Hacker News Reader

Comment by KronisLV | Hacker News Reader

That was interesting. I'm reminded of the sci-fi story called "Manna" that I read a while ago, which explored a bit of a dystopian future, where advancements in technology eventually made a lot of people unemployed: https://marshallbrain.com/manna

Actually, the Wiki page has more info on it, for those that don't have the time to read it: https://en.wikipedia.org/wiki/Manna_(novel)

The story is from 2003, but hearing about people being laid off due to AI/LLM advances suddenly reminds me of it, as well as AI gradually encroaching upon more positions. After all, letting an AI of some sort handle layoffs would almost make sense.

That's not to say that the early attempts at replacing writers or artists, or even some developers won't backfire pretty badly.

View on HN

The most unrealistic part of this is Australia being a utopia of universal basic income.

Unfortunately we've gone the way of the USA in many ways, though at least centrelink (unemployment/dole) still exists as a minimal safety net.

> The most unrealistic part of this is Australia being a utopia of universal basic income.

To be fair, a corporation had to raise a trillion dollars of capital and build that society from scratch before anyone else even seemed interested. How it progressed from that point to "the rest of Australia signed on" is entirely hand-waved away.

Just like real life, where nobody cares about programs like this and everyone refuses to help until they're already taking off. Which may or may not ever happen.

Manna(one of my favorite writings on AI), is incredibly optimistic and sugar-coated. All the "basic income" parts and Australia would have to be realistically fought tooth-and-nail with people viewing welfare recipients as "lazy" and "entitled"..

People can only view welfare recipients as lazy and entitled from the pedestal of a decent job. 80% of the white collar middle class is going to have to learn to become AI "composers" or learn trades, and I'm guessing trades will be vastly more accessible for the majority of the displaced. The mass influx of tradespeople is going to depress earnings due to competition (which already weren't great unless you were a master tradesman with employees), so we're going to end up with 75% of the population below average income and people struggling to afford rent and food. When that happens I don't expect the pushback against some form of social service or basic income to be so great.

A choice of two dystopia - one where your brain is controlled by a computer, the other where you are locked in a room.

That is like traditions chicken farming vs free range chicken farming.

Keep in mind while free range chicken farming is better to market, it is also more messy.

https://thehensloft.com/free-range-chicken-pros-cons/

Have you ever been inside of an industrial chicken house? Its absolutely, without a doubt, the most disgusting and depressing place I've ever personally seen. Even worse if you're unlucky enough to be there after a single case of avian flu is found anywhere in the area.

Chickens can be messy, but that's only an issue with free ranged birds if you don't have enough space for them. Most chickens won't venture too dar from their coop, usually on the order of a couple hundred feet. If you don't want chickens pooping all over your lawn and barn, just put their coop further out in the field (preferably near some trees for cover).

Rotational grazing is also really effective with chickens. Its more work, but you keep the birds where you want them and end up fertilizing your fields as you move them around.

It's also a joy at feeding time to see them running home by the dozens.

I've been "enjoying" watching that story come true since he published it. It's a great summation of all the trends that have been going on for a long time now.

Do you have any sources for layoffs "due to AI/LLM advances" other than layoffs-in-one-sector-to-replace-with-AI-experts which I can find plenty of, but which are not really due to efficiency improvements?

What’s the difference? AI-motivated job cuts are widespread in all sectors, and companies are phrasing them as “efficiency improvements.” This is only possible because of recent advances in LLM capability. If anything, skilled experts are slightly less likely to be replaced by AI (with notable exceptions for visual artists)

“IBM Plans To Replace Nearly 8,000 Jobs With AI” https://finance.yahoo.com/news/ibm-plans-replace-nearly-8-17...

“Around four in 10 [business leaders surveyed by ResumeBuilder] said they'll conduct layoffs as they replace workers with AI, with Dropbox, Google, and IBM have already announced job cuts for that very reason“ https://www.businessinsider.com/layoffs-sweeping-us-these-ar...

“A recent Goldman Sachs study found that generative AI tools could, in fact, impact 300 million full-time jobs worldwide, which could lead to a "significant disruption" in the job market” https://www.businessinsider.com/chatgpt-jobs-at-risk-replace...

Thanks, but I am really curious about the jobs AI itself has replaced today.

Note the use of "will", "could" and "plans" in those articles: none of this has materialized today, as was implied in the GP post.

> What's the difference?

The difference is in chasing the market potential vs actually achieving that potential today: it seems to me generative AI is not replacing any jobs today, but AI experts are being hired in lieu of traditional software experts en masse, hoping for generative AI advances to continue to the point where those efficiencies materialize (which is never a given, but I do see the business reasons to bet on it).

One statement is that generative AI is good enough to replace workers today (which I haven't seen proof of yet), and another is that it seems this might happen in the (near) future so let's invest in developing AI more (also implying that where it is today is not sufficient).

Reality is that knowing LLM and other AI tech now is going to be a long term distinct advantage for software engineers of all stripes and those that don’t know will fall behind

To me that does not sound like a "reality", but a hypothesis that many are willing to bet on today.

It may turn out to be true, but I don't see any proof of that yet, not even in the social sciences-style of proof (which is admittedly a lower bar to clear than mathematical-style proofs, but that would do for now).

I share your skepticism about AI tools being effective. However, when we look at it from a labor relations perspective, what's important is the threat of being replaced, and that's happening now. I think the threat itself is (currently) far more damaging to workers' psychological safety (thus team performance) than whatever outcomes using AI (currently) actually brings.

My point is that we're already beyond the "fuck around" threshold; your point (if I understand you correctly) is that we haven't reached the "find out" threshold yet. :)

SO where is the near-real-time country-wide system that ratchets down the standard (statutory) work week in sync with the overall decline in aggregate work ?

First there would need to be signs of greatly increased productivity where it counts (customer facing). Somehow I have my doubts we will see that with the current crop of LLM.

Governments need us busy and distracted, they'll never willingly give us free time like that.

This was an extremely good read, thank you!

Reply on news.ycombinator.com