Skilled Older Workers Turn To AI Training To Stay Afloat (theguardian.com) 39
An anonymous reader quotes a report from the Guardian: [Five skilled workers aged 50 and older spoke] to the Guardian about how, after struggling to find work in their fields, they have turned to an emerging and growing category of work: using their expertise to train artificial intelligence models. Known as data annotation, the work involves labeling and evaluating the information used to train AI models like Open AI's ChatGPT or Google's Gemini. A doctor, for example, might review how an AI model answers medical questions to flag incorrect or unsafe responses and suggest better ones, helping the system learn how to generate more accurate and reliable responses. The ultimate goal of training is to level up AI models until they're capable of doing a job as well as a human could -- meaning they could someday replace some of these human workers.
The companies behind AI training, such as Mercor, GlobalLogic, TEKsystems, micro1 and Alignerr, operate large contractor networks staffed by people like Ciriello. Their clients include tech giants like OpenAI, Google and Meta, academic researchers and industries including healthcare and finance. For experienced professionals, AI training contracts can be a side hustle -- or a temporary fallback following a layoff -- where top experts can, in some cases, earn over $180 an hour. But that's on the high end. For some older workers [...], it represents another thing entirely: a last refuge in a brutal job market that is harder to stay in, or re-enter, the older they get. For many of them, whether or not they're training their AI replacements in their professions is besides the point. They need the work now.
[...] "There's just a lot of desperation out there," Johnson said. As opportunities narrow, many turn to what Joanna Lahey, a professor at Texas A&M University who studies age discrimination and labor outcomes, calls "bridge jobs" -- lower-paying, less demanding roles that help workers stay financially afloat as they approach retirement. Historically, that meant taking temp assignments, retail and fast-food work and gig roles like Uber and food delivery. Now, for skilled workers -- engineers, lawyers, nurses or designers, for example -- using their expertise for AI data training is becoming the new bridge job. "[AI] training work may be better in some ways than those earlier alternatives," Lahey told the Guardian.
AI training can offer flexibility, quick income and intellectual engagement. But it's often a clear step down. Professionals in fields such as software development, medicine or finance typically earn six-figure salaries that come with benefits and paid leave, according to the US Bureau of Labor Statistics. According to online job postings, AI training gigs start at $20 an hour, with pay increasing to between $30 and $40 an hour. In some cases, AI trainers with coveted subject matter expertise can earn over $100 an hour. AI training is contract-based, though, meaning the pay and hours are unstable, and it often doesn't come with benefits.
The companies behind AI training, such as Mercor, GlobalLogic, TEKsystems, micro1 and Alignerr, operate large contractor networks staffed by people like Ciriello. Their clients include tech giants like OpenAI, Google and Meta, academic researchers and industries including healthcare and finance. For experienced professionals, AI training contracts can be a side hustle -- or a temporary fallback following a layoff -- where top experts can, in some cases, earn over $180 an hour. But that's on the high end. For some older workers [...], it represents another thing entirely: a last refuge in a brutal job market that is harder to stay in, or re-enter, the older they get. For many of them, whether or not they're training their AI replacements in their professions is besides the point. They need the work now.
[...] "There's just a lot of desperation out there," Johnson said. As opportunities narrow, many turn to what Joanna Lahey, a professor at Texas A&M University who studies age discrimination and labor outcomes, calls "bridge jobs" -- lower-paying, less demanding roles that help workers stay financially afloat as they approach retirement. Historically, that meant taking temp assignments, retail and fast-food work and gig roles like Uber and food delivery. Now, for skilled workers -- engineers, lawyers, nurses or designers, for example -- using their expertise for AI data training is becoming the new bridge job. "[AI] training work may be better in some ways than those earlier alternatives," Lahey told the Guardian.
