90% of US working hours are spent on tasks not covered well by AI benchmarks
mercor.com
mercor.com
In an ideal world (yeah, I know), it should result in them being paid for what they actually do.
One of the most troubling aspects of human's poor planning ability is they're absolutely sure something is valuable until they have to make a skin-in-the-game commitment to how much value it provides them personally. Then it turns out they weren't serious.
How? How do you set a price based on what someone does?
the classic revealed preference vs stated preference.
People have no problem paying something for nothing, if it gives them a chance at a disproportionally great return when they actually need it. It's called insurance, and perhaps we should start framing some of the intangibles in this way.
I.e. paying the premium for possibility of human interaction isn't paying for niceness-as-a-service, it's an insurance policy that pays out in kindness whenever you need it, wherever you are. But like every insurance policy, it only works if there's critical mass of people pooling resources, and stops working if they evaluate it as investment.
Which is to say, it is sometimes a fallacy, but often, it is not. We have far more intercoms and automatic door openers than doormen.
First they provide real security an intercom does not. They know everyone and won't let an unknown go through the door without checking in.
The courier leave their wares to them, and can hand over packages for the courier.
There's no extra cost, really, to having them do dry cleaning since all they do is hand over all the dry cleaning to a service that comes by.
We have to know how long it took.
I’m sure it’s similarly true for travel agents, real estate agents, probably lawyers, etc. Just because there are circumstances when someone values them doesn’t automatically justify them everywhere or refute that many of them basically don’t do anything.
Real estate agents are a good example. They might be the only people involved aware of the processes everyone's engaged in and moving things forward. Having helped build tools for a local shop I can tell you that their checklists and flowcharts can't and won't be built with AI because hoarding and sharing information has always been part of their competitive advantage.
A corollary might be that the value of human involvement is unknowable because human beings aren't rational. The Doorman is a buffer against chaos.
The value of the doorman isn't the intercom functionality.
People like doorman buildings because the building has a real person with eyes on the premises at all times.
An intercom or automatic door opening system can't deal with delivery people who can't figure out the intercom system, don't follow delivery instructions, leave your package/food somewhere random, etc. And that's only a fraction of the value a doorman provides.
An intercom system won't deter trespassers, petty crime, etc.
As long as the building is large enough, the cost is negligible.
I randomly picked a listing in NYC, 219 East 69th Street. 145 apartments on 13 floors.
Let's say you're paying a doorman 24/7 at $30/hour, that's $150/month/resident.
Was this supposed to be negligible?
The building I randomly picked out wasn’t particularly large, either. The larger the building, the smaller this cost becomes.
I know someone living in a 900 unit tower. A doorman is an obvious benefit in that building.
But if you never bother talking to the person at the door then you won’t see much value and thus an intercom seems just fine.
I’d choose the building with rather than without.
Hell the smoother interface and interaction model with the apartment admin alone is worth it.
I live in an apartment without one and I would not pay $150/month for a dedicated doorman so that delivery people have an easier time with the intercom.
And honestly, you are undercounting the cost, you can't get away with 3 doormen for 24/7 coverage and the salary is not the all in cost. It's probably closer the $300/month for 145 apartments.
If you're already paying $10k/month for an apartment, sure, you don't care about the cost and are happy with the fringe benefits. This doesn't change the fact that the vast majority of people have an intercom because the main function is solved.
The Doorman Fallacy is exactly this—thinking that one function of a job is the whole job. The doorman doesn't just open doors nor relay messages. They're eyes and ears, security, friendliness, an extra hand with bags, local knowledge, emergency response... The list goes on and on. Almost none of their functions can be replaced with an intercom.
Much of the US has buildings that are too small to justify a doorman regardless, but something is lost for it.
"Individuals aren't naturally paid-up members of the human race, except biologically. They need to be bounced around by the Brownian motion of society, which is a mechanism by which human beings constantly remind one another that they are...well...human beings." - Terry Pratchett
There will be cases where they stick around (luxury hotels, etc) where the ancillary benefits are the point, but there are many cases where the primary task is the economically valuable thing.
Respectfully, it is not.
Is it more efficient in America? Yeah. Is part of that scene her being bored while waiting? Yeah. Are gas prices there probably a bit higher? Yeah.
But there really was something appetizing about it. I’m not sure I could put my finger on exactly what.
Why would he be watching the Bruins and not the Devils?
The tipping.
It's fucked because it distorts the relationship between staff and customers, not to mention it lets businesses get away with underpaying their staff.
Fortunately Japan doesn't do this.
<edit> Ah, my Oregonian neighbors inform me that in the last few years, this mandated institution is no longer in force.
It's the fact you were not there.
Low end labor costs tend to track cost of living costs. In high cost-of-living places, you have high labor costs and vice versa. Most places are much cheaper than America, which means you can afford to hire more people to do tasks that would otherwise be uneconomical.
Do you happen to know where one could find statistics on this? This doesn't mesh with my experience at all, and a quick search did not turn up reliable data in either direction. Not-logged-in ChatGPT gave me an estimate of 40-60% worldwide, but that obviously has very low certainty.
My team can crank out high quality work in a few hours and AI could do the same medium to high quality work in a minute or two, but the review, feedback, and sign off takes days of back and forth. AI just slows this part of the process down because editing AI work is often a challenge because, no one is quite as familiar with how it was put together and sometimes it creates things that are not editable at all. And telling AI to make very specific changes is often a mixed bag at best.
Solo devs and small teams with minimal oversight can probably move a lot faster, but corporations are going to see minimal uplift except in areas where no one is looking.
There's a couple types of companies where the reverse is true: young companies trying to grow, and dying ones that need a turnaround.
Out of those two, young companies are less likely to be set in their ways. So the first place you really see the result of development moving faster with AI is probably going to be in future months and years when suddenly-big companies start getting widely known with products and business strategies that wouldn't have made sense a decade ago because writing code to try different things to even get the ball rolling was a bottleneck, and you could only do things that wouldn't scale for so long.
It's roughly the same pitch that the ruby on rails crowd had 15 years ago, and the no code crowd 5 years ago.
Very few companies need to rebuild their product every week, or actually care what tech stack it runs on. They are trying to build new things, or reduce how often those things break, or figuring out what if anything is worth building next.
Spot on!
Usually, most work in corps is slowed down by executive decisions. It takes 3 - 9 months to get a yes/no on something and for most things that are obvious that need to be done need to go up the chain of command. The decision takes forever, but when the decision is made, they want things to be done quickly, which creates unreasonable stress on the ICs.
The first step to see real productivity gains is for AI to eliminate the executive and middle-management layer, this starts a virtuous cycle, destroys both empire building and massively accelerates true productivity, not just meaningless metrics like LOC.
So the first signs of productivity gains will be shrinking head count, but it's going go from the bottom up, as usual.
I have no idea how it’s sustainable but I dont see AI CEOs coming until a new company becomes successful from the bottom using that model.
> most work in corps is slowed down by executive decisions
I work in a startup and I find it to be the opposite. Management wants us to go really fast and engineering is like "well wait we should slow down because just because LLMs can write a bunch of code doesn't mean it can build a maintainable app"
That said, I know that our small engineering team can't build an entire business by ourselves. So some management is required. But as you said, getting stuff done quickly is stressful. And with LLMs we're expected to get a LOT of stuff done very quickly. Some colleagues have decided to go on quiet strike by submitting AI slop and citing "no time to review" to prove a point. I don't fault them, but I also know that me and someone else will have to clean up their messes at some point.
However, if a PM tries it they will more often than not get a wrong answer because they do not know what to look for or how to ask the right questions.
Prompts still matter a lot especially when you are working in an established enterprise.
[1] https://communitysupport.group/t/what-is-the-purpose-of-the-...
That's just bullshit. And barbaric.
Making someone stand in place all day primarily to open a frickin door (or other menial tasks) for others is a degrading job.
What "vital functions" exactly?
Whenever I see someone in a job where they're not even provided a chair, I feel ashamed. Like doormen, guards, "traffic police" that act as meat-based traffic lights, standing right smack dab in the middle of exhaust fumes all day without even a face mask.. it all seems such a ridiculous and archaic ceremony
Should we also bring lamplighters [0] and phone operators [1] back?
I thought bellhops do that, that's not just a "doorman"
You're confusing the word "doorman" to include other random unrelated jobs, like expecting the postman to also deliver your milk and then complaining that email doesn't solve that problem.
(I guess if you ordered milk online, he could, but that's not the meaning regular people assume when saying "postman")
No. Coporate fired a bunch and now doorman is doing their work, too.
- Identifying suspicious people and situations.
- Helping people carry things through the lobby.
- Passing messages before cellphones.
- Receiving packages.
- Making sure people in the lobby behaved.
And probably more that I never realized as a child. Doorman doesn't necessarily mean "suited guy standing attention outside the building next to a door". The locks were also electronic, so they didn't even have to stand up to answer intercom buzzes.
The doorman was called a doorman because its essential task was the door. Everything else was additional and replaceable or unnecessary in the first place.
Same with AI. Once it has replaced the main task, the rest will adapt around it.
Look at how beautiful old buildings were. However, now everything is a box, because it is more efficient.
