> To be honest, I doubt the current state of things will hold much longer. Current AI can automate virtually all of my work as a pre-AI senior and is rapidly closing in on being able to automate the types of work that have emerged post-AI too.
reading HN these days feels like constant gaslighting. this is so far from my reality that i wonder what folks who are saying this are working on
Every time I see one of those comments, I think about two open source libraries that we use at [DAYJOB]. Both recently released a "new version" that is actually a full, ground up rewrite of the entire project.
In both cases, they have:
1. Weird bugs that a human would never write (e.g. doing async tasks in synchronous functions).
2. A migration guide that is so verbose as to be worthless. Think 50+ pages of text that give every minute change equal weight.
The old versions had some warts, but they worked fine. The new versions have been inflicted on us, and it feels like the intellectual version of digging a hole to fill it in. There was absolutely no reason for it.
I'm sure that the people in charge of those projects believe that they're massively more productive, and that what they're doing is good. But it's worse than useless, it's just a waste of time for everyone involved.
So whenever I see someone claim something like "I used to be a 10x developer, now I'm a 100x developer," I have to wonder how much they're just inflicting garbage on everyone else and they just don't have the self awareness to realize. The type of person to show up at a job, crap all over the codebase, and then leave before it catches up with them.
AI is the ultimate generalist. You can make it do a lot of things well by asking the right questions. In a generalist job, many people can ask those questions.
I don't use AI that much. When I do, it's often surprisingly competent in the topics I'm a specialist in, and it can do a lot of good work quickly. But other people in other organizations still continue building upon my work instead of asking AI directly. Apparently because they don't know what to ask.
Nobody knows how long this will last. Hopefully long enough that I can afford to retire, because I'm old enough that I would face ageism if I had to find a different job.
And if we get there, we will probably have a serious political debate on how many digits should there be in the wealth tax rate.
I have become the “AI” guy at my place, unfortunately.
For some reason I can “hold it right” and others cannot and for the life of me I cannot figure out what it is. I am not more intelligent and certainly not an amazing programmer. I currently suspect I am better at separating what is and is not important. I find many people miss the forest for the trees and overindex on irrelevant details. The more technical they are the stronger this tendency seems to be.
I regularly, like almost daily, encounter colleagues claiming it cannot do something I either did the day before or will just demonstrate right in front of them. The response is usually rationalizing and claiming it wasn’t “that hard to begin with” AKA shifting goalposts.
I have never once encountered a work problem which I truly found impossible to automate, which is insane, but the institutional inertia is not worth overcoming. Devs are a thorny bunch and usually think they are special. I am a recovering one.
> For some reason I can “hold it right” and others cannot and for the life of me I cannot figure out what it is.
I don't think we talk about this as much as we should. I've been interviewing devs as part of [DAYJOB] and there's a lot of people who cannot get anything useful out of Claude. Forget features, we're just talking bugfixes.
I also don't know what's different about the behavior. My best guess is people are overprompting, adding a bunch of noise that interferes with what they're trying to achieve.
I don't see a rude awakening for a technically competent person not having adapted to this year's technology yet. It's pretty weird to expect a functioning development process to suddenly stop working or malfunction just because an even cheaper or faster development process could hypothetically be adopted today.
I think we might already be there with security. Exploits, hack attempts are up and likely continue to stay high. Traditional development process could work but it’s probably more engineers than management wants to spend on a cost center.
I roll my eyes every time I read these comments as well. Can't imagine saying something like this without being an absolute moron with an incredibly underdeveloped model of how the world, and business specifically, works.
> this is so far from my reality that i wonder what folks who are saying this are working on
HN these days is mostly literal bots, AI company shills or people working on some shovelware AI slopware themselves, so they constantly throw out these generic statement as if it's fact. I've been reading this same crap here for the last 3 years, how "Wow, now with Newest Model Mk. 3 (tm) from Anthropic (R), everything is automated!" with very little variation.
