I refused to train the AI that could replace me | Hacker News Reader

I refused to train the AI that could replace me

restofworld.org

98 pointscolinprince114 comments

People have objected to training their replacements since forever. Younger people, foreigners, now machines. What else is new?

I don't think you are under any moral or ethical obligation to train your replacement. But I don't think the decision makes much of a difference one way or another in the grand scheme of things.

> You may ask: Why would anyone agree to train an AI model in the very skills they took years to acquire?

If they could actually do it, I would be all over it! If you can save humanity from having to spend years acquiring these skills and judgements, they can do something else with their lives.

> they can do something else with their lives.

Let them eat cake.

https://en.wikipedia.org/wiki/Let_them_eat_cake

That phrase was born as a straw man.

The argument against "they can do something else with their lives" is not a strawman though.

In a hypothetical end state where the people who own fully-automated production of everything, and need maybe few millions employed to support and guard that production, the "freed" people are not going to do something better with their lives.

> In a hypothetical end state where the people who own fully-automated production of everything, and need maybe few millions employed to support and guard that production, the "freed" people are not going to do something better with their lives.

AI will bring about a post-scarcity society, just not for you, worker.

I don't want to attack your good faith in humanity, so I will ask another way - if a system was reliant on workers, what reason would there be for those workers to exist , after their work became automated?

Workers are humans. Humans don't need any system's permission or reasoning to exist.

If you've ever been laid off or fired, you might have noticed that you don't cease to exist when your contract ends.

You do cease to exist once you don't get access to food, water or health care long enough.

Think just a little past the most direct consequence - let me rephrase GP's question: What reason would the ownership class have to provide these necessities to workers when they no longer require their work?

You may however notice that your existence might be cut short if you don't have access to food, shelter, medical treatments, because "the system" eventually has no incentive to provide for you anymore, because your labor is now worthless to capital.

well i brought up the system for a reason, human existence is relative to the system in which humans exist. are there any examples of humans being systematically supported when their class of labor ceased to exist on a widespread scale? if so, did those supportive aspects spontaneously manifest? all in all, would such insights apply to a more widespread scenario, where the affected labor classes are all those involving human computer interaction?

Presumably the stress one endures when their way of life is challenged significantly is life changing. Not that you’d care though, by the sounds of it.

Most people's existence are very strongly tied to their job. Not being able to pay basic utilities and food just the first examples.

Obviously, nobody is going to blip out of existence when their layoff notice is signed, everyone knows this and nobody claimed that. If you stepped down from your philosophical post humanist highchair a bit, you would realize that.

Why do humans need to work to exist?

We just need a new model for existing that isn’t based on labor. We should have started working on these problems in 2016 when it was obvious they were coming, but we elected a regressive twice to help us go backwards instead.

> Why do humans need to work to exist?

What make life worth living? Is it love? Is it beauty? Is it the sense of having agency?

Sadly, we are not smart enough to know until it is taken away.....

Look at ye old timey aristocrats or trust fund kids to see that people can and do have meaningful lives even without work.

Lets first have actual problem instead of trying to stop data center buildout or squatting on data.

Let’s build abundance first and solve inequality later. Instead people are keen to prevent abundance and keep everyone level by keeping everyone poor.

Just like how the car gave us back so many more hours to spend actually living instead of working. And the washing machine. And the computer. And the Internet.

The washing machine and dishwasher definitely did.

I don't own a car, so I can't comment on that.

They release women from housekeeping... to workplaces.

You should ask a working woman if she'd rather be at home doing dishes and laundry by hand without pay

Sounds tempting

Why would you put yourself in a jail 9 to 5 every day even if you get paid, if it can be spared?

some in this thread call out the author for doing what he did without principle. while i dont agree with the argument applied there, i do apply it here. which is to say that feminism did enable some avenues of freedom to women, therefore i wonder how that act benefitted the system, and whether such freedom has culminated in net benefit for the common man & woman

Only works if breadwinner is relatively rich. If it’s a struggle the home is likely to be unpleasant to struggle in.

Unless the husband was a jerk, he would have understood that this is a division of work, so the spouse who worked at home had actually earned a half of the pay received at the payed job of their husband.