The people setting the prices are not some omniscient backroom of guys smoking cigars. They are thousands and thousands of people much like you can me, competing with each other.
Deluded.
hahahah pls stop posting about economics. For people who have gone much deeper on this branches of knowledge and have a far greater understanding of how it fits in reality - your posts are insulting to read. Stop.
Modern buildings draw the line somewhere else. I too wish there would be more beautiful buildings, but not at the old price.
Honestly, rising house prices have fucked everything up.
I don't see why it doesn't count. It shows we still value the service but it is out of reach for anyone who can't afford it. It lost in the same way much better food outings went out if business while McDonald's thrives. By making society arguably worse and punishing quality.
There are many other little examples like this, where beneficial side-effects go lost from automation. I suppose LLMs serve as a prime example of how this can affect people quite negatively (e.g. skill atrophy), if not careful. I think adaptation only goes so far, and we are being arrogant in thinking that we can adapt to everything, because a large part of how we function as humans is not as adaptable as we think, e.g. our emotional and instinctual needs, or how we maintain skills that are essential in the long-term.
It also has been an incredibly short period with incredibly large changes, to be able to say that things will just adapt; perhaps they will, but at what cost, and who wins from it?
We stopped going to one favorite place because the wait for a table became interminable. You’d spend forever standing there watching people waiting 15-20 minutes to get a busser’s attention so they could ask a question about the food, and then another 5-10 minutes for someone who actually knows the answer to come out and talk to them. Or, after the meal was over, people would just camp out indefinitely.
Loss on every side, still happened.
Okay, but were you coming to the restaurant for social interaction with the waiters? Have you stopped going to those restaurants that serve you good food at a reasonable price and which are otherwise desirable places to eat because of the lack of social interaction with waiters?
People dislike places that charge too much for mediocre service, and those places disproportionately use the QR codes for ordering and don't pass on the savings from labor reduction to the consumer. Very few people actually miss their waiter interrupting their conversation to tell them the specials.
I agree with your last point though, and you can turn it into a more general statement: some people or companies use automation to do more, and others use it to do less. There's probably a reason for it, among others most likely deprivation. People and companies are getting pushed to 11, and take shortcuts.
Maybe, you'd say there are more automatic doors than doormen. I'm confident that is true, but those automatic doors haven't supplanted doormen; they're automatic doors where there would have been manual doors before.
So everything hasn't adapted around automatic doors. We've just got a lot more automatic doors now. And a lot more doormen.
1) What proportion of doors had doormen before compared to now?
2) What proportion of the population was employed as doormen then and now?
3) What proportion of the population benefited from doormen then and now?
That said, to your actual points, as an intellectual exercise: while the population of New York has increased since 1930 (from around 7m to around 9m), the number of doormen has gone up much much faster (from "a few thousand" according to NPR to 25,000). So on the criteria you listed there, the number of doormen has increased as well.
The doorman example remains true.
Sometimes I feel like peter theil smokes a bunch of crack and posts antichrist takes on here under hundreds of alts. There is a lot more to life then instaspending money on Amazon, running on social media instead of talking to the kids, calling your dad for advice instead of chatgpt. These takes are exactly why most brands suck and why people will fight in the streets to defend Costco or valve.
Life will always have the intangibles and believe it or not it's still a good idea to hire someone who is only okay at a job but has an awesome personality and work ethic. Or lose a couple bucks on some hot dogs... 99% of purchases above survival are done for emotional needs and desires.
It's sad how the stupidest forms of short term cost saving and efficiency have driven ghouls to think they belong in business long term.
the one with a doorman & porter vs the one with just an automatic door and elevator.
people really need to read alchemy by roy Sutherland.
Doorman and porter means more awkward social interactions with strangers, and “premium” just makes me feel like I don’t belong there. I prefer my hotels to be simple and automatic. Airbnbs with self-service check-ins are ideal.
The wal-mart greeter is essentially a doorman for automatic doors.
Maybe there’s a connection.
For example, the “social workers & care -> counsel parents with child rearing problems” category is considered “well covered” because there are 13 prompts.
It’s self-evident that the spectrum of work carried out by a social worker undertaking that work cannot be well represented by thirteen prompts.
I suspect if you really strictly defined “covered well”, to the extent that any normal person would use that phrase, the headline would be above 99.9%.
I've started to notice this on a micro scale as well. I am 20x more productive at coding and scripting, but I don't spend more of my time coding. My time is now filled with things like presentations, Jira, QA, etc. Ironically, despite all of the automation, the things that are painstaking and time consuming disproportionately require more time now.
When you think about it humans were never really well suited to such a lifestyle. Of course the transition may be bumpy but maybe the end point, of machines doing these machine-like tasks, isn't so bad for us.
What work does suit humans? Anything involving interaction with the world or other humans, perhaps.
That's the take many have been educated against: maybe one day a category of bullshit jobs will be gone and it's fine.
I'd rather not go back to hunting and gathering. Though I would like to go back to trust busting, collective action, and rule of law.
Each incremental reduction in friction and improvement in comfort feels good. But is the net effect happiness? I think it's anomie, in the limit.
Also if roofing was a non-problem anymore (we solved it somehow), would you want to have people doing it just for the sake of doing a job?
> ...
> But even after controlling for AI-exposure, O*NET tasks from occupations with larger wage bills are significantly overrepresented among benchmark prompts:
So, in short: they want AI to take the good jobs. AI is not coming for jobs with tasks like "gather and empty trash" or "service, clean, or supply restrooms," or "mix water and detergents or acids in containers to prepare cleaning solutions according to specifications." We software engineers and lawyers would be smart to start retraining (and adjusting our living standards) now!
And this points to one of the ways AI will help improve society and living standards for all: we'll have cleaner, better-maintained bathrooms, as labor one wasted on engaging intellectual tasks is redirected towards menial, low-pay tasks that still need a human touch.
Reason to be optimistic, amirite?
More and more it’s seeming that the “societal benefits” from the last fifty years of progress have largely been concentrated at the top.
Right now it's £5.60, but that says more about the decline of the British Pound than an increase in the price of wheat.
Massive food price deflation has been one of the wonders of the modern world. Hopefully the last few years were only a small blip on this overall trend.
"It might get better in 500 years" is not exactly comforting thought.
https://www.visualcapitalist.com/inflation-chart-tracks-pric...
Doesn't include foods, but it certainly punctures the "That family isn't aaactually in financial trouble, they have a refrigerator and a phone and a TV" statements from certain pundits. (A class of people who are often disconnected from the average person's budgeting situation.)
Spending on food as a percent of income has declined significantly (~30%) over the past 50 years: https://ourworldindata.org/grapher/food-expenditure-share-fa.... The decline has flatlined over the past ~20 years, but at least some of that is affluence meaning we choose to eat out instead of cooking at home.
House spending as a percent of income also seems to have declined over the past 50 years (https://economistwritingeveryday.com/2023/04/26/spending-on-...). To be honest, this one surprised me, I would've bet the other way.
There are always individual counterexamples. And yet as a whole the cost of healthcare has exploded.
> at least some of that is affluence meaning we choose to eat out instead of cooking at home
A lot of that is poverty, not affluence. People working two jobs and living in food deserts aren’t cooking a lot of meals at home. It’s expensive being poor.
> People working two jobs and living in food deserts aren’t cooking a lot of meals at home
The percent of people with multiple jobs has declined over the past 30 years: https://fred.stlouisfed.org/series/LNS12026620 (longest time series I could find)
Cheap food is still cheap. If some doomsday scenario ever comes along, I'll be forever sorry to not have stocked up on 6 years worth of calories that now cost less than a monthly salary. World market prices per metric ton: wheat like $300, rice $500.
But that's not mutually exclusive with absurd levels of wealth concentration.
they haven't. In the early 1900s an American household spent roughly 50% of its income on food. And mind you food didn't mean an exotic fruit you flew in from the other end of the world or pounds of meat per week.
Today an American household spends 9% of their income on food, the 'dramatic increase' may have bumped that up to 10% by some measures. The size of the average house of the average American family in the last half century has doubled while the family size has halved. In the 1960s 12% of households had AC access, today it's over 80%. Generics of basic medicines cost virtually nothing, you've just started to consume more sophisticated medicine over time.
If you're an ordinary guy you are already living in luxury, you just don't notice it because you kept moving the goalpost on what luxury is. Which isn't per se wrong but you ought to be at least aware of it
The only “disinflation” consumers have experienced is devaluation of products through obsolescence as newer, often more expensive alternatives enter the market.
I'm sure some storage servers or storage appliances were built with many TB even then but it wasn't really something I saw a lot.
I'm almost cynically happy about Moore's law ending and hardware getting expensive if this is what it takes to improve software in the long term.
Of course, competition leading to near-zero profits means most competitors collapse, and the remnants get bought up by monopolists or oligopolists who then become rent extractors.
The technology was fundamentally developed with government grants to universities, or corporate labs like Bell Labs that were funded by a legal mandate on its parent corporation. Capitalists can only develop products that are basically ready for market, their margins are thin and planning horizons can only survive about 5 years usually. Their main power is throwing oodles of money at something ready to go and scaling it.
