> because AI is so counterproductive while feeling the opposite
Wow that resonated by a scary amount.
Wow that resonated by a scary amount.
We do spend all our time either code reviewing or talking with clients but the latter was already the case and the former was just better spaced out over time as this would before take many months.
I’m a critical user, not a categorical denier. But there are certain categories where I see the current AI face a wall: fitting new ideas, or existing code into an existing architecture particularly. Or work on the backend when constraints are not strictly enforced (for various historical reasons). Schematic understanding is still an issue.
Sometimes additional chunkings and detailed planning/instructions suffice. But other time, humans are still the best vehicles to code.
It's easy if you're already an expert at software engineering and know how to leverage AI. For people like that, it's a phenomenal upgrade (this is true for me and several co-workers I chat with, all of whom are chasing cool ideas on side projects). But if you're non-technical, or not use to thinking about requirements, or think that using AI is "give me the prompt", it's a pretty big moat to cross.
Plenty of reasons, including:
1. They don't (yet) know how to use AI to accomplish what they need.
2. The ROI is still meaningfully positive and they don't want to have to do it themselves.
3. Having a third-party do the work provides protection for decision-makers. If the project fails, the third-party takes the blame and "nobody gets fired for buying IBM".
4. The vendor does bring valuable insight to the table and pairs it with the use of AI to deliver a result that wouldn't have been possible in-house.
This doesn't mean that everyone will be able to get work and maintain their rates in AI world but some people will.
People talk a lot about how "real programmers" used to have to clean up or actually deploy everyone's half-baked MS Access app, and that sort of thing will probably still be the case for a while.
So we are switching to LLMs to be fucking miserable in our jobs?
A year ago, LLMs were not useful for me as a programmer. Now they are: the models are better, they can use long contexts more effectively, and the harnesses are better at helping the models. Nowadays my job is mostly not programming, but LLMs let me organize and prepare tools in spare time rather than needing days or weeks of attention. I would not trust them on a 200k+ line project -- and Claude Opus 5 has issues even on 50k LOC projects -- but they absolutely can help given good direction and a narrow enough scope.
I do actually understand that.
I have done this very put the Access database on the web job myself. (FWIW I was well-paid for it and the firm I worked for earned a fortune, but this was in 1997)
Meanwhile they've apparently played with LLM porting of Postgres (or Postgres features?) and "would not put it in production"[2], so we can also guess the kind and scope of features being discussed here.
The code their Astra+Fable setup, automatically generated from requests customers leave in voicemails, is apparently so good they never correct it. If the stories being woven are true, the SLA is literally just a middleman's 100k cut.
At the very least there might be some particulars here that aren't universally applicable.
Indeed, welcome to enterprise software.
We can have fun and make money right? Possibly at the same time but not in that case.
Which is fine, great even (I do the same thing!), but in trying different things outside your day job, it seems like you do understand why people working on different things than you might be "reporting all these negative AI experiences"?
It's not a general argument, they made specific claims, but vague posted (and implying everyone else must be crazy) to the point that the conversation is derailed by a bunch of people trying to figure out what they meant.
> it still wouldn't invalidate the core thesis of LLMs providing business value.
lol, no, "providing business value" is not what was claimed:
>> What are people here doing exactly that they are not riding the gravy train and even reporting all these negative AI experiences?
Of course they will do it by themselves.
Because it's cheaper.
I mean it's weird that we all imagine reasons why we're still relevant when we have set fire to the thing that made us indispensable.
It's cheaper.
Very few things in this world are all-or-nothing, and not every purchasing decision is based on price alone.
Do you always buy the cheapest meal? Car? When you renovate your house, do you always choose the cheapest contractor?
Tons of developers will lose their jobs, and many more will find it hard to maintain the salaries/rates the industry has been accustomed to. This is already happening. The days where an average graduate from a run-of-the-mill CompSci program or even a coding bootcamp could sleepwalk into a $150,000+/year entry-level job are largely gone. The days where you have job security simply because you're a competent developer with 10 years of experience are in the process of going away.
This does not mean that there is no subset of developers who cannot be successful in this market. There are people who are doing just fine because they know how to articulate their value and sell themselves to employers or clients.
Because it's not (yet) that easy, especially if we're talking a complex and genuinely useful app. Agentic coding is fast, but it's not magic.
> I’m a critical user, not a categorical denier. But there are certain categories where I see the current AI face a wall: fitting new ideas, or existing code into an existing architecture particularly.
