First of all, there could be other solutions, such as B2B subsidizing individual user plans, or more fine grained model tiering per cost.

Also, yes you can get some access for free, but even today the higher tiers of proprietary models is around $200/mo for individual users, which might still be subsidized but is definitely not free, and is quite a chunk of money at $2400 a year!

I don't know what your setup is at the moment, but it's possible more efficient hardware and stack are available that you're not utilizing. Of course this depends on what models you're trying to run.

I think that smaller models will become a lot better, and hardware will become more optimized as well. We're starting to see this with NPUs and TPUs.

All this means running models will cost less, and maybe upgrading the power grid will also reduce cost of energy, making it more affordable.

I don't see any way that AI will go away because it "hits a wall". We have long passed the point of no return.

That's not nothing, but it's still not very much to pay compared to e.g. the cost of a FTE.

I would try setting the GPU to run at a lower power level. I set my GPU power level to 80% and it becomes much quieter, and only runs maybe 5% slower at most.

Also I 100% agree with the rest of your comment. We can only power the current growth we are seeing for so long.

LLMs writing prose is too robotic

LLMs output is too dependent on prompts to be interesting

LLMs take too much RAM to run effectively

LLMs take too much electricity to run locally

LLMs work locally but are a bit too slow for my taste

LLMs output mostly correct code but it isn't applicable to my codebase

LLMs make tool calls to pull in additional context

LLMs outputted code works for most developers but not my codebase <---- you are currently here

The required hardware is fits the budget for a professional developer.

Putting LLMs in the cloud allows the hardware to be utilised better and to have sharding of big models.

What multiple of the current cost do you expect? Currently, GitHub Copilot and ChatGPT for Business cost $19/month and €29/month respectively. Even a 10×–20× increase will still be economically viable in a professional setting if the tools continue to save hours of work.

for some reason it feels like people are under the assumption that hardware isnt going to improve or something?

I 100% share the codependent path too and had written a similar comment some 2-3 days ago but these AI companies which provide are either seriously negative/loss making subsidizing or they are barely net zero. I doubt that it will continue and so the bubble will burst I guess and prices of these will rise perhaps

We will see perhaps something like google which can feed on advertising can perhaps still provide such subsidies for a longer time but the fact of the matter is that I have no alleigance to any model as some might have and I will simply shift to the cheapest thing which can still provide me / be enough for my queries or prototypes mostly I suppose.

LLMs are amazing and they will change everything, and then everything will be changed.

    I can confirm that they are completely useless for real programming

Absolutely incredible tools that have saved hours and hours helping me understand large codebases, brainstorm features, and point out gaps in my implementation or understanding.

I think the main disconnect in the discourse is that there are those pretending they can reliably just write all the software, when anyone using them regularly can clearly see they cannot.

But that doesn't mean they aren't extremely valuable tools in an engineer's arsenal.

Perhaps you're doing some amazing low-level work, but it feels like you're way overestimating how much of our industry does that. A massive amount of developers show up to work every day and just stitch together frameworks and libraries.

In many ways, it feels similar to EVs. Just because EVs aren't yet, and may never be, effective to moving massive amounts of cargo in a day with minimal refueling, doesn't mean that they aren't an effective solution for the bulk of drivers who have an average commute of 40 miles a day.

This is a bit of no-true-scottsman, no? For you "real programming" is "stuff LLMs are bad at," but a lot of us out in the real world are able to effectively extract code that meets the requirements of our day jobs from tossing natural language descriptions into LLMs.

I actually find the rise of LLM coding depressing and morally problematic (re copyright / ownership / license laundering), and on a personal level I feel a lot of nostalgia for the old ways, but I simply can't levy an "it's useless" argument against this stuff with any seriousness.

You and I must have completely different definitions of "real programming". In this very comment, you described a problem that the model solved. The solution may not have involved low-level programming, or discovering a tricky bug entrenched in years-worth of legacy code, but still a legitimate task that you, as a programmer, would've needed to solve otherwise. How is that not "real programming"?

Three months ago I would have agreed with you, but anecdotal evidence says Codex 5.2 and Opus 4.5 are finally there.

I’m in enterprise ERP.

"completely useless" and "real programming" are load bearing here. Without a definition to agree on for those terms, it's really hard not to read that as you're trying to troll us by making a controversial unprovable claim that you know will get people that disagree with you riled up. What's especially fun is that you then get to sneer at the abilities of anybody making concrete claims by saying "that's not real programming".