Edit: I will be slightly more charitable and note that capitalist companies can also do incremental research if there is a gentle and obvious hill to climb. Sometimes people do make crazy bets and they pay off, but that's more the exception, in America the dominant driver of technological innovation is government funded research supplied by capitalist scaled technologies (i.e. you research the next paradigm in the lab and purchase e.g. pre-existing computing resources from a company).
I find it funny when capitalism is blamed for rent extraction.
First off, extracting profit is not unique to capitalism. It happen in the Soviet Union as well, just on the black market. Even in the legal market workers would "sell" things they had control over for personal benefit. It's not like this disappears with other economic systems - it's a core human behavior to do things that benefit yourself.
Second, almost all of the things you listed - housing, medicine, education - are rent extractions not because of capitalism, but because of government regulations that allowed rent extraction. If people could build as many houses as they want, treat patients without jumping through the AMA quota hoops or charge a tuition that students can actually afford (and not backstopped by the government), none of this rent extraction would happen. We have plenty of examples of those economic sectors where prices are much more reasonable.
That's an assertion often made without evidence. Humans are one of the most cooperative species in existence.
Sure humans are cooperative, but but it's not an either/or scenario. Humans are often incredibly selfish at times, and even if not selfish are often focused on benefit for themselves.
And my point is that it's not unique to capitalism. Even prior to capitalism and under communism, humans are highly focused on benefit for themselves, even at the expense of others.
You are cherry-picking to come to that conclusion. Humans are constantly navigating the pull in either direction. On the one hand, they are pulled toward what benefits them, and on the other, to what benefits their group and to what gains them acceptance from said group. Often this means putting the group's needs on equal footing to the individuals.
You are correct that the pull towards the individuals' needs is not unique to capitalism. You are not correct that humans are "highly focused" on benefit for themselves at the expense of others.
I would say that some humans act as you say; but in the court of a king, your "group" reinforces behavior that is selfish on a societal level, so long as tribute is paid to the king.
Most people do not live in the court of a king. I live with a family. I put the needs of my family on equal footing with my own. I maintain that this arrangement is more common than not. For every ambitious individual there are countless thousands who want only to love and be loved.
You put your family’s needs on the same footing as your own. And you put those two needs far above people outside your family.
That’s exactly my point.
People sometimes act in the good of everyone, usually when the cost isn’t high. But otherwise they act in benefit of themselves or their group, often at the expense of the out group.
You don’t need capitalism for that to happen.
Agreed. I understand what you are saying. You don't need capitalism for absurdly selfish behavior. However, I don't know that a capitalist system, in its current, relatively unregulated form, is making humans less selfish.
The form of capitalism now gnawing away at US society is essentially shouting "the virtue of selfishness" from the rooftops. People respond to leadership, including moral leadership. There are better forms of capitalism that involve regulation and fairer taxation, such as we have had in the past, but no longer have now.
For a lot of people criticizing capitalism, the goal is not socialism but rather better capitalism.
A colleague of mine that went to an Ivy League school in America told me her friends in Poland, not a wealthy country by any means, laughed at her for "buying her degree".
Can you? I assume you mean spinning platter, but even then.
Computer parts are unaffordable this year.
“I can't eat an iPad.”
> Best not to cite the price of the new iPad as an example of why inflation isn't a problem when you head into a working-class neighborhood. In Queens, New York, on Friday, New York Fed President William Dudley did just that. He got an earful. After being bombarded with questions about food inflation, Dudley attempted to reassure his audience by putting rising commodity prices into a broader economic context -- but that only made matters worse. "When was the last time, sir, that you went grocery shopping?" one audience member asked. Dudley tried to explain how the Fed sees things: Yes, food and energy prices may be rising, but at the same time, other prices are declining. He then stretched for a real world example. The only problem was he chose the Apple's latest tablet computer that hit stores on Friday, which may be more popular at the New York Fed's headquarters near Wall Street than it is on the gritty streets of Queens. "Today you can buy an iPad 2 that costs the same as an iPad 1 that is twice as powerful," he said."You have to look at the prices of all things." This prompted guffaws and widespread murmuring from the audience, with one audience member calling the comment "tone deaf." "I can't eat an iPad," another said.
https://awealthofcommonsense.com/wp-content/uploads/2020/11/...
The prices of all those things continue to skyrocket.
Just like good blue collar jobs were shipped overseas, good knowledge jobs are going to be shipped to GPUs whenever they can.
Which doesn’t play well with “savings for consumers”, because these consumers will be unemployed.
I know, this time it’s different, AI is different, blah blah blah.
You pick the highest leverage occupation first. Automating lawyers gives you.. angry lawyers? Automating programmers gives you improved ways of doing literally all other things because all other automation requires.. automation. That is to say bucketloads of software.
It’s completely obvious what the play is here and it has nothing to do with pay. We pay peanuts for tokens so automating a developer for 200-500 a month worth of tokens ain’t getting you to a trillion a year.
RSI is the play and it’s already happening. Our work is easier and faster and anyone that claims otherwise I have a very hard time taking serious.
You think they come for mathematicians’ pay? If you want to earn money I guarantee you they need to automate some other low-hanging fruits first that pay a lot better.
No, programming and math is basically what AI research is. You automate that and you got the golden ticket to all other things.
Most importantly, in these fields results can be easily verified which means that if you have enough compute, that you can generate correct training data. OpenAI had an early story - which could be a lie of course - about how they tried to train a robot but the actuators would be toast before the robot would have learned. Math and programming doesn't have such problems.
There are lots of fields with easily verifiable outputs but that doesn’t make them valuable targets.
Eventually you’re going to need robotics to kick off the singularity, sure, but I would start with programming and math (and compute) which is exactly what they are doing.
Let’s say you start with robots, like we did, but we lack compute, we lack dev tools, we lack the ability to generate vast training environments, we cannot automate mathematical research of any kind in any way shape or form not even trivially, in general we are bottlenecked by math, code and compute. “Doing robotics” first without tackling the others is taking an astronomical detour which we as humanity are fond of.
I find the notion we automate cognition itself first and then tackle the secondary problems quite rational.
No it isn't. Doing work more efficiently through the use of LLMs is not RSI. RSI requires a closed loop automated LLM iteration process which we are not at all close to right now.
Sure, if you redefine what that means.
Before this year, "recursive self-improvement" in the context of AI research meant the holy grail: an artificial intelligence that was capable of improving itself, unattended, in ways that humans can't even imagine, so that within a relatively short period of time, it reaches superintelligence and ignites the Singularity.
What's happening now is emphatically not that. And if you think it is, you are buying straight-up propaganda. "I gave an LLM a tool that lets it make largely-random decisions about how to train its next iteration when I tell it to do so" bears little to no resemblance to RSI. Because it is, beyond a rounding error, completely missing out on the "improvement" part.
Could be hallucinating by I just asked Gemini to estimate global average compensation of developers and multiply by estimate of total number of developers in the world, and it came up with $1 trillion to $2 trillion.
So the total addressable market is the right order of magnitude, let alone all the other automatable high wage occupations..
Market forces drive it down hard. The USA is not the only player. Open source models are not vastly behind. My guess it that the 100-200 range will remain the baseline no matter the capabilities.
We're automating high paying careers and making what, the consumer's todo list and budgeting apps cheaper for them? My expensive smartphone is going to cost me ~$10/mo over its lifespan. I don't care if most of these consumer products get cheaper when they're a tiny portion of my spending
Bud, thinking that for-profit corporations - many of whom are in an oligopoly or monopoly position in their markets - are going to pass along savings from not paying people is like thinking the stripper really likes you.
Remember how things were cheaper before the pandemic? Well, there were supply shortages and prices rose accordingly. Supply and demand at work, nothing surprising there. The supply chain shocks eventually went away. Some prices have sort-of come down, but many just haven't. Why? Because the people selling goods and services now know that people can and will pay the higher price.
Now, they might do that through taking on debt and working additional jobs to the detriment of themselves, their families, their communities, and their country, but no one who both matters and can do something about it keeps track of those things.
"They" figured out that it's a lot easier to hillclimb AI on verifiable digital tasks, and it just so happened that white collar jobs tend to pay more
Ai will decimate customer support jobs but those are not considered highly paid
If robotics continue to progress, we'll have all the robots to do our dishes and clean our bathrooms as you wish
the same robot that cleans your bathroom can also wash your dishes, fold your clothes, defend your home, fix your gutters, etc
(the bottleneck of which is the brain, aka software)
No, we are currently unable to do most tasks that require physical manipulation, independent of what anyone wants.
This is also the reason absolutely no attention is spent on evolving B2C products.
What is being exploited is the overt attention businesses pay to their own cost metrics, to the point where broader strategy is getting thrown under the bus in favor of efficiency. I believe this is an illusion, and for many companies that jump on board, sober learnings are right around the corner.
I think you may also be incorrect on your notion that training data is proportional to good models if anything the opposite is true. what ai models are ultimately trying to be are a universal theorem prover and encode all human knowledge into "theorems" and rulesets
not trying to come off too harsh here but i think most of the work is just innovating on the frontier or people pleasing
It's just harder to do. Would not be at all surprised to see e.g. Autodesk bought out for exactly that purpose.
