alt Hacker NewsCoding expertise is going to collapse from AI reliance
Comments
Coders should be treated as as well as as other software users. Tools like Claude code, Antigravity, and others are applying AI to solving that problem.
I know how to create a multi year P&L projection from scratch. It gives me some insight into how much capital a startup will need to get to breakeven. A good AI tool that creates that same spreadsheet in minutes should also be able to explain why different businesses have different shaped cash flow troughs. I don't think people using tools like that is bad even if they don't gain deep insight.
It's pretty simple.
You treat the AI like a junior.
You make it write pseudo code in the tickets.
Then you make it write code in the tickets for the tricky stuff.
Have it reference relevant documentation/APIs.
Then there is no lost knowledge.
(Still, sometimes it sneaks in some "helpful" belt and suspenders, but for the most part, I know everything that is going on in the code base with this method.)
Something I noticed trying to learn rust through models, they're inherently tuned to just finish everything with even the slightest provocation so it's very hard to use them to actually learn anything, they kinda just want to do everything themselves.
At some point it won’t be realistic to ask a human to solve a technical problem any more. I’m not suggesting it will happen soon but I can at least imagine that day coming. Whereas before modern “AI”, that possibility wasn’t even on my radar.
Sewing expertise is going to collapse from sewing machine reliance
This just feels like someone whining about how ORMs result in $#1tty data access code because lazy programmers don't take the time to learn how to use a database.
When that happens, (developer thinks using an ORM means they don't need to know how to use a database,) the project eventually fails. It's the same thing with AI; just like companies learned that they need to make sure developers actually know how to use a database, companies will learn that their engineers actually need to know how to program.
Software engineering >> Coding. How many times do we have to keep repeating this. Author wrote a fucking marketing piece.
Stop the FUD for profit
A list of things, that were supposed to lead to our intellectual demise:
- Writing (starting from Platos Phaedrus)
- Printed books
- Photography
- Typewriters, keyboards
- Telephones
- Recorded music
- Television
- Wikipedia
- GPS navigation
- Smartphones
- Automation in aviation, medicine, industries
I am not personally worried on this front. It is a near-certainty that something catastrophic will happen in the not-too-distant future because of the issues Lars raises in the essay. Not until that moment occurs will there be a broader industry-wide reckoning. Most likely it will get expressed in breathless language like "you'll never believe the crazy gains we got from adding tests!" (or whatever).
I feel pretty good asserting that, as always, a few smart leaders will recognize early that quality and stability are the new game now that we have speed 'for free.' And then there will be incumbents whose eventual demise will merely be accelerated.
look to past examples. when the greeks developed writing, people complained that the youth were losing the ability to memorize much content. and that is true, but with writing they were able to do better still and that old skill -- though useful -- was lost.
And so what? How do you think people are going to be working on software in 10 years. They're going to use coding agents and harnesses that work much better than the current ones (although I don't think we're getting anything but incremental improvement from here on out). They will run on commodity hardware. I first got paid for writing code in 1999, I've worked at publicly traded software companies and failed startups, in every language out there. And if LLMs "atrophy" the expertise I built up over those years, then so what? What is lost, exactly? I stopped programming in C++ about 10 years ago when Rust seemed to get good. I no longer remember a lot of the specialized knowledge I had in C++. My life goes on. I am now becoming an expert at building software with agents. The problem is still the same, you still have to apply engineering principles. You just don't have to write the code.
The boss says we should do it this way. The boss hasn’t got a clue about how to do it. But he knows what he wants, more or less. And we can fill in the details. And we don’t make a stink about him not knowing exactly how things work. That doesn’t serve anyone. And he still thinks that he knows how to do it. Because he tells us how to do it and it gets done.
Now everyone can be the boss.
This seems right to me, but I think that it's likely that we haven't fully figured out how to integrate AI into coding imo. I think there is user side thing that we haven’t fully worked out yet.
Wasn't everyone making these arguments for the last 50 years with each wave of advances and convenience? "Real programmers use [current_tech - 1]."
I mean, they were right, then, too. My meaning is that expertise already collapsed several decades ago.
--
One of my favorite talks of all time:
Jonathan Blow, Preventing the Collapse of Civilization (2019)
Hand weaving expertise has also collapsed.
I believe it. I haven't written code since March.
While true, I'm not sure it matters.
