The death of web development education
molily.de
molily.de
Over the years I have enjoyed many deep and interesting technical discussions online - both here on HN and elsewhere. I can't help feeling that those discussions have also all but disappeared over the past few months - alongside high quality free material that some experts have kindly given away but increasingly no longer do. I rarely find myself contributing any more because there is little on any of my usual online hangouts that is interesting enough to spark a deep technical discussion any more. It's 90% AI stuff - either written about AI or (usually poorly) written by AI - and most of the discussions that do initially catch my interest at all seem to end being very negative and depressing. (I appreciate the irony that I'm perpetuating that cycle right now. Looking through my relatively sparse recent contributions to HN is actually quite depressing. Although at least I actually am getting out more!)
I wouldn't mind so much if I believed the hype about the benefits of AI-generated work and thought that overall this period would be a net win for humanity. Unfortunately I don't really believe that. I fear the LLM era will eventually be remembered for producing tools that could do average/routine tasks quickly and to an acceptable standard but that were actually used to (try to) do good/innovative tasks with substandard results. But by the time enough decision-makers realise this it might be too late and the damage might already be done and hard to reverse.
Funny how our short feedback loops and greed pushes people towards deeply unstable systems, especially if the price seems so "cheap" for a reward sold as so incredible that it would be dumb not to bet on it...
Altman, Musk et al are making a play for being the default epistemic authority of society.
> The machines will do the knowing and caring for them
I don’t think it’s “the machines”, it’s the big 2-3 corporations the ones doing the thinking for us (aka controlling). These AI agents are whatever these corporations want them to be and will configure them as they see fit. We are outsourcing intelligence to a few companies.
The corporations ever growing size is what scares me the most. Why care when you are a replaceable cog in a massive sprawling machine? Why care when your effort is not particularly seen or rewarded?
People's jobs are no longer something that connects us to other people around us, or the world at large. And these massive orgs are caring less and less about their people.
I do think AI compounds this. Yes. But also, what parts of the world are actively seeking involvement, are dynamic places with apparent rife potential? The world just doesn't care, doesn't have the engagement points it did.
Tech in particular has really lost its luster as a dynamic place. Tech has gotten much more mature, much more industrial scale & less new & creative.
And then: the titans have consistently eaten all the younger children. The big companies swallow up all the interesting new places. The monoculture of these ultra massive ultra wealthy titans, who've been doing the same intermediation enshittification plays for decades now, is not exciting, is not a vibrant thing: it's tiresome to the world to endure these titans, rather than interesting and thrilling a scene to be a part of.
Just as the production and service bureaucracy of the USSR ceased to care about the value of their work (so long as they technically fulfilled their plan), we are possibly headed for a world where all the "real economy" workers (in manufacturing, logistics, etc.) will be guided by the output knowledge workers and managers driving AI, who are being judged by yet more management using AI, who are ultimately judged by investors using AI and the whole thing gets disconnected from any actual ground truth -- just as planning did in the Soviet case.
When the whole thing collapses, somebody will be around to pick up the pieces (and valuable capital equipment, land, etc.) for pennies on the dollar. I suspect the kernel of whatever comes next will come from China, but also suspect it will bear more than a passing resemblance to Putin's Russia -- a small clique of rentier-gangster kingpins semi-legally seizing everything in sight and cobbling together an autocratic oligarchy that creaks along without meaningful challenge.
We were already on track for that anyway. This just adds an interesting twist to it.
But unlike the author of this post (or cited authors) I'm not sore. AI is providing a much better education model for students. So it's us that we have to adapt, not complain that the world is changing.
For example, in Machine Learning our projects are sometimes extremely long and repetitive. Because before AI, it was a valuable skill to be able to recall APIs quickly (what is the method to drop duplicates in a dataframe again?).
But today that's no longer the truth, and for the better. I am happy students don't have to go over the same repetitive low-value processes that I had to go to. Students can now focus on more high level valuable tasks: understand the domain, compare evaluation methods, compare models, balance tradeoffs, present solutions/reports, etc.