It's getting pretty tiring listening to the shilling to be honest. The best part is that when you actually look into any of this crap, or read any of the comments [0] from people in the industry you realize how fucking amateurish and ignorant they all are
Idk, I have my fair-share of copy-paste CRUD work and I have yet to see good work come out the other end outside of a few very very isolated and constrained examples. And before "you're holding it wrong", I am constantly having to firefight and rework code produced by the most AI-pilled people at work.
Nothing devs do is interesting or hard. “Rocket trajectory optimization” is not dev, it’s math. Dev work is very boring and very simple.
I regularly deal with people that think they have amazingly hard jobs and it usually boils down to it being undocumented, arcane and/or math-y. The software itself is always uninteresting (as it should be).
Everything is “CRUD”. Name something that does not amount to taking input, performing a transformation and producing output. I am so tired of this overused thought-stopping cliche.
As if math has no interesting problems to solve, and training data exists for every possible scenario.
> Everything is “CRUD”. Name something that does not amount to taking input, performing a transformation and producing output.
This is true in the "Reductio ad absurdum" sense, but software has to be written to drive hardware that interacts with the real world. Physics is arcane and/or math-y, and many of the problems that are still outstanding in the field of software engineering (or engineering as a whole) are a result of this. The world is full of unknowns, and inferring how those work with a pattern-matcher will only get you so far.
no. this is a national security failure, and reflects poorly on the administration considering there's reporting they've known for months and still had huang in close proximity to the white house.
china has to resort to honeypots because they have few legitimate political tools to influence decisions of an american company. not so for Trump et al - there are much more effective ways for the president to coerce a ceo than telling the world they let a potentially compromised individual help make policy decisions lol
love this comment. the ceo of one of the most influential companies in the world gets caught sleeping with a spy of a rival hegemonic nation state during a period of intense technological, political, and economical change, and OP is like "pfft nothing to see here". amazing
Everyone is talking about their setup and how to manage 3+ workstreams and context switching, and so on. And here i am finishing one thing at a time, mostly writing code by hand, completing tasks faster than everyone on my team.
it really feels like everyone around me in this profession has lost their mind.
is anyone else weirded out by folks chanting this like a mantra? imagine saying this for any other emergent tech. "NFTs are inevitable", "crypto is inevitable", "excel is inevitable", etc.
I want to live a life that makes the people I love happy, makes me feel healthy and happy, and hopefully learn some cool things and skills along the way.
Bending the universe sounds predatory, like exerting what you want onto other people without thinking about what they want
Do you say that to yourself every time you stream music?
What about the labor you steal from the milkmen, butter churners, seamstresses, and elevator operators? How evil of you.
And yeah, I did it. Not you. That money would be in your bank account if you did it. But you didn't.
I personally make these things. I'm not ashamed in the least to say it, because it's true.
It's funny how I'm working around the clock using AI. Because you seem to say this isn't work. And yet I've never been so busy. So can you tell me what's going on here?
How did I manage to build so many useful things for people?
I've had over 40 million downloads of my app recently. Maybe you should think about how people value things instead of relying on your own value system. Your priors might be flawed when it comes to how the rest of society values things.
Boss it would have been easier to just link me to your github.
I read a bit your post history,
you're worse than youtube ad clickbait - you never say what you do, just flowery and bombastic language about how great whatever your vague thing is. (Thank you to sibling poster who spilled the beans, see its not too hard, just say " i made a photoshop clone" its less words than what you typed. )
And boss, I was born out of my comfort zone and never really found one, so you don't scare me one drop.
> And yeah, I did it. Not you. That money would be in your bank account if you did it. But you didn't.
> I personally make these things. I'm not ashamed in the least to say it, because it's true.
Weird. At the very least I thought you had thousands of contributors. I guess it's a good thing you're the one getting paid since apparently you're the only one who matters. Wild that you're so invested in AI when apparently it doesn't apparently do any of the work either.
Maybe you should stop jerking off on HN and get to fixing the 2k+ issues on the photoshop clone you "one-shotted" a month ago. Maybe update the prompt to say "pretty please don't make any mistakes."