It kinda seems like wishful thinking on our part as software engineers that one day we'll be able to go "Hah, miss me now?" but that just isn't the writing that I see on the wall.
Over the past year, I've AI-generated two large pieces of software over 2000 commits without having much depth in either domain, and my ability to deliver value as a veteran software engineer has only shrunk over the months or it can be trivially extracted into a reusable markdown file like "design bar: ensure things 'by construction' where possible", a lesson I learned viscerally over decades.
I was thinking about this anti AI sentiment and how it is related to American individualism. In socialist countries like China there has never been a protest related to data centres or job loss etc. unlike in the USA. Because the large scale impact of AI on the country is exclusively a net positive. The reduction of the individuals cost of labour by also reduces perceived egalitarianism which is the mental model used by ultra capitalists to allow themselves to sleep at night whilst homeless people beg for food on the streets.
The long term impact of AI on society is exclusively a net positive if utilised in the correct way. But most of the westerners lack faith in government fiscal policy and think that they are going to be left destitute when they can no longer provide for themselves. When in fact UBI is going to be the most likely outcome.
They are myopically focusing on the short term labour market effects of AI on their own profession instead of realising the long term utopian societal outcome.
"We need to ban AI because I'm going to lose my job!" - This is only said by someone who is winning in their society and wants to maintain the status quo not by those who slipped through the cracks.
[copy and pasted comment from a similar post]
I personally write lots of code manually and use AI to review it.
Sounds like a good compromise, if not even positive from many points of view.
It's surprisingly late for him to notice. But wait til you think through the consequences of the already collapsing human capacity for extended written composition, which is orders of magnitude more significant.
Of course it will. Doesn't matter though. Coding expertise is no longer valuable.
What is valuable now is productivity. Which in today's day and age is the ability to quickly and efficiently drive AI.
Hence the never-ending growth and spewing out of tooling: harnesses, TUIs, multiplexers, agent frameworks, software "factories", etc, etc.
Software engineers are chasing opportunity and expertise in the only thing that's left.
I think the hope of the business end is that coding expertise will at some point no longer be needed at all.
I used to think this was a problem. But now? I don’t think it matters.
Why? Because people are going to get so much more ambitious about the stuff they build with AI, that code expertise for the stuff they are building is already scarce anyway. People will choose difficult and less popular languages, and will replace open source packages with stuff built entirely in-house by AI.
Going forward, there is no choice except for AI to fully replace the need for code expertise. Depending on humans won’t be scalable.
Ever since I was a teenager, I started to develop cyclical bouts of existential crises.
While I have been feeling better for quite some time, I feel another starting to brew. While I was never one of the fabled 10x or rockstar developers, I still derived a lot of enjoyment and and sense of purpose from my work.
With the utility of LLMs becoming more ubiquitous by the day, what purpose can I provide anymore? I struggled greatly to get where I am today, and while it might not be impressive to any of my fellow programmers, I am still proud to be a programmer.
I obtained a degree in CS because I found the entire field to be fascinating. I had great passion for learning many of the subjects. What I learned in my degree changed my life forever. I seriously went from one of the, "What's the point in learning most math if you never use it?" to being absolutely infatuated with the subject. (I'm still bad at math, but that does not detract from my pleasure).
With the utmost certainty, I did not push through all my struggles just to prompt LLMs all day. Maybe's its my insecurities manifesting, but I want people to say, "You did a great job" and not "Claude did a great job." I understand the Buddhist were likely correct. My suffering really does come from my attachment, but at the same time, I do not want to let it go.
I have no opposition to using LLMs as tools. I think a fine balance can be achieved. LLMs can even be exceptional learning tools when used appropriately.
If programming is just a means to an end to you, then you will receive no ill judgements from me. We all have different motives in life. But programming has always meant something more to me.
Very few people know how to hand set typography anymore, but many of the conventions still exist and people don't have to understand the origin of a lot of them to use them properly.
I'm seeing it happen in real time. I've seen several folks claim they "vibe-coded" something and upon review it's some seriously low quality garbage. Unfortunately it's just "functional enough" where it's encouraging the one who talked to the robot to continue thinking they can code for some reason.
It is important to ensure that any LLM code is named and shamed.
As always, the innovators will thrive and become the new experts in the field. Everyone else will die out.