It's like in the 50s when students had to memorize tables of logarithms. Nobody misses that.
We are in the process of recreating all our projects to adapt to the new world. We'll see how it goes.
LLms dont help here either.
Part of learning is about struggle. Seems many here want to disregard that fact and just run to the LLM lmao. Yeah you can do that - the knowledge won't really stick.
The initial creativity is still with us, not the LLM. It's the Art/Craft thing - we have the Art, the machine does more of the Craft (and it always did do a fair amount of the Craft).
I'm seeing folks who couldn't (for whatever reason) work out how to code, starting to produce some amazing stuff with the LLM doing the coding now. It's their creativity, their project, their vision, still. The fact that they didn't write the code is meaningless.
From what I’ve seen, the ability to prompt quality output with AI is entirely proportional to experience and skill level.
Maybe when it comes to any coding-related learning, I can agree with that point.
However I think the future of education in general is very bleak. When we have plummeting scores for the PISA exam across the world [1], I fear people will get dumber and dumber.
There will always be those who thrive, but the easier it is to obtain information/knowledge, the less I think people have an incentive to do it. And I don't buy that AI will solve all intellectual jobs anyway.
[1] https://www.dw.com/en/pisa-2026-results-reading-math-scores/...
I'm sure AI is going to do even more damage but COVID was an important factor here as well.
Education itself has lost access to the tools that were making it cheaper and easier to teach.
Learning something is a side effect of overcoming friction. AI removes friction.
I’m using ChatGPT to learn German. I can immediately adjust its responses and language level(it says I’m at A2/B1) and tell it how I want to study.
I find it to be a far superior tool than anything else I’ve used.
If education is all dependent on the use of AI, then we're cutting off everyone but the most wealthy from access to education. Seems very regressive to me.
1. This is an assumption but I think it's fairly well established that current LLM prices are far below operating costs of inference, so prices will go up, maybe not 10x but some multiple > 1
There’s a lot of individual anecdotal responses that people come up with in these threads, and I’m happy for the positive examples.
I do think people are capable of extrapolating, especially since all of us have the experience of seeing students wanting to escape studies as kids.
It’s identical to the gym problem, everyone wants to be healthy and fit, but the effort sucks and we avoid it.
Education at scale is the same issue. We basically nurture/force crops of students over a 20 year time span.
The people who wanted and could learn by themselves always had the possibility to do it with books, audio and videos, and AI just add a layer of interactivity on top. Self-learning never replaced standard education, even in the high bourgeois class because they answer to different problematics.
ChatGPT might work for you learning German because you have a reason to care about learning it, but for most kids the goal of school is getting good grades, not learning.
This is exactly what AI companies need, so that's development should be celebrated. There are two paths to capture the value that's currently wasted on paying people:
1. Build a system that's superior to nearly all people.
2. Make people worse, so even an inferior system is superior to them.
Getting AI into the classrooms and degrading the educational system are important for path #2.
It's good not to put all your eggs in one basket, so I'm glad the AI industry is pursuing a multi-pronged strategy against their competitors. Who cares about the waste and destruction for humanity, as long the investors get good returns?
Learning to write will either help all the humans by extending their capabilities, or will make them stupid -and dependent on writing.
Socrates: It will definitely make them stupid, and I am suspicious of this new witchcraft.
Maybe he was right, but I’m kind of enjoying having both writing and AI around.
We end up with shit results and where they arent shit students are rushed though as if to prevent them from thinking about the materials.
We have libraries full of unused education formulas. Some 2000 years old. Then there is nutrition, sleep and other work/rest ratios, physical health.
How expect results without learning to teach and learning to learn?
Onto what should one anchor learnings if the suroundings are always the same?
I never hear a word about spaced repetition in school. Its so basic i must conclude they knew nothing about what should have been their area of expertise.
When things get bad enough we will [finally] have reason to improve.
Maybe the great work of Salman Khan is merely a hint?
MOOCs were something I personally hoped were the solution to scaling education. However it didn’t work out, with courses having single digit completion rates.
