If you have a job, working for a boss, you're trading your time for money. If you're a contractor and negotiate being paid by the project, you're being paid for results. Trading your time for money is the underlying contract. That's the fundamental nature of a job working for somebody else. You can escape that rat race if you want to.
Someone I know builds websites for clients on a contract basis, and did so without LLMs. Within his market, he knows what a $X,000 website build entails. His clients were paying that rate for a website build out prior to AI-augmented programming, and it would take a week to do that job. With help from LLMs, that same job now takes half as much time. So now he can choose to take on more clients and take home more pay, or not, and be able to take it easy.
So that option is out there, if you can make that leap. (I haven't)
I'm working on it. But it takes money and the overlords definitely are trying to squeeze as of late.
And yes, while I don't think I'm being replaced in months or years, I can a possibility in a decade or two of the ladder being pulled up on most programming jobs. We'll either be treated as well as artists (assuming we still don't unionize) or we'll have to rely on our own abilities to generate value without corporate overlords.
These things eventually always end up as a Red Queen’s race, where you have to run as fast as you can to stay in the same place.
A different friend who's contractor in a non-tech area told me a client of his secretly showed him his competition's bid for the same project. The competition's bid was much higher, and the reason the client showed him that was to get my friend to raise his rates and resubmit his bid.
So you're welcome to try, but as a programmer looking into the abyss, I'm looking at the whole thing as encouragement to develop all those soft skills that I've been neglecting.
How are you so sure?
It’s reasonable that people explore contingencies where the technology does improve to a point of driving changes in the labor market.
If you have your certainty then there are plenty of opportunities to short this space. For the rest of us who don’t have crystal balls, contingency planning (ideally through policy and not as individuals) will have to do.
0.2x, 2x, 5x, 50x?
There can't be things that a human can program that AGI can not program or it is not "AGI".
While I am never a true believer in AGI, it seems to go I get a little faith when a new model comes out then I become increasingly agnostic the weeks and months after that. Repeat.
Don’t hate on it, just spin up some startup with “ai” and LLM hype. Juice that lemon.
[1](https://techcrunch.com/2025/01/07/nvidia-ceo-says-his-ai-chi...)
Anyway, the mumber of tiktok users coorelates with advancements in AI too!
Before tiktok the progress was slower, then when tiktok appeared it progressed as hell!
And it seems, more or less, clear that the rate of change in the state of the art has already sharply decreased. So it's likely LLMs have already entered into this window.
Just increasing compute power will increase the performance/training speed of these models, but you also need to increase the quality of the data that you are training these models on.
Maybe... the reason why these models show a high school level of understanding is because most of the data on the internet that these models have been trained on is of high school graduate quality.
That said, I do think there is real risk of letting AI hinder the growth of Junior dev talent.
How many companies are intelligent given how many dumb decisions we see?
If we assume enough not so intelligent companies then better AI code we lead to mass firing.
Assuming the same kind of growth in capabilities isn't backed by reality.
The last release of OpenAI's model wasn't dramatically better.
At the moment it's more about getting cheaper.
https://www.autoevolution.com/news/viral-tesla-cybertruck-cr...
This is the incorrect assumption, or at least there’s no evidence to support it.
1. All problems are small- the prompt and solution (<100 LOC, often <60LOC)
2. Solving those problems is more about recollecting patterns and less about good new insights. Now, top level human competitors do need original thinking, but that's only because our memory is too small to store all previously seen patterns.
3. Unusually good dataset- you have tens of thousands of problems, each with thousands of submissions, along with clear signals to train on (right/wrong, time taken etc), a very rich discussion sections etc.
I think becoming 100th best Codeforces programmer is still an incredible achievement for a LLM. But for Sam Altman to specifically note the performance on this- I consider that a sign of weakness, not strength.
In the 90s, companies showed graphs of CPU frequency and projected we would be hitting 8ghz pretty soon. Futurists predicted we would get CPUs running at tens of ghz.
We only just now have 5ghz CPUs despite running at 4ghz back in the mid 2000s.
We fundamentally missed an important detail that wasn't consider at all in those projections.
We know less about the theory of how LLMs and neural networks grow with effort than we did about how transistors operate over different speeds.
You utterly cannot extrapolate from those kinds of graphs.