It's impossible to buy the FUD that expertise will somehow vanish because people use AI. People come in a variety of capabilities, challenges, and personal interests. Expertise varies from person to person. Those with personality traits that are more likely to take risks and persist against adversity will be the innovators and experts. We've seen this played out countless times.
You have to ignore everything you know intrinsically but can't articulate about innovation, technology, and evolution to drink this FUD Kool-Aid.
One interesting feature of vibe coding is that it allows you to write code in specialized areas where you have no expert knowledge. That makes the code effectively impossible to review in any useful way.
For example, I have a vibe coded Python monstrosity which I use for stitching together scans of large format negatives. The enlarging lens that I use for scanning has strong chromatic aberration in the red channel but is otherwise very sharp†. I got Codex to figure out some way of using the green and blue channels to sharpen the luminance while preserving the color information from the red channel. I sort of vaguely understand how this works, but not nearly well enough to review all the horrid numpy code that implements it.
For my use case, there is really no reason to look at the code, as I am only interested in the output. As disconcerting as it is, I feel that this is going to become more and more normalized. Looking at the code that your LLM generates will become like looking at the assembly output of your C compiler.
† This makes a lot of sense, if you think about the color of a darkroom safelight...
Coding expertise was on a downward trend way before AI arrived. Does nobody remember how most "engineers" were literally copying and pasting snippets from StackOverflow they didn't understand, and the snippets usually had tons of bugs? How many of you know Assembly well? How many of you can decompile a program, or write your own compiler or VM? How many of you could pass a computer science course (that you didn't take at Uni)?
We've been hearing about the demise of skills since before the invention of the loom. The only skills we've truly lost are for things nobody uses or makes anymore. What actually happens is, we develop a new tool that's better than the old tool, and we get good at using the new tool. We then have a class of people whose job it is to maintain the new tools. We don't need that many of them because the whole point is to eliminate that labor.
Your jobs as software engineers are going away, it's that simple. Your new job is to use the AI tools to make products. Somebody else's job will be to make the AI a better software engineer.
> AI models lack judgment
Very interesting is how divisive is this statement. In that almost every response is either hard agree or hard disagree.
Does agreement come from observer judgement, or lack thereof?
We're all posers - unless you can implement square root in machine code, you're not a Real Programmer™
Coding expertise is going to collapse from compiler reliance (1985 Turbo Pascal detractors)
I'm also concerned about people asking an opaque machine for an answer and stopping there instead of going to different sources, evaluating their credibility, exercising their own ability to construct their own worldview from disparate facts and opinions.
Going to? It already is.
It may bifurcate the average skill from the skill of figuring out new solutions.
Are engineers still making 100s of thousands a year still just for vibe coding while a nurse and other skilled / highly trained medical professionals working their butt off make half of that?
Do we think there is going to be a correction for the vibe coders?
Pardon I know such question is controversial but thinking ahead.
I hate to day this, but most coding work isn't hard at all - for a human or AI. There are a lot of important skills around coding, but for most people, much of the time, they are not doing any coding that is particularly challenging, difficult, it novel.
If you have a truly hard problem - sure write some code or take a hybrid approach. Still, let's not pretend this is what most people are tasked with.
"Assembly expertise is going to collapse from compiler reliance".
Ah yes. Selling courses is the oldest grift in town, which is what the author here is doing.
To be fair nobody codes in ASM and things are fine for decades.
Keep in mind also that LLMs are currenctly heavily subsidized. Once VCs money are gonna run out, you will see the real price, and your 200K per year dev is probably nothing.
I learned BASIC, Pascal and x86 Assembly language by the age of 14-15. None of which I remember (nor care to). The strict instructions (aka language) we gave the compiler was not the point. Those experiences taught me how to think logically, and over time develop a taste/structure for how software should be built, as well as foot guns to avoid. I don't know how you teach that without coding, but the "coding expertise" was always a means to an end. LLMs are just another abstraction layer above programming languages, machine language, etc.
[flagged]
[flagged]
[dead]
[dead]
[dead]
Yep.
In my experience human design -> agent code -> agent review -> release -> agent observer, does not (at all!) work. One absolutely has to be understand what is going on, how the system works, how the H/W works, etc, to catch agent errors.
The proposition of us losing those skills is terrifying.
Maybe AI will get better and we truly won't need a human in the loop. Then please also have all PRs signed by whatever model wrote them, and do not come to me for trouble-shooting.