This is for free, high quality, immaculate pedigree, learning material available at any time and any place.
Years of additional research had to be done to figure out how to raise completion rates, which creates more complexity on systems and delivery.
Unfortunately, that improvement is now moot, because LLMs obviate the friction and effort.
The point of doing a homework exercise is to provide learning friction for students.
LLMs obviate friction.
I would add Bloom’s 2 sigma problem to the mix.
COVID generation.
The pandemic was barely 3 years ago, and stretched over 1-3 years depending on where you lived.
It will likely have a material impact on most developmental statistics for another decade.
It's a real shame a leader of an education company doesn't value the actual learning process. Those "low value" processes are the building blocks to understand and master high level concepts. I'm not spending my day to day thinking about the low level memory allocation in my job, but that knowledge and work helps immensely in how I approach the higher level assignments and wider architecture. And I still kick myself because my assembly knowledge is is only 1.5 semester's worth.
And yes, I still look up how to construct an array in C++. Not really because my memory is bad (okay, kind of), but more for the practice of reviewing API's and parameters and maybe making new revelations. I probably knew internally some 12 years ago I could set an initial capacity in one of the array's constructors. I didn't really "understand" the value in that until one random work day 4 years later as I was working through some in-house container packages.
You can even say that doesn't matter if you know what resize() is, but it's those little bits of learning nuggets that builds you up from a "fine engineer" to a "great engineer". Less the knowledge and more the work ethics built to obtain it.
That’a the low value they’re talking about. And all of that stuff is accidental complexity.
To use their example, you need to know that dropping duplicates is a good idea (and why). Knowing what the syntax and api calls to do that is, in my opinion, unnecessary.
Even your example has unintentional complexity. A 'duplicate'? What's that? Do we compare the memory it points to? Through use of a custom comparator function?by an id or tag? APIs and entire languages are designed around how we define such seemingly simple terms.
Ultimately a language, be it a programming language or human one, is about communication. I find there to be value in knowing how to communicate effectively, even if employers increasingly disregard the value of such knowledge. Without it, we'll quickly find out why it's fundamentall quickly.
This is the same thing as pressing the power button on my computer; I want the thing turned on, I don’t want to have to know about memory training.
And then we wonder why apps are getting so much slower and bloated. That's fine if you don't want to think that low level, but someone needs to worry about performance. And others will need different needs for defining a duplicate in their different use cases.
I don’t have to remember assembly intrinsic since the compiler deals with that, I just describe the higher level goal and it implements it.
> But today that's no longer the truth, and for the better. I am happy students don't have to go over the same repetitive low-value processes that I had to go to. Students can now focus on more high level valuable tasks: understand the domain, compare evaluation methods, compare models, balance tradeoffs, present solutions/reports, etc.
You honestly believe that students that don't have the tenacity to stick with learning a simple set of verbs and grammar are going to have the tenacity to learn concepts steeped in jargon?
I do mean learn, which AI use has been repeatedly shown to be detrimental to.
I use AI both in the cloud (I tend to avoid Anthropic...) and with local models.
I've probably written more code in the past year than in my whole career. I can now create what I desire (I just created a telemark skiing game over the weekend based on my Strava segments). I think this has helped my teaching as well. In a recent ML course, I created many interactive examples that in the past were just doodles on my whiteboard, but are now embedded in the notebooks I give students.
I've also taught a few AI courses to clients as well. However, with the inevitable rust/ASM-ification of everything, I'm not sure how long those skills will be valid. I do, however, think that having a human in the loop for ML and data analysis is still important.
How will skill building happen when LLMs and coding harnesses seemingly provide all the answers at one's fingertips?
Having the taste and business sense to develop the right thing will be important. If it stops working, ask the AI to fix it.
People have been asking this for a couple of years already. I have yet to hear a real answer beyond "We'll cross that bridge when we get to it."
At least, so far.
https://www.themarginalian.org/2011/12/21/steve-jobs-bicycle...
You can think about how to best execute the command and that's it. Without that you don't have a military.