How tiresome.

Some years? I don't remember any agents being any good at all before just over 1 year ago with Cursor and stuff really didn't take off until Claude Code.

Which isn't to say you weren't working with agent-LLMs before that, but I just don't know how relevant anything but recent experience is.

Can you elaborate on "real programming" ?

I am assuming you mean the simplest hard problem that is solved. The value of the work is measured in those terms. Easy problems have boilerplate solutions and have been solved numerous times in the past. LLMs excel here.

Hard problems require intricate woven layers of logic and abstraction, and LLMs still struggle since they do not have causal models. The value however is in the solution of these kinds of problems since the easy problems are assumed to be solved already.

This was how I felt until about 18 months ago.

Can you give a single, precise example where modern day LLMs fail as woefully as you describe?

so I see what you're saying. he comes up with the wrong answers a lot to a problem involving a group of classes in related files

however it's Continue, so it can read files in vs code which is really nice and that helps a lot with its comprehension so sometimes it does find the issue or at least the nature of the issue

I tend to give it bug n-1 to pre digest while I work on bug n

>They never helped me solve complex problems with low-level libraries. They can not find nontrivial bugs. They don't get the logic of interwoven layers of abstractions.

>LLMs pretend to do this with big confidence and fail miserably.

This is true for most developers as well. The mean software developer, especially if you outsource, has failure modes worse than any LLM and round-trip time is not seconds but days.

The promise of LLMs is not that they solve the single most difficult tasks for you instantly, but that they do the easy stuff well enough that they replace offshore teams.

AI coding agents are analogous to the machine. My job is to get the prompts written, and to do quality control and housekeeping after it runs a cycle. Nonetheless, like all automation, humans are still needed... for now.

That is just not true, assuming you have a modicum of competence (which I assume you do). AIs suck at all these tasks; they are not even as good as an inexperienced human.

You are being fooled by randomness [1]

Not because the models are random, but because you are mistaking a massive combinatorial search over seen patterns for genuine reasoning. Taleb point was about confusing luck for skill. Dont confuse interpolation for understanding.

You can read a Rust book after years of Java, then go build software for an industry that did not exist when you started. Ask any LLM to write a driver for hardware that shipped last month, or model a regulatory framework that just passed... It will confidently hallucinate. You will figure it out. That is the difference between pattern matching and understanding.

[1] https://en.wikipedia.org/wiki/Fooled_by_Randomness

I choose to look at it as an opportunity to spend more time on the interesting problems, and work at a higher level. We used to worry about pointers and memory allocation. Now we will worry less and less about how the code is written and more about the result it built.

However I'm still finding a trend even in my org; better non-AI developers tend to be better at using AI to develop.

AI still forgets requirements.

I'm currently running an experiment where I try to get a design and then execute on an enterprise 'SAAS-replacement' application [0].

AI can spit forth a completely convincing looking overall project plan [1] that has gaps if anyone, even the AI itself, tries to execute on the plan; this is where a proper, experienced developer can step in at the right steps to help out.

IDK if that's the right way to venture into the brave new world, but I am at least doing my best to be at a forefront of how my org is using the tech.

[0] - I figured it was a good exercise for testing limits of both my skills prompting and the AI's capability. I do not expect success.

a car moves faster than you, can last longer than you, and can carry much more than you. But somehow, people don't seem to be scared of cars displacing them(yet)? Perhaps autodriving would in the near future, but there still needs to be someone making decisions on how best to utilize that car - surely, it isn't deciding to go to destination A without someone telling them.

> I feel like I’m doing the work of an entire org that used to need twenty engineers.

and this is great. A combine harvester does the work of what used to be an entire village for a week in a day. More output for less people/resources expended means more wealth produced.

Sure LLMs can churn out code, and they sort of work for developers who already understand code and design, but what happens when that junior dev with no hard experience builds their years of experience with LLMs?

Over time those who actually understand what the LLMs are doing and how to correct the output are replaced by developers who've never learned the hard lessons of writing code line by line. The ability to reason about code gets lost.

This points to the hard problem that the article highlights. The hard problem of software is actually knowing how to write it, which usually takes years, sometimes up to a decade of real experience.

Any idiot can churn out code that doesn't work. But working, effective software takes a lot of skill that LLMs will be stripping people of. Leaving a market there for people who have actually put the time in and understand software.

If you're really able to do the work of a 20 man org on your own, start a business.

