I ran across this video the other day that I thought did a great job of summing up all the reasons why I intend never to touch GenAI bullshit.
And it's from a former "vibe coder". I'll definitely be keeping this in my back pocket to share with people who ask me why I eschew GenAI.
this is a disaster that can’t continue forever
I thought the same about blockchain. And to an extent, I was right. Nobody's talking about Bitcoin or Etherium (let alone NFTs) any more except the major weirdos who are still stuck in their own weird echo chamber. But the price of Bitcoin is still alarmingly high for something with literally no value. Largely because it's being propped up by despots.
But more than that, I think the AI bubble is basically just a continuation of the blockchain bubble. Several companies with warehouses full of graphics cards eventually found it unprofitable to mine cryptocurrencies. Needing a way to salvage their profits, they took stock of their assets (lots of graphics cards) and considered how to salvage their situation. A stupid scammy bubble worked the first time. Why not a second time? And thus the AI bubble was born. A bubble which made use of... lots of graphics cards.
You're right that the AI bubble can't exist forever in its current form. But who knows what's next? Quantum computers? (That's already kindof happening.) Human brain cells in a jar used for computing? (Wired just had an article about that.) Dark factories? Some bullshit yet to be pulled from someone's ass? (Not that all of these things are entirely bullshit. Some of them probably have more potential to be genuinely profitable than GenAI. But they also definitely have the potential to be sold in half-baked (or less-than-half-baked) forms atop a mountain of unjustified hype. And to be made the new cornerstone of an even more precarious economy.)
Blockchain was a bubble. And it's... less influential than it was. (Not dead, but fizzling out.) But it didn't really "pop" the way you'd expect a bubble to pop. It turned into another bubble. My worry is that the AI bubble won't pop. It'll pivot. And it won't be the same. (I don't really fear that I'm going to be forced by my employer to start supporting Bitcoin payments on our e-commerce site.) But it probably won't be better in any real sense.
I fear we may be living in a "bubble bubble". As in, the commodity that is being greatly overvalued and propped up by despotic governments is no longer a specific niche industry, but rather the production of bubbles themselves. It's gone meta. (And I don't mean Zuck. Well, yes I do. But you know what I mean.)
(Sometimes when I comment in this thread, folks ask me "is this just you speaking or is it the official opinion of /c/fuck_ai?" Just to head that question off, this is just me talking on my own behalf. Not on behalf of the community.)
I think there's a very real threat this "bubble" could lead to mass surveillance after it pops. As for the bubbleception, yeah, I think it has to be either late stage capitalism, or the start of a general resource crisis.
Would be interesting to get a follow-up, on how management responded. Oftentimes it's difficult to quit AI at the job, because the employer will just demand you use AI and be "10x more productive". And they're the boss and it's their project.
But yeah, I think we got several reports by now, how AI coding can be very demoralizing for software developers.
Yeah, my current job isn't pressuring people on my team to use GenAI (yet), but unfortunately I'm going to have to change jobs soon-ish and I definitely worry about being pressured or forced into using it. :(
Also, another dev team at my company that my team works pretty closely with has been relegated basically entirely to the role of code reviewing slop committed by a third party contractor company. :( :(
well everything is turning to shit and they are losing their ass. All these dumb ass companies are gonna get wiped out when the market finnaly crashes.... it comming hold on. It is almost over buttttt.... thats when the draft kicks in.
I consider myself a mediocre developer, I have a security+ and a year of experience, before getting my bachelor's. I am by no means an expert in coding or such, but even then, I can't stress just the number of completely abysmal experiences I've had with AI coding tools, there are dozens of individual examples I can list but it's easier to just talk about the central problems instead.
it writes incredibly inefficient and/or broken code, following almost no coding standards or guidelines, it will also gladly take a broken foundation and build on it.
context window is either lost or slowly poisoned, the "best" models(up to 1m tokens) can barely keep up with re-writing single files, since it needs to keep the whole file in-context before editing, let alone whole project (20+ files cohesive changes are basically still a fantasy)
"suggestions" during development are just hyper idealistic google fetch results, it retrieves anything in the first page mildly related to some keywords and tells you "so, those are basically your options" otherwise it just flat-out refuses or massively overcomplicates certain implementations
incapable of noticing errors or debugging proactively, it is essentially blind to all coding errors unless you keep referencing or correcting the results, you must now sit there and QA and babysit
incapable of learning or improving, you can tell it in detail about a mistake it made, it will sincerely apologize, then moments later repeat the exact same mistake
The list can go on forever but the last one alone would be a fireable offense for any employee in a company, I'm not sure why this shitbox is getting preferential treatment, but it's whatever I guess
It 100% does all of the things you're saying. Using these tools to produce things that are not 100% rubbish requires: (1) for you to be a babysitter, (2) for you to set standards early on using automated checks, and (3) for you to generally understand on a large scale how a software project typically runs.
That's where I just don't see software developers being obsolete even with these tools. Managers and directors will not be able to produce anything but garbage that doesn't work using them.
A big issue IMO is that entry level programmers are likely also not going to be capable of running them with much success.
It 100% does all of the things you're saying.
Just fyi, the wording of this particular line is confusing (to me). It could be interpreted as you disagreeing, when the rest of your comment suggests otherwise
It's weird to me how inconsistent people's reported experience with AI coding tools can be. Someone else in this thread said it "feels like developing software with a cheat code enabled using claude code".