They don't provide all the answers. And are often wrong/mistaken. But they also produce stuff faster than you can consume it. This is great for when you know exactly what you want and can cheaply test the output and guide the LLM to the right solution. It is expensive (in developer hours) in other cases.
I have been building a new PL+VM this year and had to pause development on a few occasions because I have not had the time to go through the implementation because there are so many modules. But smaller tools (5-10KLOC), you can judge fairly quickly.
So, the answer to your question is: people need to discover the strengths and weaknesses of the tool on their own, decide what kind of expertise they want to attain, and if it is worth the effort.
That's at least what I do. Sometimes it feels performative at best, other time I get genuinely new ideas by having a vague thought turn into a tweak able viz.
I think our industry needs to transition to figure out what does it mean to build a quality system in the age of AI? It’s one thing to vibe code an application that works. But would you depend your livelihood on it? Would you get business funding and base the future of your team on it?
I think the challenge of the educational space is to figure out how to teach the judgement to build solid systems without relying on “hand coding”. For me I’ve been working on technology agnostic composable frameworks that help build that judgement. But honestly we are all just figuring this out together.
There will still be a need to be able to train system builders, the question really is what do they actually need to learn to be solid system builders in this AI era. And there’s a strong chance that pure programming language or tech stack learning may not be enough. This is what we educators need to figure out.
Yes. I stopped looking at the code about 9 months ago. It's easier/better/quicker to learn how to get the LLM to audit the code.
Investors don't care if the code is LLM-generated, they're mostly assuming that now anyway. The VC's I've talked to recently are screening pitches using an LLM, they're embracing this tech.
If I recruited a new dev to the team now, I would expect them to be able to run multiple agentic coding instances simultaneously. I would not expect them to be hand-crafting code any more.
But, investors do care about the stability of the system they invested in. And how quickly it can be extended and expanded upon while still maintaining functionality. And they deeply care about scalability. I don’t believe this is a question about whether code is LLM generated. I think this is a question of having the ability to build a system that can be up and provide the value that investors expect. And the way you use AI and your knowledge of systems does play a role in that, especially over time.
And another point: it doesn’t have to be either or. Sometimes the prudent thing is to hand code a specific highly sensitive part of the system while your agents are building other things. The fact that these agents can do things in the background frees us with different possibilities. And also, sometimes hand coding some areas in the code that you are not familiar with builds your knowledge which the ln enables you to guide the agents more effectively later on. There’s benefits of both approaches and both approaches can be done simultaneously.
I had to go back to writing code to actually have it register.
Learning is friction. LLMs remove friction.
I for one mourn the loss of these and other blogs and am not exactly sure the trade was worth it. Not with the cretins in charge.
When you hear the phrase "but first we have to go back to..." on YouTube, you're entering a filler section to allow more ad content.
Superior to any teacher I've ever had, no matter how involved they are. I genuinely don't even see what my teacher is needed for beyond a requirement for me to be there. I can ask the bot "can you clarify X section" at 3am and it'll do in 1 second. In class, the teacher reads the slides, adds some extra points and takes questions. But my bot does all this, completely personalized to me.
Times move on, and I don't see teachers surviving. I'm hoping by next year the college will offer an "at home, self paced course" option. Theres no reason for me to be there.
Teaching should turn into conveyance of taste. The teacher should help figure out what needs to be examined and how to examine; help figure out ways to close the gap between the textbook and reality. AI can't do those things well yet, and perhaps it never will be able to.
Maybe all education will get more and more similar to e.g. learning an instrument.
You can get everything from YouTube videos, sure. But a good teacher will help with things like knowing how to get a good repertoire for your skill set, what bad habits not to develop that will cripple your ability to learn more complex pieces over time, etc.
I think expertise in hard subjects will only become more and more valuable (as always), and thus harder and harder (and more expensive) to achieve.
Sucks, but its true. In my case, if there was a test that required me to disassemble an engine top to bottom, claude would be mostly useless.
I have a feeling AI is hitting all the same dopamine receptors as your attention grabbing apps and video games, is highly addictive, and is merely making you feel like you are achieving something. Sort of like DuoLingo.
