> I foresee a wave of entrepreneurship coming. AI will empower more people to provide useful services directly to other people, with less middlemen and menial work, and more direct problem solving.
All this rambling with citations but no coherent original thought weaving several conflicting ideas together. Illustrative example: AI is simultaneously going to take all the jobs yet it sucks and nobody is getting any real productivity out of it yet these evil CEOs will somehow manage to push "adequate AI" to disrupt everything to justify their unrealistic valuations yet respected economists predict only 0.X% impact over the next years etc etc...
It's just a grab bag of arguments thrown into one long rant, and as some comments here point out, some very obvious contingencies (which even have supporting data now) have not been considered.
All this and not a single mention of the very obvious thesis that the real problem is Capitalism.
Seriously, go through each point and see how it can be traced to Capitalism. Why do these CEOs keep telling everybody there will be social upheaval and yet continue building this technology? Because Capitalism demands infinite growth in profits, and if they don't build it someone will! Why do people feel lost without a job? Because Capitalism tells them they are worthless without any "productivity" (Hmm, wonder why "starving artists" were a thing. Maybe that's also why they were so derided: the system didn't like people who found self-worth independent of economic "productivity.")
Now, Capitalism is the problem, but also potentially the solution. Here's a dynamic that doesn't seem to have been covered in these discussions:
1. The majority of economic activity is generated in supporting the economy itself. We know what happens when everyone loses a job: the economy shrinks drastically (the "Bust.") TFA claims "plenty will still happen" and "GDP might even grow" but handwaves away what that would be.
2. The wealthiest are so wealthy because they essentially take a cut of the economy. So their wealth depends on the economy growing. They are ruthless capitalists and don't care about us, but Capitalism demands growth and they are monomaniacally driven by it.
3. As such their interests are aligned in keeping the economy growing and labor compensated and spending. Yes, they have their doomsday bunkers just in case, but none of them wants that life (where's the growth?!?) and they know it.
Add to that governments, all influences considered, are also overwhelmingly invested in keeping the population happy.
I don't know what the solution will look like, but all the parties involved -- the people, the elites, and the governments -- are heavily incentivized to keep things stable. Something will get worked out.
Funnily enough, my LinkedIn feed is full of managers who are ecstatic at being able to "code" again, so it applies to developers trapped in the role of a manager as well!
Great read. While I've not worked at a big company in a couple of years, this resonates with many things that I imagined is happening inside large, tech-savvy orgs, reinforced by observations of the industry and anecdotes from insiders.
If you read between the lines -- and note, the author himself probably did not intend that subtext, or was careful to avoid implying it -- you will realize that the likely natural equilibrium involves way, way less people.
I wrote in more detail in this comment: https://news.ycombinator.com/item?id=48040999 -- but the upshot is that in the future teams, or heck entire departments, will be replaced by 1 - 3 very senior ICs that do both strategic and hands-on work.
The driving force for that structure is the high coordination costs or large org, which TFA bemoans quite a few times but does not explicitly call out as THE main issue. I mean, if three teams can provide working solutions in record time, your problem is not aligning them, your problem is that you have two teams too many.
On the bright side, fewer people also means fewer organizational pathologies like bureaucracy, silos, duplication of work, turf-wars, empire building and politics in general. The downside is well, a jobacalypse, starting with juniors (future talent pipeline problem says what?) and middle-management.
This does assume very capable ICs who can do strategic as well as tactical hands-on work. This is where future job seekers want to be. They will be worth their weight in gold. Which will still be wayyyyyy less than the value they provide (or crudely, the salaries of the teams they replace.) They'll do 5x the work but won't get paid 5x as much, of course. what do you think this is, socialism?
I'm probably being overly charitable, but I would have said the same thing except in a positive context, because I've wanted to do the same thing at times.
I think it's a matter of perception because I didn't interpret any of them as being gleeful about it. If you think about it, "AI will take your jobs and maybe destroy the world" is horrible, horrible marketing -- like, your comment is a perfect illustration of how it is received everywhere -- and yet these CEBros can't stop saying it, which indicates to me that they actually believe it.
Oh, now that their IPOs are nigh they're changing their tunes (https://archive.md/s9EO3) but to me that looks more like they've decided to let $$$ prospects override what they really think.