For me its not about me or the coding assistant, its me and the coding assistant. But I'm also not a professional coder, i dont identify as a coder. I've been fiddling with programming my whole life, but never had it as title, I've more worked from product side or from stakeholder side, but always got more involved, as I could speak with the dev team.

This also makes it natural for me to work side-by-side with the coding assistant, compared maybe to pure coders, who are used to keeping the coding side to themselves.

They are pretty good at writing code *after* I thoroughly described what to do, step by step. If you miss a small detail they get loose and the end result is a complete mess that takes hours to clean up. This still requires years of coding experience, planning ahead in head, you won't be able to spare that, or replace developers with LLMs. They are like autocomplete on steroids, that's pretty much it.

I mean, AIs can drop something fast the same way you cannot beat a computer at adding or multiplying.

After that, you find mistakes, false positives, code that does not work fully, and the worse part is the last one: code that does not work fully but also, as a consequence, that you do NOT understand yet.

That is where your time shrinks: now you need to review it.

Also, they do not design systems better. Maybe partial pieces. Give them something complex and they will hallucinate worse solutions than what you already know if you have, let us say, over 10 years of experience programming in a language (or mabye 5).

Now multiply this unreliability problem as the code you "AI-generate" grows.

Now you have a system you do not know if it is reliable and that you do not understand to modify. Congrats...

I use AI moderately for the tasks is good at: generate some scripts, give me this small typical function amd I review it.

Review my code: I will discard part of your mistakes and hallucinations as a person that knows well the language and will find maybe a few valuable things.

Also, when reviewing and found problems in my code I saw that the LLMs really need to hallucinate errors that do not exist to justify their help. This is just something LLMs seem to not be accurate at.

Also, when problems go a bit more atypical or past a level of difficulty, it gets much more unreliable.

All in all: you are going to need humans. I do not know how many, I do not know how much they will improve. I just know that they are not reliable and this "generate-fast-unreliable vs now I do not know the codebase" is a fundamental obstacle that I think it is if not very difficult, impossible to workaround.

Absolutely flat out not true.

I'm extremely pro-faster-keyboard, i use the faster keyboards in almost every opportunity i can, i've been amazed by debugging skills (in fairness, i've also been very disappointed many times), i've been bowled over by my faster keyboard's ability to whip out HTML UI's in record time, i've been genuinely impressed by my faster keyboard's ability to flag flaws in PRs i'm reviewing.

All this to say, i see lots of value in faster keyboard's but add all the prompts, skills and hooks you like, explain in as much detail as you like about modularisation, and still "agents" cannot design software as well as a human.

Whatever the underlying mechanism of an LLM (to call it a next token predictor is dismissively underselling its capabilities) it does not have a mechanism to decompose a problem into independently solvable pieces. While that remains true, and i've seen zero precursor of a coming change here - the state of the art today is equiv to having the agent employ a todo list - while this remains true, LLMs cannot design better than humans.

There are many simple CRUD line of business apps where they design well enough (well more accurately stated, the problem is small/simple enough) that it doesn't matter about this lack of design skill in LLMs or agents. But don't confuse that for being able to design software in the more general use case.

They don't do any of that better than me; they do it poorer and faster, but well enough for most of the time.

Seriously. The bar is that low. When people say "AI slop" I just chuckle because it's not "AI" it's everyone. That's the general state of the industry.

So all you have to do is stay engaged, ask questions, and understand the requirements. Know what it is you're building and you'll be fine.

Planning long running projects and deciding are things only you can do well!! Humans manage costs. We look out for our future. We worry. We have excitement, and pride. It wants you to think none of these things matter of course, because it doesn't have them. It says plausible things at random, basically. It can't love, it can't care, it won't persist.

WHATEVER you do don't let it make you forget that it's a bag of words and you are someing almost infinitely more capable, not in spite of human "flaws" like caring, but because of them :)

How has free code, developed by humans, become more available than ever and yet somehow we have had to employ more and more developers? Why didn't we trend toward less developers?

It just doesn't make sense. AI is nothing but a snippet generator, a static analyzer, a linter, a compiler, an LSP, a google search, a copy paste from stackoverflow, all technologies we've had for a long time, all things developers used to have to go without at some point in history.

I don't have the answers.

ChatGPT, is that you?

It is certainly more eloquent than you regarding software architecture (which was a scam all along, but conversation for another time). It will find SOME bugs better than you, that's a given.