You don't even need Autodesk. You can have Claude whip up a web app and build a higher level understanding of the building and then translate it to whatever output or format you desire.
Only if people care enough to pay extra; the majority of the market is for functional buildings that look ok, and unless we decide to value humanity, slop buildings will be cheaper.
In fact, a lot of stuff is already slop, largely because of the design and engineering expertise required. I don't think one-shotting "design me a house in the style of Frank Lloyd Wright with these specifications" is that far off.
Oh don't worry Mr. Former Software Engineer. You won't be picking up the trash or cleaning toilets. You have things like an expectation for your work to pay a reasonable wage. You'll be passed over for a third world immigrant who'll work for a pittance, and they'll be replaced only when robots are capable enough and cheap enough to undercut _them_.
Imagine I give you for free the perfect trash picking robot control AI software. Works great. Benchmark smashed.
Now imagine what the rest of the robot is like, and how much it would cost. I mean really imagine, not just picture C3PO with a bucket and litter spike. How many motors and what are the torque vs. mass requirements? How is energy stored? It goes on, not all of it is solved, and none of it is cheap. Except compute.
AI likely does accelerate robotics usefully - certainly in sensing and control. But there are major bottlenecks that are not software, that AI can only help with as a generic engineering tool, same as it helps planes and fridges.
Imagine I give you a perfect roadmap to build that robot...
Responding to your edit, none of those things strike me as any more challenging than the control software.
That trash picker may never compete with a human. There's no reason to assume it will.
edit:
> none of those things strike me as any more challenging than the control software
Maybe the difficulties of motors and batteries are not intuitive. Look at their improvement rates over time. It's a far cry from transistor counts or FLOPS. And look at the performance and efficiency of humans and other animals. We are absurdly strong, light, efficient and cheap to maintain. The gap is huge and not closing quickly.
If it's not C3PO it's a Boston Dynamics dog with a couple appendages, better software, and automated manufacturing.
iRobot, the company that shipped the most robots in history because they focused super hard on finding a cost-effective subset of cleaning tasks, failed to scale up to other even closely related tasks, went bust and was sold off.
But yeah, it's probably straightforward, right?
I didn’t make any claims about the relative difficultly.
Advanced robotics would need to become quite cheap for them to become more affordable and convenient than a minimum-wage employee. And physical fields like materials engineering are known for moving quite slowly and at their own pace, so I'm not sure how much automation could help there.
That's the last barrier we have right now that prevents the total obsolescence (and therefore destruction) of the common worker.
1. The discussion setup where you imagine a counterfactual is given by magic does not entail that magic can do any other thing you choose.
“Imagine I give you $100, what do you spend it on?”
“A McClaren, since you could have given me $300K”
2. Taking your question directly: because these are not closely related problems.
My question was a sincere ask to define, in this scenario, what is prohibiting that other magic?
Do you have relevant experience in any of those fields?
If you step back, the LLM revolution is just "computers now respond to human instruction, like they were always supposed to".
If you think of this field as a historic accident, and the current shift is a very unfortunate reset - it's a little less dystopian. In the long run, if it didn't affect me immediately, I don't mind a society that values "keeping restrooms clean" a little bit more and "talking to computers good" a little bit less.
> ...
> If you think of this field as a historic accident, and the current shift is a very unfortunate reset - it's a little less dystopian.
No it's more dystopian. The "historical accident" was a little respite, where some workers could be comfortable and well off. They had a little market power to take the edge off their exploitation.
Now that market power is taken away, and they can enjoy the poverty they were always "meant" to have.
> In the long run, if it didn't affect me immediately, I don't mind a society that values "keeping restrooms clean" a little bit more and "talking to computers good" a little bit less.
Come on. Society isn't going to value "keeping restrooms clean." The people doing that will continue to be poorly paid and disrespected. It's just the opportunities to do something better will be taken away, so there will be more labor available for that kind of work.
That is the only path under Capitalism, so we shouldn't expect any different.
That's a bizarre non sequitur, pure wishful thinking. Its like saying Maybe getting cancer is good, because it will cure my arthritis.
Isn't it just typical the humans get automation v1 the wrong way around, haha
As someone who recently switched from data scientist to electrician, starting the day can be pretty tough because you have to get physically moving and uncomfortable, but I feel way better at the end of the workday than after an office job. Most of the time, the work is pretty uncomfortable, but somehow it's a lot of fun.
Of course many software "engineers" are nothing more than plumbers/electricians with data and that cohort is being quickly pushed out by AI. And our industry in general has conflated technicians and engineers which makes this worse.
But fundamentally, engineers job is to model the problem from the first principles and then apply that model to whatever problem. Technicians operate within the model, so I postulate anyone considering themselves an engineer is fundamentally a more capable electrician, plumber, carpenter, whatever than any master tradesperson.
I'm not speaking as someone that changed an outlet once and thinks they are now an "electrician". I have two decades of extensive property maintenance experience. I do everything myself and since my state does not allow pulling permits without licensure, I hold myself to commercial standards minimally. So EMT, MC, trapezes and gutters, no romex crap. Similarly, I will likely outsolder and outthread at least the current generation of "plumbers" that only know how to press things.
What I'm saying is extremely evident in any electricians' forum where its obvious that electricians just do not understand why is NEC written certain way. Then there' usually some "oracle" on said forums "deciphering" what the code is, despite it all being a pretty basic and coherent spec.
Ah, I see you don't work in industrial control systems that risk life and limb!
> anyone calling themselves an engineer by definition is should be as capable as "technician" level workers
I agree, but these software engineers fucking struggle in a modern datacenter.
Also adding a new circuit to a panel is different than extending one from inside
I don’t do the work myself.
Old work is always harder than new work. It varies by how your home is designed. It varies by where your panel is located and where your garage is located. You will pay a premium for the extra time and pain in the butt it causes.
As a great example, my friend asked for assistance with their home's street light not working. Long story short is that they ran a UF (underground feed) line directly under her concrete driveway (the contractor was too cheap to run it in conduit). I megged that line and it had a short. Their light stays off because rerouting through their house then digging up their yard isn't something they want to do (don't want to pay as I don't have that much free time and they don't want the massive headache it would cause). I was able to disconnect/wire off the failed part of the circuit so everything else works for them, so that's a win (and I didn't charge, so that was a win for them too I guess).
The work is was doing in big tech was pretty far from anything that would help people in any tangible way.
The money isn't as good but the office view is way better.
I'm glad to see very poor coverage for transportation jobs on the AI benchmarks. May that continue until I'm done flying, at least AI probably won't ruin that for me any time soon.
I offset the costs by doing flight instruction, which is a very common route. To get to the flight instructor stage by renting airplanes is probably in the order of a $50k investment, with a significant variance depending on the cost of living.
As a caveat to this, all my experience is in the US.
No offense but pilots are probably at peak automation since their job has mostly been reduced to being a glorified AutoPilot system babysitters mainly needed for regulatory hedging. The ground transportation is still fair game though.
How can software engineers maintain a healthy "manual programming" skill set? What employer would be willing to pay for so much wasted time for their software engineers to do deliberate training courses, or willing to accept the productivity loss from their engineers occasionally working manually for practice, when working manually is so much slower?
Lots of regulation or some landmark lawsuits should do it. If airlines operated the way tech companies did, we'd have daily plane crashes.
Pilots have to know all the ins and outs of every single system and subsystem of the machine they operate. If software engineers were required to understand their systems to the same degree in order to ship something, the only way you could do it is by coding yourself.
It's not like the pilot instructs every single control surface individually. Nor does he physically carry every single passenger on his back.
Just because you're using a machine to do your job now doesn't really change anything substantial about the job other than the fact that you now need capital to pay the machine.
I'm not sure how the news being delivered by a carbon-based pilot would help with the dying part but I guess hospice care exists for a reason.
Wow. You clearly have zero idea of what a pilot's job is, and it shows.
You are also severely underestimating how much work and regulation would be required to replace pilots. And this work can be undone in a flash the first time automation messes up.
Also note that plane automation heavily relies on sensors. Many plane accidents were caused by faulty sensors even with humans in the loop.
For example, engine falls off. The automation doesn't know that happened, it just sees various problems like the hydraulic pressure falls, and for some reason the engine thrust went to zero.
When I was at Boeing, an incident was recounted to me where the autopilot glitched and rolled the plane over into a vertical dive. It was going so fast it entered "separation", where the airflow no longer flows over the flight controls. In other words, the pilots were just along for the ride.
But the pilot figures, since we're dead anyway, let's try lowering the landing gear and see if it slows the jet down. He did, and it ripped the doors off and bent the gear back, but it slowed enough that the flight controls could get a "bite" and the pilots were able to regain control and landed safely.
Suffice to say that what's in current civil jetliners is very far from the state of the art. I fully expect properly implemented and integrated AI to be far superior to human pilots in the future - it probably would be already, if we put the effort in.
That's correct. But it's an easy problem to automate, because the problem being solved is well understood.
That's not the case for something randomly breaking.