AI training can offer flexibility, quick income and intellectual engagement. But it's often a clear step down. Professionals in fields such as software development, medicine or finance typically earn six-figure salaries that come with benefits and paid leave, according to the US Bureau of Labor Statistics. According to online job postings, AI training gigs start at $20 an hour, with pay increasing to between $30 and $40 an hour. In some cases, AI trainers with coveted subject matter expertise can earn over $100 an hour. AI training is contract-based, though, meaning the pay and hours are unstable, and it often doesn't come with benefits.
Re: (Score:3, Insightful)
Re: (Score:2)
Who the fuck is "they"? "NGOs"? You mean the Red Fucking Cross?
The only reason there's not a 100% chance you are a racist is that you might be a bot.
Re: (Score:1)
The red cross is dirty
https://www.reddit.com/r/OutOf... [reddit.com]
Re: (Score:2)
Indeed. Immigrants are, by and large, wonderful, hard-working people. They spawn new businesses at a higher rate than native-born Americans. We should be welcoming them, not confining them to steel cages.
Re: (Score:2)
Re: (Score:2)
None of the H1B issues you describe, reflect on the character of immigrants, as alleged by pete6677.
There is a charming illusion among many Americans, that if we just control the number of H1B visas and eliminate "fraud" from the program, it will keep American salaries propped up. This is entirely false. Companies that are willing to jump through the many H1B hoops, are just as willing to outsource their work to people who work remotely from other countries. "Tightening up" H1B rules, isn't going to prevent
Just wonderful (Score:2)
I'm personally not inclined to trust anything (human or otherwise) trained by a person with an axe to grind or resentful of their replacement. Someone will have trained the replacement poorly out of spite. I suspect most people have heard stories of this happening when IT started getting offshored or otherwise outsourced in decades past.
Evil (Score:3, Insightful)
Sabotaging your replacement is just evil. If you can be replaced that easily and that cheaply, the problem is not the replacement, it is that your role was never as secure or valuable as you thought.
What I have seen with offshoring is that clients often come crawling back, because in the end they get exactly what they paid for.
In medicine, deliberately entering malicious responses into a medical system would leave a beautifully documented record of your own misconduct. It would violate basic ethics, likely
There's something you don't understand (Score:4, Interesting)
For literally everything else good enough is always good enough.
Now if companies still competed against each other good enough wouldn't be good enough. Higher quality talent would produce a higher quality product overall and customers would flock to the better company.
We stopped and forcing antitrust law over 40 years ago. You can find the charts showing you the seven companies own basically everything. And those seven companies are owned by a handful of people when you look at who actually owns the stock.
So that gets you enshitification. I don't need to make a good product. If somebody comes along making a better product because they overall have better people I buy that company or I run it out of business. This is what Facebook does. You can look up the company's Facebook has bought and it's a who's who of up-and-coming social media competitors. Amazon did the exact same thing and it's how they became the number one retailer. It wasn't impressive tech it was buying their competitors and being allowed to do it by toothless regulators.
We broke one of the fundamental aspects of capitalism and they are downstream effects. In this case there's drastically less employment in our economy as a result and your wages and mine are substantially lower as a result.
It's a chesterton's fence. Don't pull the fence down if you don't know why it was put up
There is no Enshitification is a lie (Score:2)
You, rsilvergun, and others like you are obviously infected with Declinism, projecting your own decaying reflection you see in the mirror out into the world.
There is no enshitification, it is a combination of rosy retrospection, and a low barrier of entry.
For instance the quality of videos on Youtube has not declined, in fact there are way more high quality videos than when it debuted in 2007.
It takes a lot less effort today to post something on Youtube then to get in a movie theater or syndicated on televi
dystopia (Score:2)
there is no one accountable (or can be held accountable) for the bleak future. I for one welcome our EMP overlords.
Re: (Score:1)
Think of it this way, if you're right, then some of the many out of work engineers would build/hack a swarm of attack bots to target the compounds of the ruling elite.