Exactly. Which is why Joe Average still will have little to no luck vibe coding anything serious or novel. You still (IME) need a lot of active guidance, still need to push away from dead ends and propose alternative algorithms, and it takes hundreds of prompts to go from concept to what I would consider beta. (But, this is just my own experience, and it's possible I'm doing it wrong?)
The spirit of the parent comment is true. People feel like it's productive and even if they make something worse, they are going to use it.
They can do it. What's the issue?
I mean, that's exactly what's starting to happen, we have more and more clients to whom we propose a quote and their answer is "guess i'll just vibe code it" or come to use with an app that they vibe-coded and does the job, and they're content with it. So far it seems to work out just fine for them.
If you company is all about creating things that go on the internet, sure, AI is your thing. But for most companies the internet is a communications tool rather than a product. Those companies are not seeing the claimed productivity gains.
Everyone in smaller companies describes various efforts to get AI to do more that ultimately end up involving lots of corrections or nurse-maiding ChatGPT to get it to do exactly what they need, when what they need is often so formulaic they should already have a process.
Nobody I talk to describes it in terms of massive productivity gains. Everyone describes it as a puzzle they feel variously compelled to solve.
OK so, it sounds empty and miserable to me, but we can stipulate that this is how it is working out for you.
Given that stipulation: how long do you think this can possibly last, when you are in a race to the bottom with everybody else who is doing this?
Have you made plans about how you will get out of your office lease, downsize, subcontract this button-pushing to even cheaper people overseas, etc.? Are you ready to lay off everyone who works for you, downsize your house to a smaller mortgage? Because you're on a burning platform. If you can do this, so can everyone else.
I mean, I think we all are, possibly, on the same burning platform; I don't think I can fully avoid AI so I am trying to make sense of it.
But I intend to fully avoid being in a race with other phone whisperers if I can.
I think the years of software development/employment for problem solving, in and of itself, are numbered at this point. When people start to understand how relatively easy this stuff is, they're going to be in-housing everything. Not only for price but also because of increased flexibility/confidentiality/control, and even lower concept-to-live timelines.
What third party stuff continues to exist will see its comp plummet. I imagine now you can already get some really nice quality software contracted out on places like fiverr for nothing simply because the skill involved to produce quality solutions has plummeted.
But yeah — I mean, I am a freelancer who burned out for reasons that are rather more influenced by the SaaS wave. That was hard to compete with. People choose between the solution that meets their needs and changing their processes to meet the ten-times-cheaper solution that doesn't, and that decision is often not at all irrational; by and large you want as many of your problems to be shared problems as possible.
I am trying to return to work so I am looking at what AI can really do for me, but the conclusion I draw is that since I cannot simply burn token money to solve people's problems, and because bugs cost me money, I need a strategy where I remain in full control of actual code, but LLMs help me do things faster. If I can't solve that, I am out.
Perhaps AI will upend the SaaS market before it fucks the freelancer market and the balance may temporarily shift. But it probably won't.
And at that point, as a fiftysomething, even with a bit of financial security, I start thinking about living a rather shorter, happier life, instead of a longer one. Because as much as I might have ideas, I don't think there is much I can switch to where I have appropriately deep skills to survive AI there, and all those alternative jobs will be oversubscribed and less likely to hire me.
I wonder if people in this industry have understood what we are doing to ourselves, to our friends.
But in these parts of the world, making a few bucks doing whatever in the software world goes so much further. Because of that I think LLMs are bringing in something like a temporary golden age. And these parts of the world will be the last to have the lights go out on software simply because the low cost of living means there's a whole lot more slack to give before things get bad.
Quality of life will not improve, because I will be lonelier (plus I have a great quality of life here in other ways, that many places cannot match) but I am mostly reconciled to that loss. I will have a fair bit more flexibility in use of space, which means I can be a bit more of a portfolio-earner; I do have some fringe skills that could make some money still, and I can have a little workshop or studio.
> But in these parts of the world, making a few bucks doing whatever in the software world goes so much further.
I am (undiagnosed but very obviously) ADHD and I don't tend to find it easy to make a sustainable income off "a few bucks" here and there; the effort expenditure in managing it always ends up subsidising the work. It's part of why I am burned out. I need thicker strands of work and those are fewer and further between.
I have little certainty when things will get bad, but I do think it likely it happens in the next five or six years, and I do know that when things do get bad they favour the young. Which I am not. It is clear to me that I am never getting an ordinary dayjob in the tech industry again.
