[flagged]
When LLMs first showed up publicly it was a huge leap forward, and people assumed it would continue improving at the rate they had seen but it hasn't.
How do you know that? For tech products most of the users are also technically literate and can easily use Claude Code or whatever tool we are using. They easily tell CC specifically what they need. Unless you create social media apps or bank apps, the customers are pretty tech savvy.
With AI, probably you don’t need 95% of the programmers who do that job anyway. Physicists who know the algorithm much better can use AI to implement a majority of the system and maybe you can have a software engineer orchestrate the program in the cloud or supercomputer or something but probably not even that.
Okay, the idea I was trying to get across before I rambled was that many times the customer knows what they want very well and much better than the software engineer.
But I'm happy about this. I'm not that interested in or optimistic about AGI, but having increasingly great tools to do useful work with computers is incredible!
My only concern is that it won't be sustainable, and it's only as great as it is right now because the cost to end users is being heavily subsidized by investment.
But where is the S curves for programmers at?
Maybe you already understood this, but many of the "AI boosters" you refer to genuinely believe we have "seen the start of it".
Or at least they appear to believe it.
Have you ever paid for software? I have, many times, for things I could build myself
Building it yourself as a business means you need to staff people, taking them away from other work. You need to maintain it.
Run even conservative numbers for it and you'll see it's pretty damn expensive if humans need to be involved. It's not the norm that that's going to be good ROI
No matter how good these tools get, they can't read your mind. It takes real work to get something production ready and polished out of them
At my company, we call them technical business analysts. Their director was a developer for 10 years, and then skyrocket through the ranks in that department.
AI usage in coding will not stop ofc but normal people vibe coding production-ready apps is a pipedream that has many issues independent of how good the AI/tools are.
https://www.commitstrip.com/en/2016/08/25/a-very-comprehensi...
I'm not sure how well that would work in practice, nor why such an approach is not used more often than it is. But yes the point is that then some humans would have to write such tests as code to pass to the AI to implement. So we would still need human coders to write those unit-tests/specs. Only humans can tell AI what humans want it to do.
Unit tests are the correct tool, because going from an almost correct one to a correct one is hard, because it implies the failure rate to be zero and the lower you go the harder it is to reduce the failure rate any further. But when your constraint is not infinitesimal small failure rate, but reaching expressiveness fast, then a naive implementation or a mathematical model are a much denser representation of the information, and thus easier to generate. In practical terms, it is much easier to encode the slightly incorrect preconception you have in your mind, then try to enumerate all the cases in which a statistically generated system might deviate from the preconception you already had in your head.
An exhaustive set of use cases to confirm vibe AI generated apps would be an app by itself. Experienced developers know what subsets of tests are critical, avoiding much work.
And, they do know this for the programs written by other experienced developers, because they know where to expect "linearity" and were to expect steps in the output function. (Testing 0, 1, 127, 128, 255, is important, 89 and 90 likely not, unless that's part of the domain knowledge) This is not necessarily correct for statistically derived algorithm descriptions.
a) Testing that the spec is implemented correctly, OR
b) As the Spec itself, or part of it.
I know people have different views on this, but if unit-tests are not the spec, or part of it, then we must formalize the spec in some other way.
If the Spec is not written in some formal way then I don't think we can automatically verify whether the implementation implements the spec, or not. (that's what the cartoon was about).
For most projects, the spec is formalized in formal natural language (like any other spec in other professions) and that is mostly fine.
If you want your unit tests to be the spec, as I wrote in https://news.ycombinator.com/item?id=46667964, there would be quite A LOT of them needed. I rather learn to write proofs, then try to exhaustively list all possible combinations of a (near) infinite number of input/output combinations. Unit-tests are simply the wrong tool, because they imply taking excerpts from the library of all possible books. I don't think that is what people mean with e.g. TDD.
What the cartoon is about is that any formal(-enough) way to describe program behaviour will just be yet another programming tool/language. If you have some novel way of program specification, someone will write a compiler and then we might use it, but it will still be programming and LLMs ain't that.