The general public is not their customer base, they don't have nearly enough money to spend on AI. Going to the media and saying "this product can automate so many jobs" is marketing to other businesses who want to use it to cut their workers out.
There was crazy clip of Eric from Google telling a crowd of university students that in the future AI will do everything, and after the whole audience boos him he keeps pushing the point that they better accept it and get on board. The mentality these guys have is sickening. They have no humility and no humanity.
The general public may not be their customer base (except maybe for ChatGPT, which is primarily a consumer app), but it is the voter base. The AI backlash has been brewing for a while and is bubbling over in the form of data-center pushbacks, and talk about regulation has been picking up. Plus, if "AI destroys the world" does happen, even the capitalists looking to further cut out labor will not be too happy about that.
Even if this was not covered in Marketing 101, it was all pretty predictable. Sure, most of these CEBrOs probably have a god complex (probably fueled by Ketamine) but their behavior is also consistent with the premise that they see a job apocalypse coming and they must warn the world about it.
Especially never liked Eric Schmidt, and he went about it very ham-handedly, but I do think he is right. Stopping is not an option given Capitalism's hunger for growth and the current geopolitical landscape. The genie is out of the bottle and we must adapt, because Capitalism is not going to.
Knowledge worker compensation is 35 - 50 trillion a year globally (6 - 12T in the US alone.) That's a huge TAM. It's still close but 5T over 5 years seems doable.
>... unless we figure out how to make developers 2x, 5x, 10x as productive on stuff that matters, this isn't going to play out well.
The way we make ICs 10x productive is not just making each of them individually more productive, but by removing the coordination overhead of large organizations, because overhead scales super-linearly with the size of the org. And orgs will shrink automatically as AI-assisted ICs take ownership of larger and larger scopes of work, leaving much more budget for tokens.
Yes, but I do worry about junior knowledge worker job loss. These models are very good (and getting better) at the vast dark matter of "donkey work" that happens in knowledge-based industries -- work typically done by junior devs / analysts / lawyers / consultants, paralegals, admin assistants, customer success / support, etc. -- and those roles comprise the bulk of the workforce.
And worse, these are the tasks that help the junior people eventually grow into the skilled knowledge workers required to operate models, so there's a pipeline problem too.
I do too, but I think it currently has a lot more to do with the quasi-recession we've been in since the end of ZIRP and AI is a better excuse to stop training juniors than telling investors it's belt tightening, just like layoffs.
I'm already seeing tech execs/hiring managers getting very frustrated at the lack of new-senior-engineers to hire. The market will correct for this in time.
Curious if you can share any backing information from your last statement? As a senior engineer (well, that's my job title anyway), I find it encouraging.
This doesn't break it down by experience, and I can't find specific data on that, but the recent spike in demand for engineers + subsequent drop in unemployment this year is well documented [1].
The demand for senior+ engineers has remained steadier through this downturn from my anecdotal observations, with new grads being by far the most negatively affected, but even that seems to both be shifting from talking to people a handful of years younger than me + CS enrollment has already precipitously declined [2] as the narrative that programming is dead because of AI has spread rapidly.
All that leads me to think it's going to be a junk-show over the next decade for people trying to hire as the pipeline was destroyed.
However, they have reported numbers along rather inconsistent dimensions. Like, historically they've focused on number of repos and users and later PR's and issues, and often catch-all terms like "contributions" which includes all of those + comments etc... but the number of commits alone (which apparently is the main culprit now?) has been mentioned very sporadically. This has made it hard to get a consistent sense of historical growth.
Without any other information, however, it is reasonable to assume that a 14x in commits is the prime candidate for instability. Especially since commits are write traffic, which is much harder to scale than read traffic. Plus every 3 - 5x increase in scale can reveal bottlenecks in your distributed systems that you never knew existed, so they probably have like 2 - 3 "generations" of bottlenecks to figure out!
Actually, you're both right. Using AI as a supplementary learning aid -- i.e. students use AI as a personalized tutor but still do the assignments themselves -- produces better outcomes. But using AI as a crutch -- i.e. using it to do the assignments -- produces worse outcomes.
> students use AI as a personalized tutor but still do the assignments themselves.