I don't mean to sound particularly gloomy but I think mine is the generation whose lifespans will dramatically shorten. I think a lot of single middle aged people (men mostly) in the tech industry will choose the time and manner of their departure. Because we are deliberately creating both misery and job insecurity.
Since you already have that in you, I think you'll be surprised what you'll find - entire communities of interesting English speaking tech-oriented folks, many with more than a few grey hairs, would be just the start. And don't forget the teaching aspect. If you have a degree, smarts, some basic charisma, and can roll with a bit of chaos - you'll find plentiful opportunities to teach any topic imaginable. It's not just English, like many think. It's really quite fun!
I do not have the psychological constitution to teach kids so I am not going to do that.
But teaching/training adults is an industry that AI will destroy because the baseline income — the stuff nobody else wanted to do but you could earn from — will be eliminated, increasingly by policy directives from above. I have some ideas in that regard but it is difficult to see how I won't end up competing with LLMs when even open weights models are pretty good at coming up with tech tutorials etc.
Broadly I think there is undue optimism about what will be left to move to when the programming jobs dry up.
(Thank you for the discussion, though! I may sound quite negative but actually I am doing better than I have been for years, and it is always useful to provoke one's own thoughts)
I don't think AI will touch education. A good case study there is Khanmigo which was to be Khan Academy's revolutionary AI tutor. There's been a million articles written about the topic. It completely failed, in spite of receiving massive sponsorship and imposition in various educational settings. And I'm kind of surprised that Khan himself didn't understand or predict this. As you probably know, great intrinsically motivated students don't really need teachers. You could give them a book or a sort of LLM tutor, and they'd excel completely independently. But then there are the other 95% of students you have to consider.
And those other 95% tend to fall into camps of either being a bit less gifted in the cognitive domain, or lacking motivation. Depending on exactly where they stand on the balance of two, a good teacher can have a huge impact, whereas 'go learn with the LLM for a few hours' would have them tune out instantly. It's because teaching isn't just about literally teaching, but about forming a rapport and trying to gradually push those 95% into more of the habits and patterns of the 5%, but without them realizing they're being pushed in that direction. An LLM there is almost entirely nonsensical.
The days of b2b saas subscriptions doing one simple thing well for 3 or so users are however numbered.
I know fiverr will start delivering nice software in 5 years and that’s what I worry about. By that time fiverr is probably just a chatbox though without humans, as what is the point?
And your competition is not non-skilled people using AI to take your jobs. It is _cheaper people_ using AI to take your jobs. All the things that outsourcing teams used to struggle to match are much less of a struggle to match when you are actively depersonalising your own effort by handing it to Claude.
It’s spectacular for small projects, limited-scope apps (eg marketing campaigns etc) and for market-testable prototypes.
But if you don’t review and edit the code, things become unmaintainable soup very fast, with subtle logic bugs all over the place. And if you do review and edit the code, when working in large nontrivial codebases, then in my experience AI doesn’t actually go faster even if it feels like that at the start of each task.
Obviously this only holds if you have any sort of code quality standard to begin with (and I agree that with small / short-lived products you don’t need one)
I'm currently working on porting a mid-sized project to a new architecture, new programming language and of course adding new features.
Getting a new feature implemented is quite easy. You spend a few hours brainstorming specs with the agent, then ask it to implement it. This gives you extremely frequent code drops that add a new brick, add a new feature, etc. All of this with 100% code coverage (we also have mutation testing, strongly-typed code, standard and custom linters, etc.)
Then you look at the code. Code that has passed review, generally. You realize that the database schema has been broken silently, and that the agent has rewritten the tests or the golden fixtures to match. You realize that it has made assumptions that contradict the specifications and the product is going to break once it's in the hand of users. You realize that the 100% code coverage is essentially a convenient lie, because the code and tests have been written to make passing easy. You realize that none of the security golden rules have been followed, and that has managed to happen because the agent has somehow deactivated linting.
Why did it pass reviews? Well, because of deadlines. And because there is simply so much code (and so much unparsable/misleading documentation) that it's simply impossible to review all of this. And because things move so fast that nobody understands the CI pipeline anymore, and the explanations of the agent are convincing enough that surely, it knows better than you?
On the upside, bugfixing becomes so fast! Just add a new test, wait a few dozen minutes, and a new Merge Request appears. With equally convincing/misleading explanations, and something else broken.