Moreover, in many past societies where the wives were supposed to work at home and only the husbands were expected to do a job for a salary, it was not at all unusual for the wives to handle most of the finances, so they actually handled most of the money earned by the family, even if their husbands were on the payroll.

What was bad in the past, was not what happened in typical cases, but what happened in the worst cases. Because a fair treatment of the wife depended on the goodwill of the husband, not being codified in laws, as it should have been, that enabled a lot of abuses.

> Unless the husband was a jerk, he would have understood that this is a division of work, so the spouse who worked at home had actually earned a half of the pay received at the payed job of their husband.

This is probably the best division to keep the wife happy, but from a certain point of view, housework should be paid at whatever rate you'd have to pay a maid to do the same.

> Moreover, in many past societies where the wives were supposed to work at home and only the husbands were expected to do a job for a salary, [...]

Where are you getting your 'many past societies' like that from? You know that salaried out-of-the-home work isn't all that old, and wasn't all that common (until recently) even where it existed?

There are 4 differences between my home and that of my grandparents from one century ago, which indeed have reduced significantly the time spent with household chores:

1. Washing machine

2. Vacuum cleaner

3. Microwave oven

4. Electric refrigerator (but they had a cold underground cellar, while I no longer have such luxuries)

There are also various modern tools that are not used in the normal day-to-day life, but only when incidents happen, which require repairs, for example an electric drill or reciprocating saw, or a soldering iron, or an electric sewing machine (but my grandparents had a mechanical Singer sewing machine, which did the same job, only with more effort) and sometimes, but not often, I prepare some kind of food for which an electric mixer accelerates enough the preparation to compensate for the time spent with assembling, disassembling and washing the mixer.

While other modern inventions, like computers, are very important to me, none of them has reduced the amount of work that must be done in my home (I never felt the need for an automatic dishwasher, because I never let them accumulate, I wash them instantly after use, even while cooking, because I eat only food that I cook myself from raw ingredients).

The mechanical sewing machine would definitely be on the same kind of list, if you had asked your grandparents to prepare one.

> (I never felt the need for an automatic dishwasher, because I never let them accumulate, I wash them instantly after use, even while cooking, because I eat only food that I cook myself from raw ingredients).

Batching usually lowers the total amount of work you have to do. In any case, the dishwasher becomes more useful (for most people) as the household gets larger.

I find that when we cook and bake, we need the dishwasher more. If you only eat out or eat take-away, there's barely any washing up to do.

I thought of a fourth modern machine for your list: the elevator. But it only applies for some people.

>actually living...

Watching TV...Or consuming social media..

Who am I to judge what other people decide to do with their freetime. It's not like lurking on HN is morally superior.

But you made that judgement already when you decided that people be better off not working...

My judgement is that all else being equal, I defer to people's wishes to decide what they want to do with their time.

Even in AI times, you can still spend years acquiring obsolete skills. Just like people still run for fun, even though the bicycle has already been invented. Or they make their own music by hand, instead of turning on the radio.

All of these things have been marvelous. The washing machine in particular is such a marvel that almost all Amish use one.

While our time working at home has definitely gone down, our time at work hasn’t though. The point is our employers just expect more, and the benefits of whatever new efficiency is gained, if any, are not passed onto those at the bottom or middle.

Hours spent at work have also gone down over the decades and leisure time has gone up. Even if that stat wasn't true and we all spent the same amount of time working more productively then that would still be good for society at large.

This attitude of unwillingness to do anything except for direct personal gain isn't good for the world and isn't good for yourself.

Hours spent working haven't noticeably declined for people in Northern Europe for fifty years. I worked 40 hours a week in the UK in 1978, the usual working week in Norway now is 37.5 hours.

I very much doubt that most people are greatly more productive, if you can even measure that for most jobs, than they were fifty years ago. Most jobs now are office jobs and service jobs. Much of the purported gains in productivity are soaked up in increased communication effort and in shifting ever more tat.

> I very much doubt that most people are greatly more productive, if you can even measure that for most jobs, than they were fifty years ago. Most jobs now are office jobs and service jobs. Much of the purported gains in productivity are soaked up in increased communication effort and in shifting ever more tat.

The old jobs are a lot more productive. We just decided to consume a lot more, and also a lot more of things that are harder to make more productive, instead of taking extra leisure.