If I was using this AI and my grades were poor then I'd agree, that its perceieved but its provably working.
We are kind of envisioning that university will be a place for social activities, mingling, rich people meeting partners, etc. So we are moving towards more of that.
Have you independently verified your learning?
AI is moving fast, so I'm very open to the suggestion that Claude can do today what it couldn't six months ago. But my experience with folks who are learning from AI, historically, has been that they have high confidence in abilities they are not able to demonstrate when asked to show mastery through conventional problem solving.
I think that is possible with AI (I haven't experienced it), but I also feel like most folks use basic/naive/cursory techniques when using AI because they don't know what they don't know.
Part of it is shear volume of "stuff" that I'm doing on the fly while leaning on AI as a crutch for speed (otherwise why wouldn't I just do it myself), the other part is that it's ephemeral, even if you read all the lines of code, it's hard to beat manually typing them yourself for retention _and_ failing at it, and being forced to teach myself what I did wrong.
All that's to say, I think AI psychosis can happen at a group level.
I do, all the time. I get test, do good and then the college signs off that Im a certified mechanic when Im done.
Maybe its more of a critique on the school, but no one tests you on your ability to disassemble an engine and rebuild it. They simply make you write a test based off terms you learned in a textbook.
This was exactly my experience in high school, university, and now the same in college.
I would of killed for this bot during my math courses in university.
Out of curiosity, where–geographically–are you going to school?
If you're able to pass you theoretical tests without AI, it's fair to say you're learning what's being tested.
Then, for better or worse, I was pushed to honors and AP classes in high school. The same methods didn't work there. Things got more abstract and I needed to connect multiple concepts together. Math drills turned into word problems that revealed weaknesses in my older math fundamentals. reciting facts from a required reading wasn't enough; I needed to know what that meant and understand the setting of the book.
My grades suffered and you can argue the "academically optimal" way would have been to not take so many advanced classes. But those experiences definitely helped me to critically think better. Eventually (My English and History appreciation would come much later. Math and Science trained me well).
It is also a shame how much the quality of the teacher can affect this experience and perception of education as a whole. I wonder if those enthusiastic to replace the teacher ever had one that fit their needs?
You clearly have enough humility in your field to be able to soak up knowledge. That's great for you! I suggest your dedication is rare and teachers continue to be essential for society.
I never saw a teacher that managed to do that. Not to any of my classmates, nor to me.
Teachers and schools need to let LLMs do the things they are better at, and concentrate on having instructors do the things that LLMs cannot. Teachers that only read from the slides have always been a problem.
Your instructor should be having you dismantle an injector pump or bleed an air locked injector rail. You can read a book about how to do that, but you won’t really know how to do it until you’ve done it for reall
LLMs are not the solution to providing the best teacher at scale that people think it is.
If we want an education system optimised for producing capable agent jockeys, then by all means this seems an efficient way to do so. But sometimes I wonder if, for all the modern wonders capitalism and individualism have brought us, we haven't lost track of being humans - all sharing a temporary seat on a fragile spaceship, hurtling through darkness and uncertainty.
Just keeping our heads to the AI grindstone until we die doesn't really seem to be a good answer to what to do with our existence.
It's the great turn inward of humanity.
That’s a shame. I’m surprised someone so well versed in web technologies has such a brittle setup. Cloudflare Pages, or if you you’d prefer to stay CDN-less, a $4 VPS, will handle any sort of traffic this blog can reasonably receive.
Imagine: you have possibility to load desired SKILL into your brain/body - lets say it is fluency in foreign language so you can use it as native language. Will you choose it or "old way" which is hard, long and does not guarantee 100% mastering fluency? Or improving your body, making you faster, stronger with more endurance. Or SKILL allowing you breathe underwater. All of these and more are just human (of course not all humans) dreams. If you could choose, you really would choose old thing instead of new thing which makes life easier.