So your first study actually concludes the opposite. It concluded that all AI users performed worse, but the effect was smaller for students which used AI as a tutor.
The second meta analysis I don‘t quite understand. I understand they conclude that using AI tutor shows significant improvement, but I don‘t understand the methodology. I may be misunderstanding but it seems to simply count papers which shows positive outcomes and reaches conclusion that way. I think that methodology is deeply flawed as it will amplify whichever biases are present in the studies it uses. I also think the lack of control groups is a major issues. If we are comparing AI tutor to nothing, off course the AI tutor is gonna perform better. We need to compare to traditional methods. And this is especially relevant in our discussion because junior developers usually have excellent access to senior developers (via peer review, pair programing, etc.), much better then student’s access to tutors for that matter.
> The results indicated that students employing AI tutors shown significant improvements in problem-solving and personalized learning compared to the control group.
Now when I look at the control group it claims this (also in the abstract):
> Participants were allocated to a control group receiving conventional training and an experimental group utilizing AI technology,
But when I look into the methodology section I see this:
> The researchers classified the patients into two groups: MathGPT and Flexi 2.0
MathGPT and Flexi 2.0 are both AI tutors. Now I am confused, where is the control group and how was this “conventional training conducted”?
The methodology section actually tells a different story from the abstract:
> This research utilized a quantitative methodology via a quasi-experimental design.
By quasi-experimental design they mean that they tested the same students before and after AI intervention. And concluded that the AI tutor helped them improve. Now this is not what control group means, so the researchers are actually lying by omission in the abstract. This is a spectacularly bad experimental design and I wonder how it would pass peer review, so I look at the publisher Jurnal Ilmiah Ilmu Terapan Universitas Jambi. So not exactly a reputable journal.
I still stand by my no evidence for a testable hypotheses. I suspect that your first link is actually correct in that AI is bad for students and just less bad if it is used as a tutor.
I hadn't looked at that study you selected, but yeah the methodology conflicts with the abstract (Also it low-key seems to be an ad for "Flexi 2.0.") It does seems to be a shady paper, with a small N and in a journal of questionable repute.
That said, there are 80+ other studies listed in the meta-study, which is pretty frank about its limitations. (Note the snippets about positive biases in the conclusion.) It is going more for quantity over quality and is transparent about the statistical findings of each one (or lack thereof; see the count of "Not reported"s.) All these references have a myriad of results, but across the spectrum of well-designed studies at reputable venues to the other end, they follow the same themes, so I don't think this can be dismissed that easily.
This was the only study from the meta analysis that I read, and I picked it because it made the strongest claim out of all of them.
This is in the opening of the results section in the meta-analysis:
> In the final screening phase, a rigorous full-text analysis evaluated the methodological robustness and empirical validity of the remaining studies. [...] The final corpus comprised 88 studies that demonstrated robust empirical evidence for LLM applications in educational contexts.
The inclusion of the study I read does not give me confidence that this statement is true. And the fact that they reach their conclusion by simply tallying up the positive vs. negative studies makes me conclude that this meta-analysis is practically useless. They do admit this in the conclusion (which is probably why it passed peer review [assuming the peer reviewer didn’t read the same citation as me as I am 100% certain they would have asked for it to be excluded]). But that pretty much just leaves us with nothing. We are exactly where we started. No evidence that LLMs help students beyond traditional methods.
Now I am not gonna read that Anthropic study. It reminds me of Cigarette companies finding the health benefits of cigarettes. That leaves that excellent 3-study review. In their first study they found LLM has negative effects on students (in line with the first link you showed me). In the second study they found no effect. And in the third study they found mixed (nuanced) effect where using LLMs as tutor helped students in one aspect but had negative effects on others. This is by far the best study you have presented me but it still does not change my opinion. There is little evidence that LLMs (even when used as a tutor) help people learn better traditional methods.
What makes me even more against this sentiment is this quote from the conclusion of the 3-study review paper:
> Our results suggest that students prefer to use LLMs to substitute rather than complement learning activities.
So on their own, students are more likely to use LLMs in a way which is harmful to their learning. I would expect similar behavior of junior developers.
Yes, and there is already preliminary data showing this trend: https://news.linkedin.com/2025/breaking-the-trend--small-bus...
reply