Hacker News Reader

Comment by satvikpendem | Hacker News Reader

[flagged]

> I have not heard of a single person actually losing their job to AI

People are losing jobs to AI. It is in the form of layoffs so that companies can invest in AI, it is in the form of lower hiring of junior engineers because AI is supposed to do those jobs, it is in the form of silent reductions because less people are expected to do more using AI, it is in the form of entire job roles being replaced (customer success), it is in the form of second order market effects of reduced spending from people who lost their jobs, and much more.

Read my comment again. They are not losing them to AI, they are losing them to all the other factors you mentioned, most of which are economic. There is no way that an LLM for example is fully capable of entirely replacing an individual human's job at this current point in time.

Respectfully, you need to look at what I said. “People are losing jobs over AI…”

My follow up expands on it. You are arguing that LLMs are not capable of entirely replacing a human’s job therefore people are not losing jobs over AI is extremely bad faith. If a machine in a factory automates 50% of a worker’s job, it still replaces half the workers. The factory won’t employ the same number of people and the people absolutely lost their jobs to automation.

But it's not AI they're losing jobs over, that's my point, because no AI can replicate a full human employee's work. They are losing their jobs over other indirect conditions like economic ones, if they even are which I don't see any actual evidence of beyond, as I said, ZIRP ending induced layoffs and similar.

https://www.economist.com/finance-and-economics/2025/11/06/d...

I'm sure you could argue endlessly about which factors are more important to more people, but overall employment does appear to be reverting to its longer-term trends after pandemic disruptions, and these kinds of corrections have historically impacted junior positions much more than senior positions.

I know translators who did, they couldn’t make a living with it anymore.

AI cannot do translators' jobs fully, these companies hire (or hire back) people who check the translation and fix mistakes.

These review people already did exist before with human translations, and it’s also not their full-time job.

From the few translators I know, they're expected to do it all, rather than having others to review as well. I don't know how it is in big companies but I don't understand why they'd have separate review people alongside translators, seems like redundant effort.

Companies that commission translation routinely check the results, to make sure they receive the expected quality and that the domain-specific details are correct. They may check less once they gained confidence with a translator, but the same is true to some extent with LLMs.

In any case, “reviewer of LLM translations” doesn’t exist as a full-time job to anywhere near the extent that “translator” used to be.

That makes sense from the client side and I wouldn't expect that's what people mean when they say translator jobs are being automated.

The question was whether jobs are being lost due to AI, not whether translations are fully automated. If translations would support the same amount of jobs in translation review as previously for human translation themselves, companies wouldn’t save money by using AI for translation. But that’s not what’s happened: pricing has fallen dramatically for human translation, because AI makes it cheaper, and the fallen prices mean that many translators couldn’t make a living from it anymore.

Where does it show that jobs have actually been lost? That is the part I am skeptical about. The roles of my translator acquaintances were shifted to review, not lost wholesale.

In fact, for each 1 percentage point increase in MT usage, translator employment growth dropped by approximately 0.7 percentage points. Cumulatively, this effect translates into an estimated loss of about 28,000 new translator positions that might otherwise have been created over the 2010–2023 period.

https://cepr.org/voxeu/columns/lost-translation-ais-impact-t...

Talking about jobs that "might have been" created is disingenuous, the same as saying pirates "steal" copies they "might have" otherwise paid for. Even setting that aside, are there any confounding factors in that report? That would be my primary question because for many other industries, companies are using AI as the excuse, not the actual reason, to lay people off.

4000 customer support people at Salesforce replaced by AI

Which they now regret: https://maarthandam.com/2025/12/25/salesforce-regrets-firing...

(But you are correct that they did lay them off.)

> Which they now regret

for now. being early can be the same thing as being wrong.

> competently doing what the person did

AI is not there yet to replace any person's job.

> AI is not there yet to replace any person's job.

This is of course true, but it has not actually prevented any layoffs. Management generally does not understand this, does not care, or is happy to have a pretext to fire a bunch of workers to make stock go up.

Which doesn’t stop profit-oriented decision makers from trying it anyway.

Tell that to translators.

At the very least, the bosses are trying to make it happen.

I am regularly in meetings where my director or vp tells people that we should replace vendor staffing with AI and that we will get minimal engineering headcount because AI should make us more productive.

Even if the bosses are wrong, we can still see the impulse and be pissed off about it.

In those meetings, ask the director when you're going to start replacing directors with AI.

(I mean, this is terrible career advice, but it would point out the problem. Either AI is good enough to replace directors, or it isn't. If it is, then their jobs go too. If it's not, then it's probably not good enough to replace a lot of other peoples' jobs. But the problem is, the directors think that their jobs are so hard that they can't be done by AI, but everyone else's jobs are easy.)

Reply on news.ycombinator.com