I guess we as humans should definitely think about making life easier and think at bigger scale - the Universe. We still do not know our oceans, lands and so on but imagine what could be done if humans were better - in all fields you can imagine. Integrating AI features is hard problem to solve but I hope we will find out solutions. Personally, I would like to obtain many skills easier way. We are still in times where we have to learn or at least read/verify content generated by computer models. But what if we had power of these models in our brains and bodies...
If I had a question about a new library now, I wouldn't google for a tutorial, I'd just ask ChatGPT.
I'm experimenting with exactly that: https://miletus.app
I suspect a lot of people who formerly might have been interested in a web dev tutorial now just an LLM to make them a website.
I'm still doing hybrid coding, half LLM half manually. But every new LLM does more and more of the job...
I used AI to extract the raw frontend structure of the app because I know I can't design to save my life. And I still need to tweak it quite a bit because the interface was always janky (a desktop app but content is overflowing everywhere, and it definitely can all fit in a typical full screen view). But the act of structuring that stuff into a modern framework and hooking up the logic will still come down to me.
Web development, as much as some would like to think, is and has always been, complex.
I haven't tried learning Rust yet, and I'm sure as hell not going to lean on an LLM at all when I do.
I don't think what's happening is a uniquely AI phenomenon. If it wasn't AI, it was going to be something else. The ground is always shifting beneath you. It's happened to me in multiple fields through my career. You adapt or you pivot. Like a startup.
For many people who feel negatively affected by AI, my advice to them is to do what I did when my professional disciplines changed to something I didn't recognize: pivot. The great news is that AI makes it really easy to be a beginner in a new adjacent field.
This underestimates the difficulty of pivoting towards a career that won't be affected by AI any time soon: blue collar work, most likely, for us knowledge workers.
I'm keen to learn plumbing or becoming an electrician, but I won't hide the fact that I am so used sitting on my ass at home all day and get paid, not making physical effort, get callused hands and driving around in a van to make rent. I'll have to do it, but to say it's a traumatic change for someone in their 40s is an understatement.
The silver lining is that coding for fun, by hand, however I want, on whatever I damn please, after a whole week of physical activity, will be the best reward, one I haven't felt since my teenage years.
This time my approach was better and so was the capability of the model. I can make a small change in one place and have the model propagate the change throughout the codebase. I'm still working directly with the code, and the agent doesn't have to make any important decisions about structure (which remain generally terrible). My familiarity with the codebase is the same as if I had made each edit myself.
It is definitely possible to build faster and better with an LLM and not give up control or the enjoyable part of coding. The energy in this space still seems to revolve around guiding the agent or letting it take over, the key is ignoring it entirely and discovering what works for you.
just because our old-brains did it "the hard way" and now we have these LLMs that can just make it happen, doesn't mean that our little ones want to/should skip the steps of understanding HOW IT WORKS
me and my kids have been making all sort of silly games and my oldest is starting to wonder HOW DOES THAT WORK, PAPA??? and i'm like, well, let me show you this...
a magical thing! what a time to be alive!!! there is joy, but you have to conjure it!!
I'd gently submit to the author that these are not opposites. Knowledge commodification is good, and so is deep understanding.
Good article.
Rachel Andrew’s (mentioned in TFA) for example has The New CSS Layout (2017) https://abookapart.com/products/the-new-css-layout.html an excellent book I still recommend for anybody who want to learn about modern CSS layout.
But if you are this dismissive of the field of course you will fail to see the value of them.
I’ve never understood that but for some reason it is a thing
But then around 2020 I realised kids actually have no clue how to use computers because their primary computer is a phone or a tablet. And there only experience with desktop computers are Chromebooks at school.
Now I suspect my generation will be the most knowledgable technical generation by far. The only issue for us is that no one will care.
As someone who used to write a lot of tech tutorials and has published a technical book, for me it's much less a financial thing and much more just sadness that something I spent a lot of my life doing, that used to generate emails every week from people saying thank you for helping them is now just completely worthless. I've spoken about this before here, but I feel like I've lost a large part of myself over the last couple of years because almost everything I used to do in my spare time now seems completely pointless.