The problem I see is how to evolve such a prototype to more correct specs, or changed specs in the future, because AI output is non-deterministic -- and "vibes" are ambiguous.
Giving AI more specs or modified specs means it will have to re-interpret the specs and since its output is non-deterministic it can re-interpret viby specs differently and thus diverge in a new direction.
Using unit-tests as (at least part of) the spec would be a way to keep the specs stable and unambiguous. If AI is re-interpreting the viby ambiguous specs, then the specs are unstable which measn the final output has hard-time converging to a stable state.
I've asked this before, not knowing much about AI-sw-development, whether there is an LLM that given a set of unit-tests, will generate an implementation that passes those unit-tests? And is such practice used commonly in the community, and if not why not?
( variation of .. "Ours is not to reason why, ours is but to do and die" )
AI can code because the user of AI can code.
Debbie from accounting doesn't have a clue what an int is
Just today, I needed a basic web application, the sort of which I can easily get off the shelf from several existing vendors.
I started down the path of building my own, because, well, that's just what I do, then after about 30 minutes decided to use an existing product.
I have hunch that, even with AI making programming so much easier, there is still a market for buying pre-written solutions.
Further, I would speculate that this remains true of other areas of AI content generation. For example, even if it's trivially easy to have AI generate music per your specifications, it's even easier to just play something that someone else already made (be it human-generated or AI).
What if AI brings the China situation to the entire world? Would the mentality shift? You seem to be basing it on the cost benefit calculations of companies today. Yes, SASS makes sense when you have developers (many of which could be mediocre) who are so expensive that it makes more sense to just pay a company who has already gone through the work of finding good developers and spend the capital to build a decent version of what you are looking for vs a scenario where the cost of a good developer has fallen dramatically and so now you can produce the same results with far less money (a cheap developer(does not matter if they are good or mediocre) guiding an AI). That cheap developer does not even have to be in the US.
At the high end, china pays SWEs better than South Korea, Japan, Taiwan, India, and much Europe, so they attract developers from those locations. At the low end, they have a ton of low to mid-tier developers from 3rd tier+ institutions that can hack well enough. It is sort of like India: skilled people with credentials to back it up can do well, but there are tons of lower skilled people with some ability that are relatively cheap and useful.
China is going big into local LLMs, not sure what that means long term, but Alibaba's Qwen is definitely competitive, and its the main story these days if you want to run a coding model locally.
I hear those other Asian countries are just like China in terms of adoption.
>China is going big into local LLMs, not sure what that means long term, but Alibaba's Qwen is definitely competitive, and its the main story these days if you want to run a coding model locally.
It seems like the China's strategy of low cost LLM applied pragmatically to all layers of the country's "stack" is the better approach at least right now. Here in the US they are spending every last penny to try and build some sort of Skynet god. If it fails well I guess the Chinese were right after all. If it succeeds well, I don't know what will happen then.
> It seems like the China's strategy of low cost LLM applied pragmatically to all layers of the stack is the better approach at least right now. Here in the US they are spending every last penny to try and build some sort of Skynet god. If it fails well I guess the Chinese were right after all. If it succeeds well, I don't know what will happen then.
China lacks those big NVIDIA GPUs that were sanctioned and now export tariffed, so going with lower models that could run on hardware they could access was the best move for them. This could either work out (local LLM computing is the future, and China is ahead of the game by circumstance) or maybe it doesn't work out (big server-based LLMs are the future and China is behind the curve). I think the Chinese government would have actually preferred centralization control, and censorship, but the current situation is that the Chinese models are the most uncensored you can get these days (with some fine tuning, they are heavily used in the adult entertainment industry...haha socialist values).
I wouldn't trust the Chinese government to not do Skynet if they get the chance, but Chinese entrepreneurs are good at getting things done and avoiding government interference. Basically, the world is just getting lucky by a bunch of circumstances ATM.
I would agree that if the scenario is a business, to either buy an off-the-shelf software solution or pay a small team to develop it, and if the off-the-shelf solution was priced high enough, then having it custom built with AI (maybe still with a tiny number of developers involved) could end up being the better choice. Really all depends on the details.