I'm glad for my own sake to be able to say that I don't have enough experience with LLMs to know which would be closer to my experience. (Though if I had to guess, I can't imagine I'd ever feel subjectively like I was having a "good" experience with it.)
But I do wonder what the difference is between the different experiences. I suppose it could be some tools are "better" than others, but I doubt it. Maybe some people are more skilled at bending LLMs to their will. But if I had to bet, (and not to say this is necessarily the case about aesthelete specifically), there's a certain amount of evidence that people who have "good experiences" with LLMs may be self-deluded. The existence of "AI psychosis" points that direction. But I've heard "self-delusion" used as an explanation for why GenAI tends to make a lot of people less productive while making them feel more productive.
I really wish Programming. dev werent so pro-AI, makes it harder to talk about issues like these to other programmers.
I’m on programming.dev. I don’t usually pay much attention to which host people are posting from but mostly when I’m reading Lemmy from here I see just waves and waves of “this slop is slop” posts and very few “but no it’s useful” responses.
Almost no posts reach me that start out as pro-AI.
The admins sent me a warning about being anti-AI so I told them to go fuck themselves and I deleted my account on their instance. I had multiple comments removed before that under my own posts for arguing with sloppers who took it upon themselves to be slop missionaries and downvote brigade me.
Well, I mean, telling people to go fuck themselves is not the usual way to really connect with people and get a message communicated, exactly…
At least not most messages.
Go fuck yourself
Ah. I didn't even realize they were. I've blocked all the pro-GenAI communities I've run across, so it may be that it's been so long since I had enough exposure to any of the pro-GenAI content on programming.dev that I've forgotten about programming.dev being all pro-GenAI.
More than half of job postings I seen now require proficiency with AI agents and I have never used one. Our company has copilot set up to do code reviews (after which humans also review it), but I will likely be out of the job soon. It's killing me.
🫂
I get it. I really, really do. But if the tools stick around and remain affordable (and the if is load bearing) I think it will be difficult to have this stance and remain in employment in modern software companies.
Maybe there will be anti-AI software shops or you can work in places that are entirely backwards (I mean I once worked at a place where a developer was basically maintaining what would normally be a database using a series of Java collections so ... not everyone is going to be on the "cutting edge").
The posters' qualms are exactly right. You become more of like a business analyst and/or tester rather than a programmer when writing code using these tools. However, that's what will be needed for the creation and maintenance of software if these tools stick around, remain in use, and remain affordable.
It 100% feels like developing software with a cheat code enabled using claude code and friends and I think it likely leads to apathy about what you're developing in probably most cases, but 🤷. Most software written before this was seemingly made by people who didn't care at all about the people using it so I'm not sure that changes much of anything.
I do, however, think the "my website has a chatbot" and "try to one shot a bug fix using my token workflow" things that have sprung up everywhere are entirely useless and will quickly find a graveyard when the bubble bursts.
@lemmy.world
"We did it, Patrick! We made a technological breakthrough!"
A place for all those who loathe AI to discuss things, post articles, and ridicule the AI hype. Proud supporter of working people. And proud booer of SXSW 2024.
AI, in this case, refers to LLMs, GPT technology, and anything listed as "AI" meant to increase market valuations.
go to feed...
@lemmy.world
"We did it, Patrick! We made a technological breakthrough!"
A place for all those who loathe AI to discuss things, post articles, and ridicule the AI hype. Proud supporter of working people. And proud booer of SXSW 2024.
AI, in this case, refers to LLMs, GPT technology, and anything listed as "AI" meant to increase market valuations.
go to feed...
I think people shouldn't worry too much about it because this is tied to a real financial crisis that is more concerning.
OpenAI and Anthropic are losing fuck tons of money and, even though they have hyped the fuck out of some desperate losers, they will run out of investor funds.
Purely from a financial perspective, this is a disaster that can't continue forever. Maybe not even another year.
That doesn't mean LLM tools will disappear... But they will be priced accordingly. You want to slop some app up? Great, but you'll pay for it, and likely still need engineers to read and understand that code and not rely on slop tooling to do so.
The other side to this is a generation of parents all told their kids "just get a STEM degree, preferably learn to code" because they saw programmers living well, and then a ton of people who don't care all that much about code flooded the tech industry and are learning maybe they don't want to be software engineers.
Compound that with slop tooling that is priced below cost right now and a looming financial crisis, and you have a financial disaster that will hit hard and change the industry forever regardless of whether you choose to use LLM Tools or not.
I think there will be a future for those who use LLM tools and those that don't, and it will look like a different job, and be preferable for different industries in different ways, and LLM will be a lot more carefully used due to cost being normalized. No more, "implement mergesort". People will eventually get looked down on for using it for basic shit. It will be expensive.
Maybe some corporation wants an iPhone app but doesn't want to build a whole team to make it, and it's skills none of them have. They will consider outsourcing, or slopping it. Slopping will mean quick and dirty code that can put an app together well, as long as you give it a good schema and design up front. It will cost money, but you'll have v1 in a day. And then you won't need that iPhone app team.
That's where I see LLM ending up, some prototyping tool that helps you pump out a dumb app out quickly if provided good instructions up front, and priced accordingly.
save