LLMs with HTMX and plain old CSS and JavaScript (maybe Typescript) is the future. Getting back to little to no build times and frontends that don't take up 1+ GB of RAM on the clients
I think traditional web frontends might be dead for any applications with servers that can afford to keep server-side state of a logged-in user.
(Notably this includes almost any kind of AI application, because the LLM costs dwarf the web hosting costs anyway.)
And for the stateless ones, there's always htmx
1. LLM webpages in my experience have definitely not gotten simpler, nor load faster. It just pushes out more of the same junk underneath without any of the oversight we'd give a human
2. I'm not even sure the webpages are the big resource hog anymore. Chromium itself has had the reputation for eating RAM for well over a decade at this point.
3. In many ways, we did this to ourselves. In the name of shipping faster, we (or perhaps the executives) outsourced solutions that would sacrifice the simplicity of the code for "more mainainable, cross platform" frameworks. Electron was a solution for software companies not wanting a dedicated Windows and Mac (and perhaps Linux) team. React was a very specific solution for one of the largest websites in the world, that the rest of the industry adopted as a trend.
I pulled up the QR code menu website on my phone and grimaced.
“This is AI slop!”
My friends paused their scrolling. “What’s that?”
With AI everything has been upended.
And by "everything" I mean everything: how we teach, learn, test, and most importantly, the kinds of jobs we will have in the near future.
With AI, all the "cogs in a machine" jobs are essentially gone. There is no economic reason to employ people when AI will do 80% (and rising fast!) of the work at a fraction of the cost without pesky things like perks, leaves, raises, promotions, or heck, even sleep. Humans will be needed only for the 20% (and dropping fast!) of work that AI cannot do and to keep it honest.
And unfortunately, to get to that 20% we had to go through years and years of the 80% to "pay our dues" and learn enough to get up there; the exact 80% that is usurped by AI. Not only is education hosed, so is expertise building through on-the-job experience.
But! On the flip side, we can now have a dedicated, personalized, omniscient, always-available tutor! Do we even need school or college? Like, literally a 100% of children are naturally curious and exploratory; just that formal education (shaped by understandable economic pressures) is very effective at beating it out of them. What if we free children from the restrictive rails of rote, conventional education and let them explore wherever their interests take them? They could stay home to spend more quality time with parents who outsource work to their agents.
And AI makes it very easy to just do things, and we learn best by doing, so let them do! Would that not be a good way to nurture motivated, practically skilled talent that is truly passionate about whatever discipline they choose, which will be the exact quality needed to drive them to push the boundaries where AI cannot yet reach?
We have to critically examine with brutal honesty what jobs will be needed in the future, wholly rethink from first principles how we should prepare young talent for those, and start adjusting our society and economics to safely transition to that future. I do not know what that looks like, but I know it will be tumultuous and that we must start immediately.
This is Schroedinger's Future, simultaneously terrifying and glorious.
It is already here. That is the current reality.
"I knew one day I’d have to watch powerful men burn the world down – I just didn’t expect them to be such losers"
Good.
I mean, it's either that, or becoming another interchangeable meat proxy.
I bought some books to support A list Apart, but most of the content seemed to be an aggregation of stuff that authors were already writing about before/after.
In that case, why not just find a good Udemy course after maybe a few hours of research and throw $11 at it when it's "on sale" (I hate the psychology there, but that's a whole other bag of worms) when I just got back from paying $12 at Starbucks?
Of course, I'm not a web dev and don't need full books whenever that tech inevitably comes my way. I love books for my own domain, and several of them are well worth their weight in gold, since all the amazing tech out there isn't just freely available to view with an inspector.
I have no idea how it will feel to be forced to write code and learn data structures, and SQL, and infrastructure, when it's all right there in 15 seconds with the AI agent.
Does anyone have any experience with this in 2026 specifically? How are the kids doing?
At least with a medical degree you HAVE to study, and memorize. It's part of the job.
I am convinced that there is no fundamental difference between artificial neural networks and biological ones. Eventually, anything a human can do, a computer will do better.