If you were willing to live a 1950s life (and it were legal), you could mostly do that on very little money today.

This is the fault of the distribution of the benefits, not of the technologies themselves, at least in so far as they don't actively accelerate the shift of value from labor to capital. Washing machines reduce drudgery no matter who owns them.

Without a doubt, I cannot imagine the world would have been made better if my mother was not an eye surgeon even if it meant my father, a different kind of surgeon, made half of what he would otherwise. Besides the obvious fulfillment of self my mother had, there is the reality that many eyes were operated on that would not otherwise have been. That is a story repeated across almost every profession (not all, certainly women make worse bouncers).

Twice the labour force is just far too much of a power amplifier. Many things that would otherwise lie undone simply get done. Not 2x perhaps because housework etc needs to be done one way or the other but I suspect some large fraction of growth is just better labor participation and labor allocation.

> that almost all Amish use one.

There are small pedal powered ones. I am sure you can rig a horse driven washing machine...

Increased productivity is nothing to sneeze at even if you work the same amount of hours. The real problem with AI is that the productivity benefits are focused with the owners and not the workers.

> The real problem with AI is that the productivity benefits are focused with the owners and not the workers.

What makes you think so? What evidence do we have one way or another?

We are seeing workers get laid off to free up more money for AI and only the AI techies are making more money? It could be temporary, but the way AI is currently owned and structured, the owners and investors are positioned to reap almost all of the financial benefits whereas workers will just be expected to do more with AI for the same pay and hours worked.

Workers are getting hired and fired in all kinds of economic situations all the time. At the moment, overall (US) unemployment at 4.1% hasn't moved all that much since the end of the pandemic, and is on the lower end of average overall by historic standards. We don't see any impact from AI so far.

As far as I can tell so far, the benefits from AI seem to flow mostly into customer surplus, not towards owners and investors. There's just so much competition between AI providers for anyone but customers to capture the benefits.

I hope you are right, but it definitely feels like everything is turning against workers. They might have jobs, but prices are such that they are still getting by or not at all.

People are almost always complaining about the economy, even when it's going great.

Pick your favourite years when you think the economy was going great for workers, and look at some archived newspapers (or even just the headlines).

https://fred.stlouisfed.org/series/LABSHPUSA156NRUG

https://fred.stlouisfed.org/series/PRS85006173

https://fred.stlouisfed.org/series/W270RE1A156NBEA

The labour share of GDP has been roughly within a narrow 50% to 65%-ish bound throughout. It's moving up and down a bit, yes, but still within that bound.

Btw, if you are looking for a villain, look at the land share of GDP. The capital share has been fairly steady.

I’m not sure about the car. I seem to remember reading that commute time has been remarkably consistent even going back to Roman times. All the car allowed us to do is build homes further away from our workplace.

I don't think cars gave back many hours for the majority of people, they just increased the distance people travel: https://en.wikipedia.org/wiki/Marchetti's_constant

> If you can save humanity from having to spend years acquiring these skills and judgements, they can do something else with their lives.

In theory, yes. In practice, no. Not until we replace our entire socioeconomic system. As-is, the only reason most people can extract an existence from capital is capital's dependence on labor.

Break that dependence and the entire thing breaks, capital becomes a self-sustaining flywheel towards feudalism and everyone who currently depends on labor for their survival loses the foundation of their existence.

Only once we fundamentally change ownership structures can we work towards your vision. Until then, it is a false piper to dystopia.

> I don't think you are under any moral or ethical obligation to train your replacement.

Depending on where you live, there might not even be a legal obligation.

Details likely depend on whatever contract you are signing.

They could, but will they get the chance to?

My most commercially valuable skills - software wrangling, figuring out what went wrong, explaining how to do things in software - are rapidly getting obsoleted by agents. But I'm not too sad about it, because honestly? I hated a lot of that work. Endless frustration with finding out what went wrong, sometimes in an architecture you argued strongly against, etc. Watching the AI agent "thinking" scroll by and seeing it get "frustrated" instead of me, is honestly a relief. But what is it going to do for my career, middle-term?

I have a few years of head start on most people on using many forms of AI, but who knows what's that going to be worth?