Historically, it would seem that often lowering the amount of people needed to produce a good is precisely what makes it cheaper.
So it’s not hard to imagine a world where AI tools make expert software developers significantly more productive while enabling other workers to use their own little programs and automations on their own jobs.
In such a world, the number of “lines of code” being used would be much greater that today.
But it is not clear to me that the amount of people working full time as “software developers“ would be larger as well.
Not automatically, no.
How it affects employment depends on the shapes of the relevant supply/demand curves, and I don't think those are possible to know well for things like this.
For the world as a whole, it should be a very positive thing if creating usable software becomes an order of magnitude cheaper, and millions of smart people become available for other work.
Counter argument - if what you say is true, we will have a lot more custom & personalized software and the tech stacks behind those may be even more complicated than they currently are because we're now wanting to add LLMs that can talk to our APIs. We might also be adding multiple LLMs to our back ends to do things as well. Maybe we're replacing 10 but now someone has to manage that LLM infrastructure as well.
My opinion will change by tomorrow but I could see more people building software that are currently experts in other domains. I can also see software engineers focusing more on keeping the new more complicated architecture being built from falling apart & trying to enforce tech standards. Our roles may become more infra & security. Less features, more stability & security.
hmm outsourcing doesn't contradict Jevon's paradox ?
That's completely disconnected from whether software developer salaries decrease or not, or whether the software developer population decreases or not.
The introduction of the loom introduced many many more jobs, but these were low-paid jobs that demanded little skill.
All automation you can point to in history resulted in operators needing less skill to produce, which results in less pay.
There is no doubt (i.e. I have seen it) that lower-skilled folk are absolutely going to crush these elitists developers who keep going on about how they won't be affected by automated code-generation, it will only be those devs that are doing unskilled mechanical work.
Sure - because prompting requires all that skill you have? Gimme a break.
At some point the low hanging automation fruit gets tapped out. What can be put online that isnt there already? Which business processes are obviously going to be made an order magnitude more efficient?
Moreover, we've never had more developers and we've exited an anomalous period of extraordinarily low interest rates.
The party might be over.
I was working with developer training for a while some 5-10 years back and already then I was starting to see some signs of an incoming over-saturation, the low interest rates probably masked much of it due to happy go lucky investments sucking up developers.
Low hanging and cheap automation,etc work is quickly dwindling now, especially as development firms are searching out new niches when the big "in-IT" customers aren't buying services inside the industry.
Luckily people will retire and young people probably aren't as bullish about the industry anymore, so we'll probably land in an equilebrium, the question is how long it'll take, because the long tail of things enabled by the mobile/tablet revolution is starting to be claimed.
The job is literally building automation.
There is no equivalent to "working on the assembly line" as an SWE.
>Not so many lower skill line worker jobs in the US any more, though
Because Globalization.
ooohhh I think I missed the intent of the statement... well done!
And overall fewer farmers with more technological skill sets than back in the dustbowl days.
Here (Western Australia) the increase in average farm size by product can be plotted over time along with the fall in numbers working that land.
I think it’s a reasonable hypothesis that the amount of software written if it was, say, 20% of its present cost to write it, would be at least 5x what we currently produce.
I get your point, hope you get mine: we have less legal entities operating as "farms". If vibe coding makes you a "developer", working on a farm in an operating capacity makes you a "farmer". You might profess to be a biologist / agronomist, I'm sure some owners are, but doesn't matter to me whether you're the owner or not.
The numbers of nonsupervisory operators in farming activities have decreased using the traditional definitions.
You aren't going to going to do that to AI systems. If, after a couple of weeks you hit the limit of what the AI could do in a million+ LoC, you aren't going to be able to hire a human dev to modify or replace that system for you, because:
1. Humans are going to be needing a ramp up time and that's damn costly (even more costly when there are fewer of them).
2. Where are you going to find humans who can actually code anymore if everyone has been doing this for the last 10 years?
Look, I dunno what they will do, but these options are certainly off the table:
1. Get a temp dev/team in to patch a 1m SloC mess
2. Do it cost-effectively.
If the tech has improved by the time this happens (I mean, we're nowhere near this scenario yet, and it has already plateaued) then perhaps they can get the LLM itself to simply rewrite it instead of spending all those valuable tokens reading it in and trying to patch it.