1. What kind of programming do you do?
2. How would you describe the difference between AI coding in 2023 vs 2026?
3. How confident are you in your prediction of AI capabilities?
> Does anyone have any experience with this in 2026 specifically? How are the kids doing?
Here's what it looks like from the perspective of Europe's largest tech company as of 2 months ago [0].
0 https://news.ycombinator.com/item?id=49229412
SAP stops most travel and hiring because of AI's soaring cost (404media.co)
102 points | 53 days ago | 68 comments
You need to know about database normalization, about algorithms and techniques, ideate solution to problems etc because as impressive as those 1-shot youtube videos on model release look, once you go beyond that, that's where the value ad comes in my experience.
I'm talking about learning/training for the job!
My experience with Manning, OTOH, tends to be good. Most of the material I've seen from them was human-written.
If it weren't for the job losses I'd never want to go back to 2015 or so to go back and forth on SO questions and answers and spend days on a boilerplate issue.
By BS, i mean tsconfig files and caddy boilerplate.
“Nobody’s buying your course bro” has been a meme for years
things people didn’t want are disappearing, the people that would actually take initiative in education have another solution. those error-prone chatbots that the article mentions are good enough. most people aren’t trying to get jobs in the profession, they are trying to make minimum viable web products and landing pages that fit on free tiers, and succeeding
You want it structured? You need the book - "AI" won't solve it for you (well, if we discount cases of word for word reproduction... which we probably shouldn't, but this does not seem to be the argument here).
The kind of people, that read the books, are the kind of people that largely won't settle for a chatbot as a replacement.
Can there be another reason for the sales drop? Unemployment, "death of the junior", wages decline?
There are a vast number of books that I purchased not for any interest in the topic, but to solve a particular problem. I have zero interest in understanding Docker. I find it tedious. But I do have a book on it as I did need to know it at some point.
I wouldn’t buy that book today, as AI means I no longer need to know Docker.
> The traffic to my blog and my books (which were free to read online) increased beyond what I can currently afford. Virtually all of it comes from AI crawlers, so there is no ad income.
I know correlation is not causation, post hoc ergo propter hoc etc. etc. But these circumstantial evidence are too strong to simply ignore... As it stands AI is the obvious culprit. And with the current evidence, it is perfectly reasonable to blame AI and nothing but AI.
Personally, I read way less traditional books when it comes to learning stuff. Especially the “utility”/“tutorial”/“handbook”. Top nitch books talking about general principles I’d still read.
Especially for books that need to teach me a subset of a certain discipline - in the past I’d get 4-5 (or at least samples), and try to find the parts that are of interest to me in a style that fits me. Novadays I’d just ask Claude to explain things in a form that I like, with pretty illustrations from Imgen :)
E.g. I don’t think I’ll read the animal books from O’reilly again. But an equivalend of “Thinking in C++” or “Pearls of programming” of a new field? Sure.
AI for sure impacts ad revenue. It's an absolute mess trying to measure proper web traffic anymore, and it' been shown quite a few times how AI summaries in search impact page clicks.
It sucks for the people caught by the change. Honestly, I’m sorry. But rapid, disruptive change has always been part of working in technology. The industry doesn’t owe any particular job category permanence.
I think the more interesting conversation is what happens to the learning industry now. If AI changes how software is built, then teaching people to become developers has to change too.
We shouldn’t be asking how to preserve the old pipeline of teaching people to write code so they can get an entry level programming job. We should be asking what it means to teach someone to become a developer when AI can increasingly write the code.
The scarce skill may shift from producing code to understanding what should be built, decomposing problems, designing systems, evaluating tradeoffs, debugging complex behavior, verifying AI generated output, and understanding whether the resulting system is actually correct, secure, maintainable, and fit for purpose.
That doesn’t necessarily mean fewer people need to learn software development. It could mean many more people can become capable developers because the mechanical barrier to producing software has fallen dramatically.
The educational model just has to catch up.
The technology changed. The answer shouldn’t be to pretend it didn’t. It should be figuring out how to use it to produce the next generation of developers.
That’s going to be contentious for some people, even a little insensitive. But it’s just how it works in our industry.