Review code better than you? Seriously? What you're using and what you consider code review? Assume I could identify one change broke production and you reviewed the latest commit. I am pinging you and you better answer. Ok, Claude broke production, now what? Can you begin to understand the difference between you and the generative technology? When you hop on the call, you will explain to me with a great deal of details what you know about the system you built, and explain decision making and changes over time. You'll tell about what worked and what didn't. You will tell about the risks, behavior and expectations. About where the code runs, it's dependencies, users, usage patterns, load, CPU usage and memory footprint, you could probably tell what's happening without looking at logs but at metrics. With Claude I get: you're absolutely right! You asked about what it WAS, but I told you about what it WASN'T! MY BAD.

Knowledge requires a soul to experience and this is why you're paid.

Think of yourself as a chef and LLMs as ready to eat meals or a recipe app. Can ready to eat meals OR recipe apps put a chef out of business?

Once you learn that it's mostly about interacting with a customer (sometimes this is yourself), you will realize the AI is pretty awful at handling even the most basic tasks.

Following a product vision, selecting an appropriate architecture and eschewing 3rd party slop are examples of critical areas where these models are either fundamentally incapable or adversely aligned. I find I have to probe ChatGPT very hard to get it to offer a direct implementation of something like a SAML service provider. This isn't a particularly difficult thing to do in a language like C# with all of the built in XML libraries, but the LLM will constantly try to push you to use 3rd party and cloud shit throughout. If you don't have strong internal convictions (vision) about what you really want, it's going to take you for a ride.

One other thing to remember is that our economies are incredibly efficient. The statistical mean of all information in sight of the LLMs likely does not represent much of an arbitrage opportunity at scale. Everyone else has access to the same information. This also means that composing these systems in recursive or agentic styles means you aren't gaining anything. You cannot increase the information content of a system by simply creating another instance of the same system and having it argue with itself. There usually exists some simple prompt that makes a multi agent Rube Goldberg contraption look silly.

> I’m basically just the conductor of all those processes.

"Basically" and "just" are doing some heroic weight lifting here. Effectively conducting all of the things an LLM is good at still requires a lot of experience. Making the constraints live together in one happy place is the hard part. This is why some of us call it "engineering".

Those twenty engineers must not have produced much.

Orchestrating harmony is no mean feat.

What doesn't help is that the current state of AI adoption is heavily top-down. What I mean is the buy-in is coming from the leadership class and the shareholder class, both of whom have the incentive to remove the necessary evil of human beings from their processes. Ironically, these classes are perhaps the least qualified to decide whether generative AI can replace swathes of their workforce without serious unforeseen consequences. To make matters worse, those consequences might be as distal as too many NEETs in the system such that no one can afford to buy their crap anymore; good luck getting anyone focused on making it to the next financial quarter to give a shit about that. And that's really all that matters at the end of the day; what leadership believes, whether or not they are in touch with reality.

What we do know is this. If AI keeps improving at the current rate it’s improving then it will eventually hit a point where we don’t need software engineers. That’s inevitable. The way for it to not happen is for this technology to hit an impenetrable wall.

This wave of AI came so fast that there are still stubborn people who think it’s a stochastic parrot. They missed the boat.

Care to share a few of these examples?

I feel like LLM's are specifically selected to impress people that have a lot of influence. People like investors and CEO's. Because a "AI" that does not impress this section of the population does not get adopted widely.

This is one of the reasons I think AI will never really be an expert as it does not need to be. It only needs to adopt a skill (for example coding) to pass the examination of the groups that decide if it is to be used. It needs to be "good enough to pass".

But so many problems (programming, but also physics and math) start as informal word problems. A major skill is being able to turn the informal into something formal and precise.

Ones that want their application to work? :) the last piece of software you should be knocking is MS Excel, in my 30+ year career the one common thread just about everywhere I worked (or contracted at) has used Excel to run some amazing sh*t

Are they?

I feel like AIs fill in a lot of the blanks with reasonable assumptions rather than the input being precise

We had "free clothes" for years, decades now. I don't mean cheap I mean literally free, as in $0.0 software. Cheaper software isn't new.

Also there are still clothe designers, fashion runways, and expensive Patagonia vests today. The clothing industry is radically different from back then but it's definitely not gone.

Writing code is not at all the most time consuming part of software development.

If llms can churn out software like Toyota churns out cars, AND do maintenance on it, then the craftsmen (developers of today) are going to be displaced.