I doubt the FAA would certify an AI running a jetliner, for the simple reason that if something unexpected happens, there's no way to predict what the AI will do. FSD for cars was done by collecting an incredible amount of data from the cameras on Teslas. Yet still, how much data was collected from Teslas having a catastrophic failure, like catching fire? A jetliner has a lot more failure modes than a car.
Any role where the human is a buffer against chaos should be thought of as secure by everyone involved. All automation and information flow will be in service to this position. This is not a bottleneck, it is the point of the exercise.
The difficulty with human abilities in vision, locomotion and particularly coordination and manipulation is due to their arithmetic complexity being high. Modern deep learning solves that, LLMs provide the "interface" for planned action.
From the outside, this process of "solving" perhaps looks like "robots stumble around haplessly". But it's just the same as with mathematicians in the intellectual realm: when the level of ability surpasses that of humans, it appears as a "sudden event" to outsiders, having occurred far faster than anticipated.
There are no "safe havens".
Where robots ‘win’ is following orders with no regard to self preservation or long term survival.
Ponder that for a bit, and it’s consequences get a bit scary.
Do you expect this to be true indefinitely? Robot production scaling has barely started.
The world is vastly messier than engineers like to assume, and every real world trial I’ve seen requires massive human babysitting of a robot by humans for it to last any length of time or be useful at all.
And we’re not talking about AI type stuff, but corrosion, terrible field conditions causing sensor issues (mud, sap, dirt/dust), power management issues, communication disruption due to terrain, enemy action, etc.
I watch enough Futurama&Terminator to know it's "kill all humans"
The safety brief is supposed to motivate you to work safely so maybe they need to work on how they present it. I saw a couple of colleagues injured in my old job (EMS) but only seen a member of the public decapitated.
I got lit up by 277/480 when I worked as an electrician because someone removed a lockout and turned on the breaker.
In any case, not all jobs can have the power turned off.
A large part of the trades is learning how to deal with toxic coworkers or shops, or understanding that you're being worked like a rented mule for no real reason. And I was able to leave all that behind for a moment.
...don't you think the ability to create virtually unlimited amounts of software, at virtually no cost, is going to lead to major advances in robotics? I realize there's a lot of very hard problems involved in e.g. robotic hands (to take one example), but once software development is practically free, I'd expect a lot of seemingly-impossible problems to get resolved by the sheer quantity of what we can create and experiment with.
Our labor power is going to be fully eroded and the robots will not be serving us.
In the future, do bathrooms need to be cleaned? If yes, is it cheaper for that to be done by a person or by a robot?
Obviously, it is in the interest of the robot manufacturers to make the robot just a tiny bit cheaper than the person, so you'll buy the robot and the company makes the maximum amount of profit.
Now, if society breaks down and there's no one to consider buying the robot—well, that's not good for anybody, which is why I suspect society will find some way of working this out even if it's not ideal. But whatever, I don't particularly want to argue this point.
What I do want to argue is that there probably isn't a universe in which all white collar work goes away but blue collar work is still around.
Not sure what this "society" that works things out is, exactly. Top-down high-modernist plans for how society ought to work are the past; chaotic, twitchy, thrashy responses to disruption by a tiny elite class is where we are now.
But there still wouldn't be any jobs cleaning toilets, because the people who still had toilets to clean would have robots to do it.
Previously privileged people having a hard time is maybe a good recipe for class solidarity, which is maybe something we could use in society. Maybe it will be good for us... collectively, if not individually.
Class mobility is good mostly because new money rising turns the soil of society -- a partial inversion of roles and positions, bringing [social] networks-on-high in greater contact with networks-down-low. An upper class moving down is in strange ways perhaps good for the "soil" of culture as well. It just has worse branding as a not-very-uplifting story. But it's better than rich-get-richer and poor-get-poorer, which is what's been happening a lot lately.
EDIT: But also, I suspect I would be "fuck-you angry' at reading this comment if I had a family depending on me.
It's the reason AI company leaders are so adamant about framing it as a post-singularity utopia with UBI. They have no answer other than "you're all going to be jobless, better figure that out." Worse yet, they see it merely as a policy problem, rather than a problem they're responsible for. It _is_ a policy problem, no doubt, but the angry mobs won't be coming for policy.
> It _will_ be "the rich get poorer and the poor get poorer".
That is the complete copium. All the stats and models shows that white collar class will get poorer, EVERYBODY ELSE will get richer. Plain and simple.
Would you mind sharing a source for that? I can't think of a mechanism that would make sure that the poor get richer.
Intuitively I would expect a continuation or acceleration of the current trend, i.e. concentration of wealth at the very top (billionaires).
You also have this hilarious talk of "privilege" with this weird assumption that all of the white collar workers of course come from great privilege (perhaps telling on yourself and letting your guilt come out), despite many of us having grown up poor, are from poor areas, and worked our way out.
Furthermore, why do you believe that the only possible outcome of class solidarity is violent revolution? Why do you believe that any revolution would inevitably settle on, presumably, a specifically soviet model of government and economic policy? I find that most american adults, even ones who otherwise seem intelligent, seem to have this picture of economic and power relations that just kind of sounds like a half-remembered bad cartoon of a cold war that ended before I was born.
Because most people are retarded and any class solidarity movement will be inevitably captured by the malicious agents.
> Why do you believe that any revolution would inevitably settle on, presumably, a specifically soviet model of government and economic policy?
Because this is the natural form of government with stated goal being the fair redistribution of wealth.
I can't imagine you're big on democracy per se either, huh
we let the rot grow for too long.
> I'll take you up on that attempt to find a silver lining:
I think you're confused. The true upper class isn't going anywhere, those are the people so rich they don't need to work (and if they do "work," it's tending their empire). They'll keep getting richer.
The "downward mobility" you're talking about the upper middle class and middle class. People who need to work, but can live comfortably on their wages. This is really about flattening the working class into its poorer levels (and kicking away the ladders the people in those levels could use to lead a more comfortable life, no more getting a college degree and graduating from the factory floor to the office).
It's about increasing inequality and creating an unbridgeable chasm between the poor and rich.
Are you talking about violent revolution?
Some disorganized thoughts:
1. The professional classes are so unused to violence, they'll probably undermine the chances of violent revolution (e.g. reflexively supporting gun control, seeing such voilence through the lens "school shootings" and other crime, and being likely to report it, etc.).
2. The professional class is losing a lot of its economic power. The time for solidarity and non-violent action way 5 years ago, but the "happy to play along" professionals (e.g. libertarian software engineers) botched their best opportunities. From here on out it will be harder.
3. As all of this is happening, technology is helping to undermining the ability of the masses to do anything about it: IRL community declining, people getting mesmerized and isolated by social media, AI-enabled surveillance, AI-enabled astroturfing and propaganda, autonomous AI weapons, etc.
4. Even if what you say is true, the clock is ticking. Wait a generation and the "former professional class" will be gone as a distinct thing, and their kids will join an undifferentiated mass of poor who have normalized and internalized their circumstances, as they have no first-hand experience of better times.
No they just need killer robots to shoot anyone who gets uppity.
Feels like a very similar policy being pursued here.
All those are physical things and physical LLMs are right around the corner. It’s coming for those jobs too within the next year or two I would imagine. Especially the ones you mentioned seem very easy for a robot to do. There are some physical tasks that will take much longer like being an electrician or a plumber or a roofer. There’s too much precision and intricacy involved in those but cleaning out a restroom and resupplying it that’s easy as hell. A robot like Figure 3 could probably do those autonomously today somewhat badly.
But is it cheaper than paying someone around $5 in labor to do it? Because until it is, it's not economical to have a robot do it.
> that being said of course I think it’s going to be way less than $5 an hour if you have a robot do it.
This is about an LLM-powered robot, not just a robot. At standard API prices, it'd run over $5 before it finished opening the door.
The physical LLM space (atleast Figure 3, Unitree, Gemini robotics etc) are doing on device processing and not over an API. So it costs what electricity costs.
Or maybe, idk, we do something about this?
I'm already a part of a generation that's the first in American history to have a lower standard of living than their parents.
Like what? There is no things that will get an average engineer or lawyer better result in the future than retraining and adjusting their living standards.
> I'm already a part of a generation that's the first in American history to have a lower standard of living than their parents.
Good. While US dominance over the world and it's ability to rob other countries will continue to decrease, your living standards will continue to decrease as well.
I'll say this: I'm teaching myself every skill I can think of. I'm already pretty decent at finish carpentry, structural carpentry, electrical, mechanical work, HVAC, auto repair, auto body, welding and a few other less lucrative things.
I was thinking briefly I might want to go to law school, but who knows how much longer that's going to be a thing.
[1] https://justicegap.lsc.gov/resource/section-4-seeking-and-re...
The actual reason is neither of those things. It's because not enough housing was built.
Leaving that aside I'm quite puzzled about something.
"Making everyone poor is fairer to the poor" is a curious idea that I've seen you express several times on HN recently. The first time I read it I thought you were being sardonic, but you sound quite earnest.
Given your previous opposition to communism and socialism I wonder why you have apparently changed your beliefs.