It is only generating based on training data. In mature code bases there is a massive amount of interconnected state that is not already present in any github repository. The new logic you'd want to add is likely something never done before. As other programmers have stated, it seems to be improving at generating useful boilerplate and making simple websites and such related to what's out there en masse on Github. But it can't make any meaningful changes in an extensively matured codebase. Even Claude Sonnet is absolutely hopeless at this. And the requirement before the codebase is "matured" is not very high.
99% of software development jobs are not as groundbreaking as this. It’s mostly companies doing exactly what their competitors are doing. Very few places are actually doing things that an LLM model has truly never seen crawling through GutHub. Even new innovative products generally boil down to the same database fetches and CRUD glue and JSON parsing and front end form filling code.
The simplest version of that is some CGI code a PHP script. Which everyone should be writing according to your description. But why so many books have been written to be able to do this seemingly simple task? So many frameworks, so many patterns, so many methodologies....
It can't do anything in these random Phaser games I'm making and even translating my 10,000 line XNA game to Phaser. It is totally hopeless.
Phaser has been out forever now, and XNA used to be too.
This is not the case anymore, current SOTA CoT models are not just parroting stuff from the training data. And as of today they are not even trained exclusively on publicly (and not so publicly) available stuff, but they massively use synthetic data which the model itself generated or distilled data from other smarter models.
I'm using and I know plenty of people using AI in current "mature" codebases with great results, this doesn't mean it does the work while you sip a coffee (yet)
*NOTE: my evidence for this is that o3 could not break ARC AGI by parroting, because it's a banchmark made exactly for this reason. Not a coding banchmark per se, but still transposable imo.
This is the fundamental delusion that is driving AI hype.
Although scale has made LLMs look like magic, actual magic (AGI) is not on the scaling path. This is a conjecture (as is the converse), but I'm betting the farm on it personally and see LLMs as useful chat bots that augment other, better technologies for automation. If you want to pursue AGI, move on quickly to something structurally and fundamentally better.
People don't understand that AGI is pure speculation. There is no rigorous, non-circular definition of human intelligence, let alone proof that AGI is possible or achievable in any reasonable time frame (like 100 years).
That's a very bold claim. We are already seeing plateu in LLM capabilities in general. And there is little improvement in places where they fall short (like making holistic changes in a large codebase) since their birth. They only improve where they are already good at such as writing small glue programs. Expecting significant breakthroughs with just scaling without any fundamentally changes to the architecture seems like too optimistic to me.
After 20 years in tech, I can't think of a single company I've worked for/with that would fit the profile of an "intelligent" company. All of them make poor and irrational decisions regularly. I think you over-estimate the intelligence of leadership whilst simultaneously under-estimating their greed and eventual ability to self-destruct.
EDIT: you also over-estimate the desire for developers to increase their productivity with AI. I use AI to reduce complexity and give me more breathing room, not to increase my output.
Says nothing about companies and everything about you
> you also over-estimate the desire for developers to increase their productivity with AI. I use AI to reduce complexity and give me more breathing room, not to increase my output.
I'm the same. But I expect that once many begin to do this, there will be some who do use it for productivity and they will set the bar. Then people like you and I will either use it for productivity or fall behind.
>I expect that once many begin to do this, there will be some who do use it for productivity and they will set the bar.
Yeah, probably. I've had companies so pinpointed on "velocoity" instead of quality. I imagine they will definitely try to expect triple the velocity just because one person "gets so much done". Not realizing how much of that illusion is correcting the submissions.
No one is making this claim.
My comment was a bit terse and provocative, rude, deserves the downvotes tbh. I'll take them.
To elaborate ~ I've got a lot of empathy for the poster I was originally replying to. I've fallen into that way of thinking before, and it sure is comfortable. Of course, companies and their leadership make poor and irrational decisions. Often, however, it's easy to perceive their decisions as poor and irrational when you simply don't have the context they do. "Why would they x ?? if only y!!" but, you know, there may well be a good reason why that you aren't aware of, they may have different goals to you (which may well be selfish! and that doesn't make them irrational or anything). Feels similar to programmers hating when people say "can't you 'just' x" - well yes, but actually there's a mountain of additional considerations behind the scene that the person spouting "just" hasn't considered.
Is leadership unintelligent, or displaying poor/irrational decision making, if the company self destructs? Perhaps. But quite possibly not. They probably got a whole lot out of it. Different priorities.
Consider that leadership may label a developer unintelligent if that dev doesn't always consider how to drive shareholder value "gee they're so focused on increasing their salary not on business value". Well actually the dev is quite smart, from their own perspective. Same thing.
