion and incredibly lifeless and soulless.
It would be the same the other way around btw. Me writing a piece of text with significant Danish cultural context and humor, slang and references would be translated into total nonsense for an English speaker, I’m sure.
Oh man is translation not possible with AI.
i mean, it’s pretty good at it? A lot of human translators even struggle with the same problem, the AI is just a lot faster, and significantly more versatile. That’s arguably one of it’s strongest areas of performance, is translation, because it’s so well suited to it.
You think it’s OK because it spits out grammatically correct language on your end, but if you spoke both languages you’d get how it fails. Look at translations of Korean comics if you’d like to see how badly mechanical translation is when it’s a connected story across multiple chapters, I was reading a comic where a character said he liked the elegant and sophisticated sound of calling a lightning strike skill ‘‘bolt’’ instead of whatever he was calling it ‘‘lighting strike’’ I think. It took me a while to realize what or whoever translated it didn’t know how to look at the context of the translation and find a English word that English speakers would find at least old fashioned if not archaic and of course longer or more poetic sounding. It’s like the whole thing when JRPGs can’t figure out if they should localize names by just spelling out the phonetic sounds in Roman letters or actually translating the meaning of the name, or a thing no one’s ever done and find a name in a European language family that has the same meaning.
Just like the AI art, it’s not replacing good translation, it’s replacing hack job translations, it’s replacing mediocre and predictable art. I really don’t care if someone uses AI in the pre-production or some post production functions, just not the part you need a human for, the actual creativity, there’s an adage in 3D animation ‘‘it you let the computer do it, it’s gonna suck.’’ You can let the computer do inbetweens, but you better be giving it nothing near a key frame. It has to really be the very least important frames.
AI currently completely does not understand the context of translation when it comes to visual media. Whereas a human translator can use that for additional interpretation
It’s still doing a consistently poorer job than a skilled translator, because it has no concept of nuance or tone. I encounter people getting themselves worked up over information in AI-translated news articles, so I go back to the source material and discover it’s mistranslated, under-translated, or just completely omitted parts of sentences. It’s very Purple Monkey Dishwasher.
The quality is better than it was a decade ago, sure, but that’s a pretty low bar. Back then it was gibberish, nowadays it’s natural-sounding phrases with incorrect translations.
And yet, translators are losing their jobs left and right, from what I hear. Sure, quality has gone down, but most people don’t seem to care. Plus, in a lot of cases, instead of the AI doing all the work, translators proof-read AI generated texts and correct the worst mistakes. Fewer translators can translate more at a lower price this way.
Does the quality still go down a bit that way? Probably. But again, who cares? Not the people spending money on translations, that’s for sure.
this is true, but for the average person, who just wants to translate something to make it make somewhat sense, it’s great.
Though yeah, you can’t really trust it, there’s a lot of intricacies.
No one said that AI was doing the jobs it’s replaced people in well.
Amusingly, cook is probably the safest of those positions for the time being. The physicality and necessity of presence makes it harder to automate. Lawyer, doctor, and teacher can be done remotely, and is based largely on knowledge, so they are prime targets. People are already trying it. Drivers you could see being done remotely if we had faster, more ubiquitous, net connections, so it’s doable as well. It’s basically already happening. But cooking… AI doesn’t seem like it would give you the right kind of inputs and outputs to do that any easier/faster/cheaper. It’s already possible to make a food vending machine. The limitations of vending machines aren’t really that they need an easier interface on their database. AI won’t really help there. And to go beyond that and try to make an AI powered restaurant probably wouldn’t be profitable. It’s barely profitable to run a regular restaurant most of the time. If you try to put in the probable millions to automate a restaurant, it’d probably go the same way as the self-checkout lanes at stores, which is to say poorly.
Actually have all of the jobs I would think the safest are doctors and lawyers. When your life and liberty are on the line you really don’t want an emotionless machine you want a human.
Years ago I had to have surgery on my neck to remove a benign tumor, and I absolutely wasn’t worried, I was definitely worried it would hurt but I wasn’t worried it would go wrong and I’d end up getting a major artery cut, because I trusted the person doing it, because they came and talked to me. I wouldn’t absolutely not trust a robot to do surgery, even if logically the robot would probably be better than the human.
Watch Prometheus by Ridley Scott, there is a scene in that movie which is on topic of the subject discussed here
If we were not ruled by tech oligarchs, and the control & benefits of AI were not concentrated among a privileged few, AI replacing our jobs would be a good thing.
Everyone thinks their own line of work is safe because everyone knows the nuances of their own job. But the thing that gets you is that the easier a job gets the fewer people are needed and the more replaceable they are. You might not be able to make a robot cashier, but with the scan and go mobile app you only need an employee to wave a scanner (to check that some random items in your cart are included in the barcode on your receipt) and the time per customer to do that is fast enough that you only need one person, and since anyone can wave a scanner you don’t have much leverage to negotiate a raise.
This is the lump of labor fallacy. The error you are making is assuming that there is a fixed quantity of work that needs to be performed. When you multiply the productivity of every practitioner of a trade, they can lower their prices. This enables more people to afford those services. There’s a reason people don’t own just 2 or 3 sets of clothes anymore.