If the tech is not up to it, then their options are effectively:
1. Use it as is till the end of time
2. Throw it out, and start again
3. Pray
That's not the case for IT where entry barrier has been reduced to nothing.
The craftsman who were forced to go to the factory were not paid more or better off.
There is not going to be more software engineers in the future than there is now, at least not in what would be recognizable as software engineering today. I could see there being vastly more startups with founders as agent orchestrators and many more CTO jobs. There is no way there is many more 2026 version of software engineering jobs at S&P 500 companies in the future. That seems borderline delusional to me.
Doesn't mean it will happen this time (i.e. if AI truly becomes what was promised) and actually it's not likely it will!
> AI changes how developers work rather than eliminating the need for their judgment. The complexity remains. Someone must understand the business problem, evaluate whether the generated code solves it correctly, consider security implications, ensure it integrates properly with existing systems, and maintain it as requirements evolve.
What is your rebuttal to this argument leading to the idea that developers do need to fear for their job security?
It might be not enough by itself, but it shows that something has changed in comparison with the 70-odd previous years.
Meaningful consequences of mistakes in software don't manifest themselves through compilation errors, but through business impacts which so far are very far outside of the scope of what an AI-assisted coding tool can comprehend.
That is, the problems are a) how to generate a training signal without formally verifiable results, b) hierarchical planning, c) credit assignment in a hierarchical planning system. Those problems are being worked on.
There are some preliminary research results that suggest that RL induces hierarchical reasoning in LLMs.
What previously needed five devs, might be doable by just two or three.
In the article, he says there are no shortcuts to this part of the job. That does not seem likely to be true. The research and thinking through the solution goes much faster using AI, compared to before where I had to look up everything.
In some cases, agentic AI tools are already able to ask the questions about architecture and edge cases, and you only need to select which option you want the agent to implement.
There are shortcuts.
Then the question becomes how large the productivity boost will be and whether the idea that demand will just scale with productivity is realistic.
I think you are basing your reasoning on the current generation of models. But if future generation will be able to do everything you've listed above, what work will be there left for developers? I'm not saying that we will ever get such models, just that when they appear, they will actually displace developers and not create more jobs for them. The business problem will be specified by business people, and even if they get it wrong it won't matter because iteration will be quick and cheap.
> What is your rebuttal to this argument leading to the idea that developers do need to fear for their job security?
The entire argument is based on assumption that models won't get better and will never be able to do things you've listed! But once they become capable of these things - what work will be there for developers?
It's extremely hard to define "human-level intelligence" but I think we can all agree that the definition of it changes with the tools available to humans. Humans seem remarkably suited to adapt to operate at the edges of what the technology of time can do.
It had required a ton of ordinary intelligence people doing routine work (see Computer(occupation)). On the other hand, I don't think anyone has seriously considered to replace, say, von Neumann with a large collective of laypeople.
I mean they are promising AGI.
Of course in that case it will not happen this time. However, in that case software dev getting automated would concern me less than the risk of getting turned into some manner of office supply.
Imo as long as we do NOT have AGI, software-focused professional will stay a viable career path. Someone will have to design software systems on some level of abstraction.
you mean "created", past tense. You're basically arguing it's impossible for technical improvements to reduce the number of programmers in the world, ever. The idea that only humans will ever be able to debug code or interpret non-technical user needs seems questionable to me.
Also the percentage of adults working has been dropping for a while. Retired used to be a tiny fraction of the population that’s no longer the case, people spend more time being educated or in prison etc.
Overall people are seeing a higher standard of living while doing less work.
There are lots of negative reasons for this that aren’t efficiency. Aging demographics. Poor education. Increasing complexity leaves people behind.
So, yes, reasons other than efficiency explain why people aren't working, as well why there are still poor people.
Now we can set arbitrary thresholds for what standard of living every American should have but even knowing people on SNAP it’s not that low.
The cost to participate in society is much greater.
Yeah we do have more cars. But you also need to buy one to go to work.