They won't be able to get the resources (Score:2)
I don't think folks have realized just the scale of poverty that the Epstein class has in store for us. G
Re: (Score:1)
LOL you built an LLM rsilvergun post generator. Nicely done!
Re: (Score:1)
You must be forgetting that a couple of decades ago during the Iraq war a handful of guys with IEDs were able to do a shit ton of damage to the world's most powerful military.
Re: (Score:2)
Can you help me understand the bleakness you speak of?
Like every new technology, and especially every new automation technology, AI has a dark side. But automation also frees people from mind-numbing jobs that nobody wants.
These days, who actually wants to work on farms? Yes, a few, but less than 3% (down from 70%). Or what about factory jobs? Who wants those? Certainly not me! And when it comes to programming, I *love* using GitHub Copilot. I can get things done much faster than before, with a lot less typ
Mixed feelings, actually.... (Score:2)
On one hand? I think there's considerable evidence this AI bubble is going to pop; maybe in 1-2 years from now? If that's the case, the tech workers who manage to get paid training AI models still walk away with that money when it gets shuttered due to lack of funds.
On the other? I also get how distasteful it is to "train your replacement", especially when the replacement is just computer software.
I think much of this depends on how things *really* pan out. I'm not seeing big I.T. job losses due to AI impl
Re: (Score:2)
Predicting a bubble in 1-2 years ... that's not how bubbles work. If it is a bubble, nobody knows what's in 2 years. Bubbles are a short-term effect.
Re: (Score:2, Interesting)
"I think there's considerable evidence this AI bubble is going to pop" Citations please? I've read lots of speculation, but where's this evidence you claim?
I think many of us may wish there's a bubble that may pop, but 'if wishes were fishes..."
There IS evidence that job losses are occurring, it's not just hypothetical, just because you say "I'm not seeing..." perhaps in your immediate world, evidence is to the contrary (just a few I found)
https://www.goldmansachs.com/i... [goldmansachs.com]
https://econofact.org/factbrie... [econofact.org]
ht [adpresearch.com]
Re: (Score:2)
No, there is no AI bubble. A bubble happens when valuations (stock prices) wildly exceed revenue - that happened during the dot.com bubble - companies with *no revenue whatsoever* were valued to the moon. And no one seemed to notice or care.
Todays situation with AI is not that. Revenue is massive, is growing wildly, and stock valuations are really not overextended - given the reasonable estimates of future growth for the companies. There is a huge amount of money being used to build out more AI, but tha
How to vote for your interests (Score:2, Informative)
2) Get rid of the 2 party system. Support any and all measures to do ranked choice voting which gets rid of 'vote so the other guy doesn't win'.
3) Nov 2028 vote in completely in your own interest without having to 'vote so the other guy doesn't win'
From middle class to this (Score:2)
Nothing says 'survival' wages or "you're disposable", like the absence of health-care. Once again, it's a US employer's market and they are ensuring the employees cut their own throat. With trained AI machines for their workforce, they can stop paying wages and start buying more politicians.
I'm 59. I'm doing a different kind of AI training (Score:2)
In the past 6 months, I learned how to build web apps using Vue.js...using AI. I started out by telling GitHub Copilot what I wanted, and watching what it did. As I went through this process, I learned Vue.js myself. Now, I'm teaching others.
Also in the last 6 months, I learned how to build search indexes using Lucene (the engine behind ElasticSearch). Again, I did it with AI. Now that I've take the code through multiple revisions, I actually understand how Lucene works.
Just today, I built my first MCP serv
Re: (Score:2)
Re: (Score:2)
I fully agree!
I know what I want the software to do, I just don't enjoy figuring out the exact spelling of the API or method calls I need to make to get it done. When I read the code AI produces, I instantly know whether it's right, and I'm just glad I didn't have to type it all!
Re: (Score:2)
Re: (Score:2)
Not this one. (Score:1)
Combinatorial explosion and irreducibility await.
P != NP