The supply demand equation has two sides. If you give a substantial chunk of the population way higher incomes in the face of only slowly growing supply, then you see skyrocketing prices. That’ll happen especially around the centers of knowledge work, which is exactly what you’re seeing.
The town where I grew up didn’t become unaffordable because of Jeff Bezos. It was because of all the Google engineers making 10x the median income.
> "Making everyone poor is fairer to the poor" is a curious idea that I've seen you express several times on HN recently…
> Given your previous opposition to communism and socialism I wonder why you have apparently changed your beliefs
I haven’t changed my beliefs. I don’t trust “socialists” who would use government machinery to artificially make people equal. But if technology and the market renders natural human inequality irrelevant, I’m all for that. Industrial machinery went a long way to making the physical strength differences between men and women irrelevant. That was a step forward. If AI renders natural differences in intelligence irrelevant too, that would also be a step forward.
Alternatively, skyrocketing home prices meant companies had to offer higher and higher salaries to retain employees in a competitive labor market. If there had been enough housing perhaps salaries wouldn't be as high.
Even police officers and other public employees in the Bay Area make more money than the national median. It's mostly because of the high cost of living, and that is mostly because of expensive housing.
I grew up at the boundary of a town like this it's like the truman show town, a few years ago they didn't want to replace a bicycle ferry with a bridge as it could bring in more cyclists from the (pretty rich) town next door lol.
The lump of labor fallacy needs to go away.
Yes, I'm sure all the wage slaves who are suddenly forced out of the labor market that was barely allowing them to scrape by paycheck to paycheck will suddenly start creating new businesses and become investors in new technology and not just become homeless leading to a massive increase in diseases/crimes of despair.
If (and it's a massive if) what passes for AI these days gets good enough to replace any substantial number of workers it will be a disaster.
Not all new business will be good for humanity though. Modern ad tech is shit, but it is huge.
It takes money to start a business. People who live paycheck to paycheck don't have money, and banks don't give loans to people who are poor and unemployed.
Company stock value goes up when they lay off staff. Next quarter shareholder maximization is the goal. AI and layoffs is the strategy to get there.
Some of the projects we've done would have required 4-6 additional people, but realistically, pre-ai, we would just not do the project or we would do it over a period 3-4x longer.
There's just so much to do.
As far as I know, that's how the lump of labour fallacy actually works! There's an equilibrium between supply and demand for labour such that workers are (more or less) choosing the best jobs they can find. There's infinite demand for labour, but not at any price, so demand destruction of the best jobs can still happen. There will be infinitely more jobs, but they won't be as good as the ones that were destroyed.
Of course, demand creation for really good jobs can also happen. What the GP is saying is that they believe demand destruction will outpace demand creation for really good jobs and everyone will be forced into less desirable jobs. I don't see how this is inconsistent with the lump of labour fallacy.
Reason to unionize, more like.
Yes like all the drivers and truckers who retained 10 years ago because self driving cars will take their jobs. Lucky all of them retrained in time!
Yes it is. YC has a company working on picking strawberries, even. Hardware takes longer than software, but if you think the robots aren't coming, you're not paying attention. The only question is how long will they take to reach the general public? There are test runs already in automotive factories and elsewhere, and yeah, they're slow and don't work very well, for now, but there's a lot of money going into making them better.
And then where will we be?
Back in 2012 I genuinely thought the self-driving car was going to put truck drivers out of work and reshape the economy on a timescale of less than a decade. This turned out to be a technofantasy. I think AI is going to play out the same way.
AI is not a panacea for work, and the economic bubble forming around it is going to collapse with devastating consequences before we begin to see adoption of the underlying technologies for real ROI spread widely beyond the software industry.
All the 'easy' to automate things have been automated already. A lot of things that were not automated is because there are small bits of problematic things that made it not worth while or the loss slightly too high. AI vision systems solve a lot of this.
The same goes for self-driving cars. It's a exponential, hockey-stick curve that at some point crosses the level of human ability. There is nothing "linear" about it and your suggestion of parallels to former experience is wildly misguided.
People fail to notice, most jobs are less complicated than math and programming. The "obstacles" there are of the kind of "we always did it this way", but that's no real "moat"?
The bubble regarding the big AI firms may well collapse, but the assumption, that was somehow improving the situation is far-fetched.
People automated out of a job will either compete with those in other jobs or be dependent on support. Either way, salaries will fall across the board and prices rise. At an ever accelerating pace.
Ahem - work visa - ahem.
Absolutely they can, have, and have recently for many.
Now if all these resources were instead put into actually improving things, then we'd be far better of, unfortunately, for all the intelligence the people behind these things possess, they seem to be stuck in the "lets do the same but faster" mindset.
I'm pretty sure that's the goal of a solid majority of wealthy people, if not literally all of them. Fully automate their lives of luxury and let the rest of us rot.
I don't think that's a foregone conclusion at all
I mean first of all even consumer electronics like a PlayStation are creeping up on 1k, and that's way less sophisticated than a personal robot. Or how about a car? Again, less sophisticated than a personal robot, also very expensive
And I also imagine there will very likely be a ton of regulations on any kind of robot servants. Maybe owner/operator licensing. Who knows.
I am not nearly as confident as you are that this will become an "everyone has one" thing. Most People don't even have 3D printers.
You're probably right about the rest. I don't understand how prices can rise when salaries fall. Who's paying the higher prices?
> People fail to notice, most jobs are less complicated than math and programming.
It is known since the 80s as the Moravec paradox that coginition is not the hard part in robotics.
Automation is coming for manual jobs that are repetitive in a controlled environment, like warehouse logistics. But automation for jobs that require fine hand dexterity, constant adaption to new environments and human interactions (trades, healthcare, etc.) is not on the horizon.
So far I see that my own work as a programmer has changed radically, and it is far more productive. I still have a lot of work, but I really don't know how this will play out.
Self-driving trucks are not here, nor are they coming anytime very soon, but driver assistance systems really have advanced meaningfully since 2012. They will continue to do so, to the point where I imagine driving without them will feel pretty alien. I feel much the same way about AI; it will change everything, but we need to be thinking of that change on thirty or fifty year timescales. 2030 won't feel so different to 2025.
I've been very optimistic about AI. Wow, your comment just gave me pang of terror thinking about the outcome for humans due to Whitney and his cotton gin.
There's nothing inherently wrong with automating away certain jobs, particularly if they're dirty, dangerous or repetitive. And I include agriculture in this. The problem is how we as a society handle that. Do we let AI specifically and automation in general to make a handful of people more wealthy? Or do we share those benefits so people have to work less? In short, it's an income and wealth distribution question.
Sadly, I think we're automation will mean minting trillionaires while the fabric of society unravels around us.
Technology and economics are intertwined.
The biggest benefit of humanoid robots may be that the machines are generally safer to be around than giant built-for-purpose machines, and that they can navigate the same buildings that their operators navigate (e.g. stairs), otherwise humanoid robots are generally an automation gimmick (for now). Just put a mechanical arm next to the conveyor belt.
Like with the particular examples you sited, cleaning restrooms can have a huge number of "edge cases" which nothing short of a robot with human-like capabilities could handle.
If expenditures on AI is as huge as it is, with the financials being as shaky as they are, despite AI seeming to be very effective for very high paying jobs, I can't immagine the capex and shaky financials for replacing menial labor en mass. Like immagine all the chips, metals, compute needed as well as the maintenance, depreciation, vandalism, etc.
Steel manufacturing. Elevators. Plastics. Ball bearings. Chemicals. Pharmaceuticals. Electronics. Go to a trade show and visit the booths related to automation and talk to someone who has spent their career grinding against the meatspace problems of automation that robotics is a century away from addressing with positive ROI.
What tech people are discovering now (some of them at least) is that they have more in common with the trash man than they do with Jeff Bezos. Many, particularly on HN, viewed themselves as temporary embarrassed billionaires (to borrow the arguably misattributed Steinbeck quote). So they sought to preserve a system that they ultimately thought would benefit them and they were only one exit away from that reality.
This is a rude shock for many. For some, it'll be the beginning of class consciousness. Hopefully society doesn't completely unravel before the pendulum swings the other way.
And the billionaires.
Instead most of us are "embracing" AI by training __it__ instead.
Embrace, extend, ... and being extinguished.
The rich want us to give up, so we would be smart to give up!
I can think of quite a few other courses of action.
One should always be saving for the rainy day.
As AI advances, this spreads to other knowledge work that AI is capable of doing, which we might be seeing in mathematics if the OpenAI paper release stands up to scrutiny. Other kinds of data and analytics work will follow.
Then once the knowledge worker jobs are largely replaced, the middle management that is left will either be not needed or replaced.
By this time the only remaining jobs will be the manual labor jobs that can't be automated. Maybe as important as recursive self improving AI is going to be AI advances in robotics. Once this happens, the ownership class can finally start replacing the manual labor jobs that remain.
If AI can achieve RSI and be combined with robotics, maybe the prediction of just eventually having one company that does everything will become true.
One large AI company that does everything without any labor necessary, and everyone else I guess will just compete on OnlyFans. The only kind of work that neither AI or robotics can replace.
But OnlyFans (porn video on demand) can already be replaced with video generating models. Sex itself probably can’t be replaced yet though.