The way your half-assed broken slop is taking attention away from actual, functional open source projects is a crime.
Please stop being so toxic. Why is it that you always harass me on HN?
> Weird. At the very least I thought you had thousands of contributors.
We do, and they're awesome.
The paid app I built prior to the Craft apps got to millions in revenue off of my work. Although an anti-AI person like you would say that wasn't my work and that it the revenue just created itself.
> I guess it's a good thing you're the one getting paid since apparently you're the only one who matters.
We're going to build open source infrastructure and weights, and we'll need to earn money to subsidize the build out.
> Wild that you're so invested in AI when apparently it doesn't apparently do any of the work either.
I'm working harder than I've ever worked in my life. You anti-AI people don't understand this because you don't try.
> The way your half-assed broken slop is taking attention away from actual, functional open source projects is a crime.
We are open source, and we're working on adapting open source to the software factory era.
Software factories are the future, and engineers steer the factories.
I'm not harassing you, I certainly don't "always" harass you.
I am criticizing you, and I'm far from the only person doing so.
If your ego can't allow you to see a difference between the two - and your comments suggest it can't - that isn't my problem.
>We do, and they're awesome.
Cool. So you admit you didn't personally, entirely, as a "solo engineer" do it all. That at least is something.
>Although an anti-AI person like you would say that wasn't my work and that it the revenue just created itself.
OK so we're just arguing with strawmen now I guess. Fair enough, it's my fault for engaging. I should know what I'm getting into.
>We're going to build open source infrastructure and weights, and we'll need to earn money to subsidize the build out.
But open source doesn't need that. Gimp doesn't need millions of dollars in investment to subsidize infrastructure. Neither does Krita.
You do because you've created a "software factory" that requires the expense of burning tokens, which of course most open source projects can't afford.
It doesn't seem like you're building open source, it seems like you're building a moat.
>I'm working harder than I've ever worked in my life. You anti-AI people don't understand this because you don't try.
Brother everything you do - or rather everything your AI does and that you take credit for - is built on our work. The work that you mock, deride, and gleefully try to undermine while bragging about the money you make doing so.
At some point the hype is going to die and people are going to judge your products on their actual stability functionality. So far everyone who reviews them at the very least finds them fundamentally broken, and there is no professional adoption of any significance, at least not outside of the AI sphere.
The arrogance is astounding coming from someone whose photoshop clone can't even open a photoshop file.
The whole article is worth reading, but this part in particular caught my eye:
"Despite billions of dollars that firms around the globe have invested in AI over the past few years, major returns remain to be seen: 90% of executives say AI efforts have yet to translate to productivity gains at their companies"
I suppose that depends on tooling as well, and the user. Two scenarios:
1. Throw your entire 200k SLOC codebase in and ask for a bug fix.
2. Tokenize your codebase (AST -> SQLite) and have the LLM explore and reason about subgraphs, and then ask for a bug fix.
One takes expertise but yields a better signal to noise ratio. Though, at that level of expertise one encounters what’s effectively the library of babel: you get back what you write and / or already know, and coherent text outlining future events might be wrong until they transpire
the actual answer to this is accreditation. you will see safety standards rise across the board if engineers risk losing their license for shipping unsafe features.
one only needs to open a history book to see that this country used to be much, much more corrupt. while this country has problems, we are far from the first generation to face them.
Mind being more specific? You have a rapist, demented president who previously failed to coup his way back to power, who has gutted all institutions and is siphoning as much wealth as possible to the pockets of himself and his cronies. You have open financial corruption of the whole administration by industry leaders. You can now literally buy pardons and citizenship. The Supreme Court is a full-on political activist court, and created a whole concept of executive immunity that never existed before, to protect Trump. Every single week is a new scandal by this administration that should be career ending.
Please educate me, when was the country more corrupt?
reading HN these days feels like constant gaslighting. this is so far from my reality that i wonder what folks who are saying this are working on
reply