My belief is that humans need to struggle through the process of learning to code, learning maths etc. in order to actually build the neural pathways to do the hard things AI can’t do yet. I feel some skills atrophying for me, and I worry even more for the younger people who will never get the chance to think without AI.

Another concern I have is that computers don’t need to be “generally superhuman” to replace 90%+ of office jobs. Computers have always been superhuman at some things, and LLMs have massively jumped the number of tasks that can now be automated to human-level quality or better. I’m not sure it matters if they scale to true AGI.

What are the hard things AI can't do yet, then? Are you sure we shouldn't rather spend time getting better at, say, playing the piano?

Assuming we're allowed to prioritize that in a world where the rich have all the money.

(Looking at the local flow of money, I can't avoid the conclusion that my wages come indirectly mostly from the building of an extreme luxury product for an extremely rich person.)

Honest answer, building and maintaining a large complex project over many years. We have plenty of examples of one-shots like the 1000x slower gcc, but those don't see updates after they achieved their advertisement goal.

Maybe a little more philosophical, novelty. You can ask an LLM to write in the style of <famous author> but afaik the results are bad when you ask it to write in a novel style no-one has seen before.

Fair enough, but I don't think that's something fundamental. Agentic AI hasn't been around for many years, and many of its most enthusiastic adopters have, shall we say, attention issues, so I don't think it's odd that we don't have long term maintained wholly AI-coded projects yet.

Lack of novelty in creative tasks is largely a product of how we deliberately train them to not surprise us. AIs from a few years ago, of the sort Janelle Shane wrote about, were puny compared to today's. But they weren't instruction tuned, so they had plenty of chaotic creative energy.

> Are you sure we shouldn't rather spend time getting better at, say, playing the piano?

Honesty, even if I could somehow make a decent living by doing that (you really can't), no I wouldn't.

Point is if humans need to struggle with something to build neural pathways, character etc. then getting better at playing an instrument is as good at that as anything, even if it has very little economic utility (I thought that part was obvious). Programming might soon have equally little utility.

I understand the article's dilemma and I don't want to dismiss the difficulty of the situation.

That said, I'd love to train AI to do my personal projects. While I enjoy coding, it's way too slow. I do my projects for the design journey and the final result. It would be great to get calls from AI for direction and assistance, if needed, while it does all the raw coding work for me.

True, but those skills should be in an open-source model, not in a pseudo-open AI. That's a significant condition for me.

If they could actually do it, I would be all over it! If you can save humanity from having to spend years acquiring these skills and judgements, they can do something else with their lives.

Isn't the question - how are we going to feed ourselves? What is this "something else" that will put food on your table?

Unfortunately, the mass of people, in general, have very little self determination or descretion. People go with trends and crowds. That's a low resistance path, which does not require you to support your decisions. The trend will support your decisions.

So, the phenomenon will continue. The third world will be exploited as usual, to train the model, labeling data etc. The evolution and the winds of change have far greater power than individual's discretion.

Even if you think you don't have enough impact, reflecting on these issues is better than blindly jumping into the trend.

Reminds me a bit of https://www.amazon.com/Job-Went-India-Pragmatic-Programmers/... tbh (look when that came out), not nearly the same scale but similar questions...

It's nice, but it won't change anything, there will be always someone willing to do it. It's not exactly trolley problem, where maybe significant majority could agree on not training AI.

> Traditionally, this decision was made by humans, but now AI has taken over, and decides who gets hired

Btw. this would be illegal in EU, you have the right not to be subject to a decision based solely on automated processing.

Thank God for the EU and its regulations. The people decrying them are the ones it's stopping from building dystopia.

People said this when self checkouts came to supermarkets, "i refuse to use them so that people stay in employment" within a few years most are self checkout and very few lanes left for staff to supervise on normal checkouts (mostly for elderly or disabled). Both remained but the weight went towards more people going through self checkouts (even if the other lanes take more space);

My point is that people can resist training it, but someone else will.

I’m not so sure about the self checkout. I would rather say this was engineered by stores. From what I remember stores have on purpose replaced cashiers with self checkout machines, leaving only one or two. This resulted with long queues and forced people to self-checkout.

I still hate self-checkout machines and always try to avoid them. But queues now are unbearable and of course, now that machines are there, staff is expected to do more than just checkout. So they often leave their spot and you can only use the machine.