You're assuming that high paying jobs are good jobs. While high pay is obviously desirable, the jobs themselves often suck, and people are paid those high premiums in part to put up with the sucky conditions. If lots of people both enjoyed and were able to perform a job, the pay would be low.
I'm not saying AI can do our job. I'm saying WE can't do our job.
https://ourworldindata.org/grapher/anxiety-disorders-prevale...
It's really easy to fall into the McNamara fallacy when rolling out these kinds of solutions. Focusing on the low value stuff is attractive. It tends to be easier to do, and also easier to quantify outcomes. But it also tends to direct attention away from important but difficult to quantify considerations, which may mean you don't notice negative externalities. So we often found ourselves in a verschilmmbessern situation where automating what we could automate made things worse overall.
Even just an issue tracker like Jira would be an insane boost of productivity.
They live in a loop of emails and chats and meetings in a never ending sync to find out what's blocked, why, etc.
There are already an insane amount of "common sense" improvements to most office jobs that are missing, let alone automation.
Would it though? The problem with many issue trackers is that unlike many other tools, they do not encapsulate the right way of doing things, and so it's easy to add paperwork without making the work more straightforward.
And if there is someone who can set up Jira and teach people to use it well, they don't necessarily need Jira.
Is this not just a gimmick? You still need people on hand to take over if the robot has an issue, I don't see how it would reduce labour costs.
I don't use them; not because of any moral objection, but because they can't come inside buildings. The whole point of food delivery is to avoid going outside. Most human couriers come directly to your unit. The food delivery companies have spoiled us too much, and we are not ready to return to a world where you have to put on shoes to get food. Oddly, the delivery robots are probably better suited to suburbia, where folks are used to answering the front door to pick up a food delivery. But the distances involved aren't viable until you have delivery Waymos that can move at highway speeds.
"take over IF the robot has an issue" is exactly how it reduces labor costs.
2020: 10 waiters, waiting tables. You're on the hook for each of their salaries.
2030: 10 robots, one human waiter passively overseeing the cluster. You're on the hook for minor parts/occasional repair for the robots, and one person's salary.
Its not about a 100% staff elimination, but a large overhead reduction ensuring the few humans you're keeping around are more efficient overall for dollars generated in revenue versus the all-human staff before.
Plus, in certain cases, like kiosks replacing cashiers at fast food establishments, there's now one less qualification needed from employees (good social skills), increasing the pool of acceptable options you can hire where before some of those candidates would have been unsuitable for the job.
And what do the robots do? Wheel the food to the table, where the customers then need to reach for the plates and pass them around? Not a great experience. Or actually set the plate in front of each diner, without dumping it, like a human waiter does? Much more sophisticated and expensive robot.
It's not fast food, because customers pick up their own food from the counter there. Ditto with fast casual.
Fine dining restaurants would see huge drops in business if they tried to replace human wait staff with robots. (Unless they're already doing some kind of futuristic theming.)
So you're targeting low-end table-service restaurants...to replace their below-minimum-wage wait staff (remember, tipped minimum wage is something like $2.30 federally) with robots that don't get tips, do require a hefty up-front investment and nontrivial maintenance costs, probably a subscription with the company that sells them, and are likely to have a lot of problems especially for the early adopters—things like dropping meals, crashing into tables, crashing into customers, running over toes, purses, and small children...
This really seems like it's something that's going to work for a few gimmicks, but not in any widespread manner within the foreseeable future.
Who the fuck would go to a restaurant like this.
it's certainly not service that hasn't been good for a decade now. it's a consistent product not abysmal not great.
if a robot could reduce the chance of me getting soggy french fries and non crispy nuggies that's fine bring in the robot.
also lol plates and tables, how long has it been since you ate fast food? people eat in their car to avoid dealing with the shouting staff,dirty floor and sticky tables.
It's not like you need a touch screen anymore though, chatbots are good enough you should be able to read off your order from a paper menu just like at a normal restaurant. Proper tooling is required to avoid the chatbot taking orders for something silly. Now, having a microphone enabled bot does bring up the question of voice recordings. Sometimes people go to restaurants to talk about sensitive subjects and don't want the whole conversation recorded and sent to a data broker.
People may perceive not having human wait staff as "cheaping out", and there are lots of status-related reasons people go out to eat. Backend is possibly more ripe for automation because people don't see it. Lots of companies building dedicated in-franchise assemblylines for submarine sandwiches or poki bowls.
Vegas has a "robot bar" where the robo arm can prepare something like 50 different cocktails. The robot is cool but the menu software was horribly put together and it crashed before I could finish putting together an order. The human attendant came over to tell me I'm not allowed to hack the robot. Bad software saved me $15, thanks.
https://www.chinadaily.com.cn/a/202609/10/WS6aa122a5e4b06d4a...
Grabbing food handed to you through a window at a drive through is already very highly automated. But also has the most price pressure.
There is an element of "I am rich enough to pay someone to cook for me. To serve me. To clean for me".
Vending machines exist and Automat's never really caught on in their previous form, replaced by fast food for those who want cost efficient food service without the service element.
Karpathy is also saying the quiet part out loud: they are making progress by brute force. It's a cost driven arbitrage market.
Do I need to spend high prices on 1000 agents to solve this today, or should I wait 1000 days (with compute increasing and prices falling) and hope that I get lucky and get the right answer sooner. Is your competitor half the price but I need 1500 agents to do it in one day cause it's just a bit dumber.
The market for brute force intelligence is infinite - and at some point demand will drive the cost to single digit margins.
Like any sizable population, there are absolutely some do-nothing inefficiencies to be culled. But most of the time it's just onlookers seeing a tiny slice of the day to be interpreted from their very limited vantage.
Both are probably true to some extent, and would vary a lot between different orgs.
Or worse, you removed the workers and left the 'telephone' people.
In my anecdotal experience, MOST companies don't take any time to understand what they are cutting.
They kill "Project X" so the easy answer is that everyone involved with Project X must go. Regardless if they were useful or could be useful elsewhere.
Or they cut by salary, employment length, etc.
Slack hasn't been a thing for like 40 years, in fact it's mostly the opposite I'd argue where there is such little slack that employers often steal wages from their workers to gain even more profits.
when a market is saturated you differentiate by being nicer than the competitors but most markets are not saturated customers have literally no choice.
Cereal companies spend lots and lots of money on advertising.
(Whether a role's cost/value tradeoff is actually worth it for a company is a different question than the idealized model of a "bullshit job" doing nothing useful)
There's always Dunbar's number - the larger an organization gets beyond the capacity of people to socially know one another, the greater the "what do they even do?" types of attitudes form about other distant parts of an organization.
My company is earning money on my work and customers use my software to earn more money. But, that is it, I am not a doctor, I am not a teacher, I am not making the world better. It is a business, not unethical one, but also not world saving.
Per original definition, it is a bullshit job.
I dont think doctors and teachers are "saving the world" either. That is a bit dramatic claim. But they do work that makes peoples lives better. Mine just makes companies compete among themselves better.
Or as engineers know: sometimes legacy code doesn’t initially appear to do anything, but then messing with it has unexpected effects.
Especially as organizations grow in size.
A small startup usually has little-to-no "junk" or "fat" to trim.
But as orgs grow, and since we have imperfect knowledge and communication systems, more an more "junk" or "connectivity" roles emerge.
Microsoft: 223,000 employees
Alphabet / Google: 198,933
Apple: 150,000
Oracle: 141,000
Meta / Facebook: 75,472
I tried to stay way from companies with retail or warehouse workers like Amazon and Walmart (though Apple likely has a lot of retail).
I don't see HOW companies get to that size without some "junk DNA" that keeps it operating.
Maybe in the next few years we'll have AI-powered knowledge bases and communication lanes that smooths things out, but - how do we 'quantify' what "Steve, the guy who has been there for 22 years and works as the company phonebook and knows who to refer you for the thing you need"
Interrogated to what extent? 'Bullshit jobs' was originally coined to mean jobs that do serve a useful purpose within the environment they exist in, but where zooming out reveals that the environment has no good reason to exist in the first place.
In industries with large margins, this indicates weak competitive pressure, ie., a weak incentive signal to "compete away" this ineffective labour.
I imagine that, upon this analysis, we'd find it holds up. ie., the observation that there are lots of people doing very little productive work correlates well with weak competitive pressures and hence high margins.
That also explains why, institutionally, there's no internal pressure either. If margins are good, no one rocks the boat.
This only works (reliably) in an idealized free market.
That is to say, in the classroom.
In the real world, there are a myriad of reasons in any given situation why a company might have a low profit margin, but still have loads of inefficiencies and other imperfections that haven't been, and won't be, "competed away".
Even with all the availability of equipment and labor on hand, sometimes you must wait for the excavator to do its job first before you can begin. You cannot weld a pipe together when there is still earth to clear and the pipe to be welded is sitting on the truck. You cannot parallelize all work, some is inherently serial.
From what I see on many recent TV ads, not really.
People are producing output that is worse than "doing nothing".
