Erik Brynjolfsson says an AI "job apocalypse" is unlikely
wpintelligence.washingtonpost.com
wpintelligence.washingtonpost.com
The job disruption is real and it will be an job apocalypse for plenty of people.
If a company is now paying for ai tokens and you now have 5-20% higher cost but you are relevant more efficient, you will reduce your team by 5-20%.
It is already a job apocalypse for people who learned language and no longer get a job for translating or text analysis. People who want to be an illustrator have a job apocalpyse.
And we have not even SEEN what it will look like in 4-14 years. Physical AI / Robotic is progressing like crazy too.
I routinely talk to people at my current non-tech job who don't even know what Claude is.
Technologist talk about tokenmaxxing while I can't imagine more than 10% of the non-tech workforce could even explain what a token is.
This level of automation is going to have an extremely slow roll out because of regulatory and sociological reasons.
There is no equivalent to the contagion of AI coding agents outside of programming. I am still fighting just to get access to Claude Cowork at work. I have quite a few automation ideas but there is not much in it for my manager. They have to sign off on something they don't understand with the best outcome their department shrinks. I can understand from their perspective, the status quo is perfectly fine. The risk is long term to the whole business not being competitive but neither my manager or I will be at this company by the time that materializes.
There is this cultural sorting function too that if you don't like change you probably don't work in tech. The culture that creates is night and day different to the non-technology workforce culture. Most people outside technology hate change. Time and time again, I have seen that even the most simple and obvious changes don't get made, even if it makes the task easier because of this bias against change.
Agentic AI Layer: moving more knowledge into the software stack itself will reduce the amount of people you need for 'software'.
And the second thing: We are now in a time of commoditisation: Everything which isn't optimized becomes optimized. That small company not being efficient at all? Another young person with automatisation / ai in mind will try to disrupt that market because we are so saturated.
With white collar job people getting dispersed, will lead to more pressure on blue collar
We truly do have distorted view on anything outside our bubble.
The reality is, hiring severely overshot in 2021-22 with a degradation in quality. And 2-3 years of remote first showed that you can offshore with little-to-no operational impact.
Dario's messaging made sense as the underdog - which Anthropic was until a year ago. Their moat used to be FedRAMP until Emil Michael called their bluff.
What firms are going to do now is probe deeper and disinvest in crappy projects that - many that most likely generated negative returns upon closer inspection.
We've been doing this since 2022. Most layoffs overlap with Sales, Engineering, and Ops associated with cost centers.
Like, there was no reason Uber had 3 internally competing and built chat apps with dozens of engineers allocated as headcount.
> much of the cash flows are coming from investments made yonks ago
Not necessarily, but you can't train an old dog to do new tricks.
Yes they are.
The marginal ROIC for tech firms that are in mature growth is shit. Made shitter by continuing to field many people for no good reason.
And marginal ROIC is just one metric amongst multiple to benchmark against.
Also, if you had the Benoit [0] you should use your normal HN account instead of a throwaway created for this topic.
As a side note, if you think we haven’t been doing this for quite some time already, you may want to look into this topic a bit more!
Also Brynjolfsson: It will likely impact entry-level jobs. There are huge implications for business and the economy, including some increases in living standards, but also, potentially, job disruption. We must act now to address this gap. We can’t just sit back and wait for the tsunami to hit us.
Also Brynjolfsson: If companies don’t hire people at the base of the pyramid, then they’re not going to have those people later when they need them. Now, AI can do a lot of those basic tasks. Some companies, such as law firms and investment banks, are not hiring as many people at the base. In the short run, I can see that logic, but, in a few years, where are they going to have the most experienced people? So they need to think harder about that.
So there's that. Basically what he's hoping is if everyone does what he recommends re: regulation and policy, then maybe the job apocalypse will become unlikely.
LLMs didn't really change anything about that but people like to give it credit for stuff.
In the long run it meant seniors commanded high salaries. It's the tragedy of the commons i guess except where the apocalypse is you need to build a free sushi restaurant and on site laundry for your spoiled developers.
Imagine when jobs started adding the requirement entry level positions require ability to read? That cut out so many people when that requirement was added yet the economy survived
You used to have a career as a software engineer or devops person if you were good at debugging. Today, I challenge you to take any past production outages, toss the traces into Claude Fable, and see if it finds the bug faster than humans can.
I think this is the way the rest of white collar labor goes.
I'd love to see one someday.
It’s great at being a search engine for our docs and communications. It can also find bugs quickly in small repos with static analysis and coverage tools.
But when it comes to larger projects or interconnected systems, the time it takes to correct its “assumptions” is often longer than the time it takes to solve the entire problem myself.
(There have been a few times where it has refused to debug due to safeguards. This is actually the biggest problem.)
And yes, before anyone brings it up, I've read Manna and if you haven't you probably should[0] but I don't think it will be as dystopian or utopian as that. Humanoid robots would obviously replace some of the proxy need but there will still be a lot of things that only a living human can do even if it's just purely for legal and regulatory reasons.
I also think there will be a lot more roles that are purely human in nature and which will have an entirely different set of requirements but that's a different topic.
[0] https://web.archive.org/web/20120224063109/http://marshallbr...
If the number of jobs stays constant but a lot of those jobs have been downgraded to fungible operators of AI, that's very bad.
Smells of hypocricy. And if the world didn't really change for gig workers before, why would it now?
All well and good, but the age old question. How will people get experience if entry level is AI ? But maybe for a preview look at COBOL, there is huge demand for experience people.
Aka, pushing training costs onto labor instead of capital.
We’ve been doing this in an accelerated fashion for decades and this is the next step on that treadmill.
US companies used to hire at entry level and train employees in the high level skills they needed.
You’ll find many people wailing and gnashing their teeth about how those entry levels will jump ship as soon as they get trained. Nothing is stopping them from paying market rate for those newly trained people.
Nothing is also stopping them from making a contract that you have to work for them for X years post training or return the training costs other than the allergy US firms have with making any sort of contract with labor that doesn’t let them fire them at a moments notice.
Is there? I actually looked into this once, and nearly all the Cobol work I saw mentioned was being outsourced to software consulting companies in India. And it didn't sound like an especially good job there, either.
If anyone is genuinely paying disgusting amounts of money for deep Cobol expertise, I'd love to see any public evidence that this is happening on any significant scale!
And maybe due to falling population by then whoever left would have something useful to do nonetheless.
General society: [notices this is what AI companies are explicitly trying to do.]
AI companies: don't worry about it bro, it'll all be fine -- better than fine in fact. And not just for us, promise! Actually we'll somehow create even more jobs, while still being able to massively increase productivity, even though this is contradictory and impossible. Because nothing is impossible with AI!
And this will continue to go on and on and on. This means that this idea of large scale AI lay offs is complete delusion.
There is all this accumulated, inefficient structure we are maintaining.
The next recession will have a strong amount of economic autophagy because of the amount of structure to eat through.
I suspect this is how automation materializes too at scale. It materializes in the leanness of companies on the other side of a recession. We are viewing LLMs as not producing this autophagy as proof of something not working. I don't think this is reality.
I think it is related to this childish view of instantaneous market efficiency. A recession is what will perturb the system and start the automation chain reaction. Inside the current economic regime, automation is just a risk and a cost because the tools were derived inside this long expansive regime. Nothing much happens until the shock and regime switch. Then we never come back to this regime.
I woke up everyday thinking how someday I will look back on today with such fond memories of being the good old days. The regime switching is going to be quite miserable.
The Work from Home programmers surely aren't, since they've been displaced.