When you multiply the productivity of every practitioner of a trade, they can lower their prices.
I’m sorry, but that’s some hilarious Ayn Rand thinking. Prices didn’t go down in grocery stores that added self-checkout, they just made more profit. Companies these days are perfectly comfortable keeping the price the same (or raising them) and just cutting their overhead.
Don’t get me wrong, if there are things they could get more profit by selling more, then they likely would. But I think those items are few and far between. Everything else they just make more money with less workers.
Are you sure self checkout is actually a labor-saving device? Does it actually save costs on net, once you factor in increased theft and shrinkage? Remember, just because companies adopt something, doesn’t mean it’s actually rational to do so. Executives are prone to fads and groupthink like anyone else. And moreover, this is a bit of an inappropriate example for two reasons. First, the demand for groceries is relatively fixed. Even if the price of groceries was cut in half, you probably wouldn’t suddenly double the calories you consume. Second, self checkout is a small marginal cost to the cost of goods in grocery and retail stores. Self checkout doesn’t improve the actual production process of the goods being sold in a store.
But I’m sorry, yes, you can cherry pick a few examples. But the general rule is and always has been that increased automation leads to lower prices. This is the entire story of the Industrial Revolution. People used to own only two or three outfits, as that’s all they could afford. A “walk in closet” was an absurdity 200 years ago. The clothing industry industrialized, and the cost of clothing was driven to the floor, completely contradicting what your model predicts. The 19th century textile barons didn’t mechanize production and then simply pocket the savings.
Hell, the only reason you can afford any kind of consumer electronics is because of automation. The computer, phone, or tablet you’re using now? It would cost 100x as much without automation. This is why niche electronics like specialized lab instruments cost so much money. If you’re only building a few of something for a tiny market, you can’t invest in large scale automation to bring the cost down.
Look at how quickly and dramatically the price of LiDAR has declined. LiDAR was once the purview of specialized engineering and scientific instruments. But because of driver assistance technologies, the demand for LiDAR has exploded. This allowed LiDAR manufacturers to invest in more automated production chains. They didn’t automate and keep charging the same price, as you would assume.
For an example of this in a white collar field, consider something like architecture. How many people actually hire an architect to custom design them a home? Very few. Most people buy mass produced tract homes. Tract homes benefit from a lot of automation and economies of scale, so they’re cheaper than one-off custom-built homes designed by architects. Yet if an architect could rely on specialized AI systems to vastly lower the number of hours required to design a set of home plans, they could charge less. Many more people would then be able to afford the services of an architect.
Yes, you can cherry pick a few examples of industries that have little competition or fixed demand, where they automate without substantially lowering prices. But even those big box stores with their automated checkouts are examples of automation lowering prices. There’s a reason the giant chains can charge less for products than small mom-and-pop shops. A giant grocery chain is big enough to invest in a lot of automation and other economies of scale that a small co-op can’t afford.
In some extent this is true and correct, but when it comes to automate individual thought and creation then ethical problems arise which should be looked at and asserted carefully and with dignity, because there should be boundaries on how much automation can extent in human life, in the end humanity does not compete with anybody except itself, we are humans trying to live and most of all communicate with each other, Jobs are also a way to communicate and socialise but as we already saw they try to take that away in any way they can.
And that’s a good thing, if and only if you provide pathways to other jobs or phase workers out slowly i.e. by retirement.
Or provide UBI to share the wealth generated by increased societal productivity
I have had a number of conversations with relatively reasonable conservatives, where I’ve brought up the dangers of so many jobs moving toward automation with no additional job creation. And steering the conversation carefully, I got them to at least consider the idea of UBI funded by taxing any and all automation. I also got them (with the “everybody should have to work, people shouldn’t get life handed to them for free” mentality) to agree that the rise in automation should mean people working less hours each, so everyone still has jobs (basically, UBI and changing “full time” to 25 or 30 hours, where people get overtime past that… creating more jobs while peoples needs are still covered).
It’s amazing, sometimes, how starting with some similar premises (people should have to work, which I mostly agree with) and shared threat (automation taking jobs) can lead to some more open minds for things that they would otherwise be adamantly against.
Teachers, drivers, and lawyers are all very replaceable by AI. And, with some investment in automation, so are cooks.
drivers
or…
TRAINSSSSSS 🚂🚃🚃🚃💨
If you mean proper definition of the word AI, then of course, everyone are, AI by definition can do everything human can.
If you mean modern slop generators or narrowly trained models, then no, some professionals can use it to make their lives slightly easier, but that’s it.
Just to be clear, the proper AGI doesn’t exist, and we aren’t closer to the understanding how to achieve it than we were in the age before we discovered electricity. Possibly further, if everyone will continue to be mesmerised by a chatbotYeah interestingly I watched a video where a robotics specialist said they believed AI would take jobs long before the new generation of robots do. Robots are hard.
Can’t we skip that and go straight to replicators instead?
Star Trek or Stargate?