And if every company you've ever worked for truly has poor leadership then, yeah, it's probably worth reassessing how you interview. Do you need to dig deeper into the business? Do you just not have the market value to negotiate landing a job at a company with intelligent leadership?
So, two broad perspectives: either the poster has a challenge with perception, or they are poor at picking companies. Or perhaps the companies truly do have poor leadership but I think that unlikely. Hence it comes back to the individual.
@y-c-o-m-b sorry for being a bit rude.
Cheers for reading
>And if every company you've ever worked for truly has poor leadership then, yeah, it's probably worth reassessing how you interview.
No need. I work in games. There isn't a major studio in the industry that isn't like this. An industry used to churning workers and releasing them the moment the project ends.
In some ways it's a path I chose, but at the same time it means I need to be more cynical to defend myself from their inevitably orthogonal actions. I have an exit plan, but I need more time and money first.
If the industry wasnt so secretive with its techniques and knowledge, maybe I could have side stepped it altogether. But alas.
I've worked in big tech for a combined total of 6 years. Several of the other companies are Fortune 500 members. I've also worked at mid-size and small companies across mortgage, healthcare, fin-tech, point of sale, HR/payroll, and more. I surmise that you would categorize most of these as "intelligent" companies, which negates your argument about this being a "me" problem. Let's take Intel - where I worked for 3 years - as an example. Some would consider this an "intelligent company". I was there when Brian Krzanich took over for Paul Otellini. I think you will have a very easy time finding huge swaths of people/employees that consider Krzanich's leadership decisions to be very poor. In fact over the last few months, you'll find threads here on HN that directly pin the decline of Intel on his decision making.
We can argue the semantics of "intelligent" all day and make excuses for why leaders make irrational choices, but my point still stands. I don't think this is a "me" problem for one simple reason: If you take me out of the equation, the issue still exists.
Just look at the over hiring during covid and the methods used to cull that workforce after they realized their mistake. Back handed and inhumane. Executives are more followers than a junior dev is. They just have a lot more terminology to obscure that fact. But they are basically professional bullshitters, like consultant firms.
This is excluding executives with vision. But the market and corporate structure bias towards eliminating those leaders as they are not consistently profitable over every month.
self interested and short sighted says nothing about intelligence, irrationality, or poor decision making. Back handed and inhumane covid hiring and firing is probably not a mistake from their perspective. professional bullshitting is a form of intelligence (I hate it too, I've been done in by it, but I respect it)
It's not even necessarily about intelligence but about the simple concept of unknown unknowns. If everyone had perfect knowledge of the current reality and could perfectly describe what it is that they want immediately, without spending any time investigating, producing proof-of-concept work, iterating on a product, etc. I would agree that it could be feasible for AI to replace a lot of programming work.
As it stands, what I just described above is THE BULK of development. Coding is the last thing that happens and it also happens to be the fastest, easiest and smallest part of the entire process.
Not my experience. I spend as much time reading through and replacing wrong AI generated code as I do writing my own code, so it's really wasting my time more often than helping. It's really hit or miss, and about the only thing the AI gets right most often is writing console.log statements based on the variable I've just assigned, and that isn't really "coding". And even then it gets it right only about 75% of the time. Sure, that saves me some time, but I'm not seeing the supposed acceleration AI is hyped as giving.
Where are your facts? What is your basis on these future prediction? Sure small code snippets have already improved since gpt2. What about larger applications where layers and layers of abstractions coming from private company sensitive data?
> At the end of the day product is just one area of company development, build the complete e2e ultimate solution with 0 distribution or marketing will not help.
this sounds like PR garbage to me why use "e2e ultimate solution" provide technical details so we can verify what you're saying is true
1. AI code ability will be the same as is today
2. Companies will replace people for AI en masse at a given moment in time
Of course both these assumptions are wrong, the quality of code produced by AI will improve dramatically as model evolves. And is not even just the model itself. The tooling, the Agentic capabilities and workflow will entirely change to adapt to this. (Already doing)
The second assumption is also wrong, intelligent companies will not layoff en masse to use AI only, they will most likely slow hiring devs because their existing enhanced devs using AI will suffice enough to their coding related needs. At the end of the day product is just one area of company development, build the complete e2e ultimate solution with 0 distribution or marketing will not help.
This article, in my opinion, is just doomerism storytelling for nostalgic programmers, that see programming only as some kind of magical artistic craft and AI as the villain arrived to remove all the fun from it. You can still switch off Cursor and write donut.c if you enjoy doing it.