After ~4 months, we had a bare bones deliverable, which we're now steadily expanding. If we had had to write the product manually, I suspect that it would have taken us at least one year, possibly two. So, that's the productivity increase. The productivity decrease is that what we have is not a product but a glorified demo, something that will work very nicely on the happy path, but on any other path, all bets are off.
> and the explanations of the agent are convincing enough that surely, it knows better than you.
I feel this in my bones. I also get to watch the misalignment feedback loop close itself when the next agent sees that security rules aren't followed because of a hallucinated 20 line justification in a code comment, and then it decides that the project _is_ a demo and then confidently writes even more security holes into the codebase.
Then when you catch the issue, the agent pushes back against the fix because it would need a schema change and production DB migration.
> Then you look at the code. Code that has passed review, generally. You realize that the database schema has been broken silently, and that the agent has rewritten the tests or the golden fixtures to match.
First is don't trust the code review from your local harness, even if it uses sub-agents; externalize it into another system.
Second is to get your most critical human code reviewers to encode their heuristics in markdown files and feed those to the code review agents.
Third, if possible, is to bring the code review "into the loop" so that it's not only running in the PR, but also running in the coding loop so the coding agent has immediate, external feedback. Final PR code review is a backstop.
This pattern [0] works well because it solves for some team level problems where folks are using different harnesses or different models (consistency issues) and it means that code reviews don't just sit at the end of the loop; it actively alters the code production cycle.
They are very good at that.
> "Why did it pass reviews? Well, because of deadlines. And because there is simply so much code (and so much unparsable/misleading documentation) that it's simply impossible to review all of this. And because things move so fast that nobody understands the CI pipeline anymore, and the explanations of the agent are convincing enough that surely, it knows better than you?"
I have come to realize that AI is so "successful" because the system in which it is being deployed was designed to push product as fast and cheaply as possible from the start.
Humans are usually overworked and stretched to their breaking point, which I originally saw as the source of our broken software woes, which, like our streets in the US, just get a new layer of asphalt to cover up the crumbling bits each year instead of rebuilding the infrastructure with reliability and longevity in mind.
My former employer was using both Claude and Codex for firmware that was driving an over-burdened power circuit that itself was partially designed with ChatGPT. All of the individuals involved approach LLMs with god-fearing reverance because they do not understand _how_ the LLM works, just that it _does_ in a "good enough" way and they can offload their thinking, which is something we all wish we could do because thinking is hard, time-consuming and costly. I get it.
But like you mentioned, tests were being passed, not because the code was sound, but because the tests were altered to match the results. This is not necessarily the fault of the agent, either; it's just interpretting the prompt(s) - written by a flawed human, btw - with stochastic mechinations that seem to make a great deal of sense on the surface, but remain unable to be followed or repeated by the brains of (most of) its users.
As a rresult, I had to deal with product that work great in the field...at least at first, before it start literally catching fire, ruining its own powertrain because everything the agents touched became too complex with too many subtle cracks in the veneer to review properly. The system (read; capitalism) demanded viable product quickly to please investors, and the burnt-out humans who decided to try this AI thing ended up trusting it nearly completely, so any ideas of repeatable and complete testing, diagnostics and root cause failure analysis morphed into a sloppy "it works on the bench" checklist before being sold to a customer who had come to trust that their deceptively simple product would just work as advertised.
I'm going to die on the hill that AI as a replacement for our brains is precisely how we will make ourselves go extict, but I am old enough to already be regarded as a crufty dinosaur who is stuck in his ways, and I'm made peace with all of that. What I can't get my head around is watching people use this awesome tool (and it is, admittedly, awesome) to literally just speed up all the mistakes they were already making. Perhaps it is because I am aging, but slowing down and having a think seems more valuable to me now than it ever has, especially when creating something new. AI is powerful and, like any good tool, could be useful in the right hands, but more often than not I see it being used as an accelerant for all the worst parts of product development to appease a market that has suddenly been told they can now pick all three points on the Iron Triangle instead of just two. This makes about as much sense to me as taking a laxitive when you already are suffering diarrhea.
At the end of the day, LLMs are still tools.
LLMs are just tools indeed.
I don’t know what to say to those types anymore. Live and let live I guess.. or in this case, not live I suppose.
Useful talk and meetings all day is fine, but these are just to write busy/billable hours for all these useless folk with no value for the project. And socially I like listening and talking, just not for this.
Convenient way of saying "you're holding it wrong".
But that's boring and you can't build a YouTube audience around it.