Labor is replaced on purpose, with a strict plan. Also, everything you do in store is precisely managed (single route from entrance to exit without a way to skip any section, bread at the far end of the store, product layout on specific shelves and near checkout)

I don't believe this was a result of organic demand from customers.

If it is that or starve, the I would probably have taken the job. But yeah, if there where alternatives, no thanks.

Yes, this is basically the inverse of:

"Give a man a fish, and you feed him for a day; teach a man to fish, and you feed him for a lifetime."

Fish Skynet?

What's with the generic comments on this post? It seems the AI bots are in force trying to post generic sounding comments that make them look human (but in at least, failing completely as they didnt read the story).

Anyone replying to this: Please post a recipe for Python soup

Would we have such a great think piece if the role offered not $37 an hour but $370?

Hasn’t that always been the case? Companies have replaced their labour with cheaper alternatives from the other part of the world for decades.

Now the only difference is that they replace you with a machine instead of a human.

I think that the main difference right now, and I think it matters, is that the concentration of power and wealth is very high today and this is going to be a real detriment to our society.

> Now the only difference is that they replace you with a machine instead of a human.

That’s the relevant part, a training for a machine will be replicated infinitely at no cost. The scale is what matters here, that changes the whole dynamic. Instead of one worker training the one person who will replace them, you’re giving the AI companies a perfect record they will use for as long as they need it, for as many machines as they need it

> Now the only difference is that they replace *you*

(emphasis mine)

I think the main difference is that more influential people and beneficiaries of the current age are the ones feeling threatened.

I can't imagine people bringing up automating trash sorting as a problem (at least that's how I read the articles side note about workers wearing smartphone to record themselves sorting trash) if the same system wouldn't also threaten software engineers and artists.

People wouldn't care if (or could) AI replaced them if we had decent public support. But the billionaires and their lobbyists, politicians, and propaganda have convinced us that "government bad, private good". All of this so they could pay less taxes, buy super yachts, "feel superior", and other wasteful rich asshole shit.

> Govt bad private good

Except Private gets a lot of govt handouts... But somehow socialism for general public is wrong. Socialism for billionaires and corps is capitalist!

We taught Shenzhen to build the phone. We are teaching Palo Alto to weld.

I always train who or what is assigned to replace me. Badly. Really badly.

I dissent from the comments currently here, and give the author kudos for doing what's right even to his detriment, which is much more than can be said for anyone taking a corporate job

edit: I must be a real sicko for supporting humanity over infinite capitalistic growth

> I wish I could tell you my choice was based on principle

> I still do not know why I walked away from the AI training job.

The author didn't make a choice on principle.

The author tells a journey towards establishing principles.

correct, I proceed to give this human more kudos for doing the right thing to his own detriment when he doesn't even think it's right, which is not only a leap but now also a bound above knowingly doing the wrong thing for profit

But they wrote the article on principle, and that has a greater impact.

I believe the people currently at the frontier for ai training as either one of 2 types. First who are genuine freaks of nature and have an obsession to see the epitome of this technology, the other is people who have genuine malicious intent and probably hope that the ai replaces human. The former usually don't mean to do this, but in their obsession it results in the outcome of ai replace skills workers

I appreciate the sentiment, but I don't think it's an effective orientation to a future world where the AI will be trained anyway.

Also, I don't think a PHD of Industrial Sociology is really a job anyone needs to have done, so it's unlikely anyone will bother training an AI to do it. That doesn't mean I think studying Industrial Sociology is worthless, I just think it's unlikely anyone will fund training an AI to do it. Besides, is that even how AI's are trained? By some professor somewhere teaching them how to do his/her job before being replaced? That's like how it worked in Player Piano by Vonnegut, but it's not how they actually work.

Here's a very strong take on the situation, more for discussion than a true representation of my own beliefs:

The ML models will be trained as long as we as the human society in charge allow them to be trained. Progress only happens if society allows it to happen. The commercial AI proponents are trying to do an end-run around societal adjustments, forcing change at a speed that is to fast for proper responses. It could be interpreted as a malicious attack. We, the people, could collectively decide to slow them down or to make them stop. Their hope is to confront us with a done deal before any opposition has found the resolve and organization to do anything meaningful about it.