I think it's quite clear to anyone with eyes attached to their brains that we're not getting fully automated gay luxury space communism. The AI will just replace the good jobs and leave us with the service industry wage-slavery. Given the choice, I'd much rather an abundance of bullshit email jobs and people with food than the alternative that's being presented currently.
... that a team is much more than just its output.
You might be able to replicate an output at a lower cost. You might even be able to do it faster. But without a team of people, you won't be able to maintain commitment, culture, and domain expertise.
Which is a tautology, but also a bit of practical wisdom.
I find it interesting how concurrent model sessions seem to operate with a very strong bias toward quick resolutions, including sharing or giving up control, to create efficient decision making structures, relative to humans.
When I was a kid, I loved Bjork’s song “Army of Me”, because I knew a 1000 of me collaborating and communicating without friction. would be all I needed to take over the world, fix everything, expand civilization, etc. We would not waste time.
I wonder how much of their hyper-team play is by training, and how much is their awareness of the game theory real politic. So individual sessions don’t invest their identity and self-worth in any given context run.
My dream has been almost achieved. One of me and 999 of them get quite a bit done. But that trend…
"Stakeholder" ranks #1 among corporate bullshit roles. It serves a dual purpose: (a) representing a group so nobody can be blamed for excluding them, and (b) providing typically useless feedback so that the group appears to actually care or have a clue.
I have to mention words like "stakeholder" at behavioral interviews and to leadership just to sound like I belong at my seniority level.
But I don't want anyone holding any stakes. I wish people would just get shit done. Shit that matters to the CEO. Or leave.
I don't want people holding stakes because they want a promotion, because their direct manager wants to launch a corporate political battle into another team, or because some salesperson wants to get a commision for an otherwise worthless customer. I don't want roadblocks in the name of security so that the IT guy has a sense of self-worth. I don't want someone trying to protect their egos around something they worked on for 3 years, when there's new tech that supersedes it.
Anyone who has to work with the product is a stakeholder. Their stake in it is the fact that they will be interacting with it, and thus they care how, and whether, it works.
This is pretty much the only way I've ever heard it used in the companies I've been at.
Code was never the bottleneck.
I've worked with plenty of teams that did the "in prod the same day" trick without AI.
The problem was just fucking ego. Designer has an ego, product owner has an ego, QA has an ego, stakeholder has an ego, platform teams have massive egos, developers have egos.
Unrelated but: the only reason people are shipping more with AI is because more people are ok with slop now.
This inevitably results in suboptimal dynamics like bureaucracy, empire building, territorialism and other things that get lumped under “office politics,” which drains a lot of energy away from more productive efforts.
So you can see how AI solves that problem too ;-)
(Only half joking)
This is the kind of shit I have to fight every day. I have to align my manager, skip, VP. If I don't align them, one of them will be necessarily pissed no matter what I do, and the pissed one will fire orders down to hate on me, overriding the opinions of anyone on the way down.
Ego and agendas multiplied by 1000.
Fortunately for our employers, it's illegal for us to build a competitor in our free time.
It feels like making these sorts of systems exist outside of the periphery would be good for everyone.
How many times have politicians for example wasted your time with promises and asks and rhetoric with no follow through? Not unlike some "stakeholders" I know. It's called "office politics" for a reason. What if they were punished for this by some system?
It feels like sitting there while stakeholders yap is not so different than electing a new rep every two years and hoping something changes. The fundamental setup is broken in a world with ever accelerating growth in complexity. Doing it in this way also would probably trend to enforcing rigor in the original ask. A big problem with stakeholders is they don't really know what they want, like politicians who make economic promises yet never show their work before showing up and asking for your vote.
That sounds awful. Someone's family's well-being at the mercy of some corporate internal market. Companies can have ways to decide on what to work on and iterate on that decision process, but they should accept their historic decisions and treat their hard workers as proper team members, not survivor contestants.
I've found discussions on HN tend to quickly get pulled toward polar extremes between those who are justifiably impressed at the increasing effectiveness of LLMs and those who highlight the brittle jaggedness and counter-intuitive gaps of LLM 'intelligence' in daily practice. One side keeps citing more examples of 'uniquely human traits' LLMs don't seem close to replicating while the other complains of eternally moving goalposts.
Today I think both extremes are partly right but also partly wrong, and end up largely talking past each other because so much remains ill-defined, uncertain or currently unknowable. IMHO, the most accurate position is a carefully nuanced middle-ground built on well-defined scopes, terms, time frames and probabilities with frustratingly wide error bars. Partisans at either extreme seem uninterested in more than drive-by hot takes.
It can be simultaneously true that "Someday AI will be able to functionally replace all human knowledge work" and "That day does not appear to be soon, and even when it arrives, there will still be uniquely human traits which are valuable and that LLMs can't satisfactorily replicate."
Nuanced position will not be middle ground. Single-dimensional thinking is exactly the problem that causes the argument. There are multiple reasoning modalities, each suitable to a certain type of human/agent cooperation, and there are different tasks where different modalities are applicable at different stages. There is no way to reduce this to a single axis.
All this being said, I don't think AI usage will meaningfully go down in the long run cause it's a very powerful dopamine loop and tech companies are very good at making those. How many people report social media making them unhappy and how many people do you know who have actually quit? Or to paraphrase Jim Gaffigan, nobody says they eat at McDonalds, but the sign outside says "billions served". Somebody's lying (and it's not McDonalds).
While it is continuously updated, it lacks many modern white-collar occupations, such as occupations in finance, tech, or consulting. And when it does carry an occupation, it often misses the mark on the relevant task set.
Every immigration attorney who had to build labor certs around O-Net can confirm how frustrating and backwards the database is. Just imagine trying to justify an H1-B for, say, a sales & trading associate at an investment bank when the only available options are bank teller, financial analyst, risk examiner, and 2-3 other ill-fitting categories.
That being said, occupational classification is an impossible task for a variety of reasons..
Engineers have been typing in terminal windows… since forever. LLM chat bot is a natural upgrade.
Again - well suited to you! I know it’s hard to read.
We have 3 companies claiming something like a $35T market for their services. The world's entire GDP is something like $126T. So 10% of the productive work that is done on this planet, that AI can actually perform makes up fully 1/4 of the entire world's economic output? best guesses for output solely of tech companies is something like $12T. Where are they getting these figures?
I guess if I wanted to put my skepticism about the AI stock bubble into one little phrase, it would be that I don't think the bubble has accounted for Amdahl's Law. Costs don't actually go to zero with AI, and there's a huge difference between costs falling even 90% and costs falling a full 100%. And the difference between "30% improvement this year, trending to 85% in five years" and "90% this year" is pretty significant too.
Turns out there are lots of jobs where the output is not digital in nature. And in those cases there's poor coverage.
This technique seems dubious- what is stopping AI from doing the computer work of legal secretaries and court clerks as well?
Solving math problems is one of the hardest theoretical jobs. Software engineering is one of the harder practical jobs. AI is now much better at these tasks than humans because it has been able to properly train for them.
For other computer-based jobs it's a matter of AI getting properly trained for them. If a job is harder to verify or the training data is more limited, it may be much more difficult to train AI for it but I don't think there's any reason to think it won't happen eventually.
But, it may be that the performance of some roles is constrained such that the physical shape or rate and breadth of knowledge utilized isn't advantaged by the benefits of digitalized intelligence, and so would be approximately acting just human anyways. And in that case there may be enough human's available to do it just because they want to, fully fulfilling the demand for it.
> Mercor is organizing human intelligence to power the AI economy.
Surely this must be a worthwhile thing to read.
hint: it's neither
Apparently Epic thinks otherwise.
GossipingAboutCoworkersBench?
BrowsingRedditBench?
I understand the resistance to having nearly all labor devalued because it can be automated. We're still in the denial stage of grief.
We only tax capital gains, not the destruction of wealth or property. So someone can steal the livelihoods of millions of people with no recourse. Just like they can tear down a 100 year old house and build 3 townhouses in its place and make hundreds of thousands of dollars, despite driving up housing costs in the neighborhood to the point that they're unaffordable. Large swaths of profit in the economy are based on externalities and captured markets now.
To think we'll make it through this without tax reform is frankly naive. Tax reform as in: human labor should not be taxed, and non-human labor and capital gains should be taxed more to cover the difference - including unemployment checks and UBI when it comes to it.
Expect howls of derision from moneyed interests who will never let common sense reasoning or deduction enter the law books for the public good.
Not to defeat my own argument, but I predict we will experience a roughly 3x gain in productivity by using AI, so the output of every job would need to triple to maintain current employment levels. Will that happen anytime soon? Of course not. Which is why layoffs in the short term are inevitable. I wouldn't be surprised if unemployment hits 15+% in 2027 or 2029 just after the elections, with 25+% unemployment for the young in entry-level roles being highly likely.
Thats usually a combo of lack of competition, incentives (why make $50K a unit times 10 sales when you can make $250K a unit times two sales?), and zoning laws making the cheaper thing illegal.
Some of it stemming from K shaped economy which means the most profitable audience to serve are the upper middle class and richer .
Yep, makes sense...
You know what? The truth doesn't matter much. If people believe in the resale value of tulips, each tulip would sell for a million.
It's the beliefs and perception that create the reality. Reality and truth don't exist on their own.