We have education, but you need 22 years to be employable.
It’s probably not with continuing the discussion if you don’t believe poverty exists as a concept.
Poverty still exists, but vast inflation of what is considered’a basic standard of living’ hides a great deal of progress. People want to redefine illiteracy to mean being unable to use the internet not by the standards of the past.
How would you describe the level of wealth of those Americans outside of metro areas?
Yes, but it’s not why there are fewer adults in the workforce.
I actually didn’t say that. And the twisting of words is the source of confusing
The first line made me laugh out loud because it made me think of an old boss who I enjoyed working with but could never really do coding. This boss was a rockstar at the business side of things and having worked with ABAP in my career, I couldn't ever imagine said person writing code in COBOL.
However the second line got me thinking. Yes VB let business users make apps(I made so many forms for fun). But it reminded me about how much stuff my boss got done in Excel. Was a total wizard.
You have a good point in that the stuff keeps expanding because while not all bosses will pick up the new stack many ambitious ones will. I'm sure it was the case during COBOL, during VB and is certainly the case when Excel hit the scene and I suspect that a lot of people will get stuff done with AI that devs used to do.
>But the job of understanding what to build in the first place, or debugging why the automated thing isn't doing what you expected - that's still there. Usually there's more of it.
Honestly this is the million dollar question that is actually being argued back and forth in all these threads. Given a set of requirements, can AI + a somewhat technically competent business person solve all the things a dev used to take care of? Its possible, im wondering that my boss who couldn't even tell the difference between React and Flask could in theory...possibly with an AI with a large enough context overcomes these mental model limitations. Would be an interesting experiment for companies to try out.
I find SQL becomes a "stepping stone" to level up for people who live and breathe Excel (for obvious reasons).
Now was SQL considered some sort of tool to help business people do more of what coders could do? Not too sure about that. Maybe Access was that tool and it just didn't stick for various reasons.
I certainly hope so, but it depends on whether we will have more demand for such problems. AI can code out a complex project by itself because we humans do not care about many details. When we marvel that AI generates a working dashboard for us, we are really accepting that someone else has created a dashboard that meets our expectation. The layout, the color, the aesthetics, the way it interacts, the time series algorithms, and etc. We don't care, as it does better than we imagined. This, of course, is inevitable, as many of us do spend enormous time implementing what other people have done. Fortunately or unfortunately, it is very hard to human to repeat other people's work correctly, but it's a breeze for AI. The corollary is that AI will replace a lot of demand on software developers, if we don't have big enough problems to solve -- in the past 20 years we have internet, cloud, mobile, and machine learning. All big trends that require millions and millions of brilliant minds. Are we going to have the same luck in the coming years, I'm not so sure.
And that hits the offshoring companies in India and similar countries probably the most, because those can generally only do their jobs well if everything has been specified to the detail.
but the actual work of constructing reliable systems from vague user requirements with an essentially unbounded resource (software) will exist
The skills needed to be a useful horseman though have almost nothing to do with the skills needed to be a useful train conductor. Most the horseman skills don't really transfer other than being in the same domain of land travel. The horseman also has the problem that they have invested their life and identity into their skill with horses. It massively biases perspective. The person with no experience with horses actually has some huge advantages of the beginner mind in terms of travel by land at the advent of travel by rail.
The ad nauseam software engineer "horsemen" arguments on this board that there will always be the need to travel long distance by land completely misses the point IMO.
imagine being an engineer educated in multiple instruction sets: when compilers arrive on the scene it sure makes their job easier, but that does not retroactively change their education to suddenly have all the requisite mathematics and domain knowledge of say algorithms and data structures.
what is euphemistically described as a "remaining need for people to design, debug and resolve unexpected behaviors" is basically a lie by omission: the advent of AI does not automatically mean previously representative human workers suddenly will know higher level knowledge in order to do that. it takes education to achieve that, no trivial amount of chatbotting will enable displaced human workers to attain that higher level of consciousness. perhaps it can be attained by designing software that uploads AI skills to humans...
I was imagining companies expanding the features they wanted and was skeptical that would be close to enough, but this makes way more sense