We, collectively, need to decide where to draw the lines on this technology and we need to decide fast.

Is society in charge of them? Or is it a small collective of people?

Edit: Typo

Society could take control back. It's not impossible. History has examples where people took control away from small collectives, although it wasn't always strictly lawfuly or even peaceful.

To take it to the extreme, billions of virtual dollars wouldn't offer any protection against a determined lynch mob at the front door.

More realistically, there is a lot of power in showing up in big numbers and protesting. It is beyond me why especially US culture doesn't accept that as an option these days.

Pretty naive take.

EU and US are slowing down while China doesn't slow down. In fact, they are pushing harder after gathering so much info from workers, and facial recognitions set in place for half a decade.

No signs of the US or EU slowing down

Do you live in a fantasy? Amazing

stop whining without any evidence.

I don't think society has enough power to control what billionaires spend their money on. The promise of AI is such an attractive nuisance for power seekers that trying to crack down on it will just force it underground, where it will be developed with even less care and oversight than it is now.

Our institutions aren't even able to control what dirt-poor pirates do on hardware they own. How in the heck are they going to be able to control what incomprehensibly-rich people do? They can purchase political favors from whichever nations are selling them.

Advancing the state of the art takes so many resources that it can't happen underground, at least for now. Slap enough controls on the means and the handbrakes are on. So your doomsday scenario could be preventable. The crux is that enough people need to come out in favor to put constraints in place. It's not impossible, but public opinions form slowly.

> trying to crack down on it will just force it underground, where it will be developed with even less care and oversight than it is now.

With way less access to infrastructure, hardware, capital, people. That’s literally what we need, it should be constrained by having way less resources

Having fewer resources such as security engineers will not lead to better outcomes. A skeleton team with gigawatts of compute in some middle east bunker is MORE likely to unleash something poorly aligned.

Gigawatt of power doesn’t come from nowhere. You cannot have that much power consumption without being visible to regulators (of course that assumes a government that isn’t as corrupt as the US).

You can create bioweapons today, from a small lab. Doesn’t mean we allow it

> We, collectively, need to decide where to draw the lines on this technology and we need to decide fast.

I hope you, collectively, don't start deciding what deodorants I can wear. (Like Bernie Sanders infamously brought up.) Or what toys my daughter can play with. (As Donald Trump wanted.)

Society abides by a lot of rules that you also learned to follow, mostly while you were raised. Why else would you even wear a deodorant?

I wear deodorant to get along with the humans around me. Concrete folks I can point to (even if I don't necessarily know the guy next to me on the train by name nor will see him again).

No nebulous society involved that makes collective decisions in the abstract.

You got it backwards. Society is simple the collective that forms transitively from you, your connections, theirs and so on. You are pointing at a tree and missing that it's standing in the forest.

> Society is simple the collective that forms transitively from you, your connections, theirs and so on.

Where do you draw the line? Does the collective that makes these collective decisions you are so fond of have to include, say, people in Russia?

You can default to existing political structures and voting systems as a preliminary answer to get going. And they work for some things. But ain't nothing simple nor god given about that.

There is a subtlety here: on the one hand there are decisions that someone decides on and tries to enforce. On the other, there is collective behavior and what it does and doesn't tolerate - societal norms. The later forms in various ways and can overcome enforced decisions.

It's not on me to declare who is and isn't included or what consensus is being reached. These things normally play out in a self-organizing way.

But some companies try to disturb that process by not allowing it the time it needs. And I think this needs awareness.

Ok, I'll start shunning people who work at AI companies from now on. Let's see if that makes a difference or if they care.

Name one technology society has decided to stop and it stopped. There isn't one. Lots of people have made angry sounds about it, and made doom and gloom predictions about them, but none have ever been stopped. Nuclear hasn't been stopped so don't think that's your great comeback.

Some examples:

- CFCs are completely banned (Montreal Protocol)

- drag-reducing swimsuits

- some Persistant Organic Pollutants were completely banned (Stockholm convention)

It can be done.

Based on https://annas-archive.pk/blog/physical-destruction.html I wouldn't assume being in a niche is enough.

Could you please elaborate?

Reply on news.ycombinator.com