"AI" is a failed technology, in the sense that it's neither fit for purpose, nor a likely path to its stated end goal of "AGI", but a lot of people's profits currently depend on pretending it's not.
-
@crazyeddie The AI bros have been selling the line for a while now that readable source code is nice to have, but not particularly important, because of how fast you can start over with the same prompts. So like, basically the prompts are the real source code. The fact that they give different results from one run to another, or because of a change to the LLM, is just a nuisance we deal with in order to reap the wonderful benefits of vibecoding -- obviously, having to run the prompt twice is the same as when you had to delete build artifacts and recompile your whole app, sometimes, in the old ways.
So @skavish thinks @Gargron is a fool for reviewing the actual code, when he should be looking at CLAUDE.md and maybe some tests or whatever
@clayote this is a strawman. I never said readable code doesn't matter, that prompts are the source code or that reviewing code is pointless. quite the opposite.
I use AI agents to build real software (bigger codebase than mastodon btw), and code review, tests and engineering discipline are still essential. the difference is that I don't assume the output has to be spaghetti just because an agent wrote it.
you are arguing against a caricature of how you imagine people use these tools, rather than how they can actually be used.
-
"AI" is a failed technology, in the sense that it's neither fit for purpose, nor a likely path to its stated end goal of "AGI", but a lot of people's profits currently depend on pretending it's not. I am curious to see how long the hype will last once the Nvidia ouroboros ends and money stops being pumped into marketing.
@Gargron I like to ask pro-AI folks, "Where is your closest data center?" "How many acres of land did it use?" and "How much electricity will it consume?"
Many will not know the answers.
Here is a drone fly-over of a data center that's currently under construction and about 30 minutes from me. It answers many of the details everyone should be familiar with, whether you live by one, or simply utilize the technology.
https://youtu.be/52x2qBh4vsM?si=-9w4JBOBP3h1wlgO -
@skavish @Gargron It's weird to note that you have admitted he has expertise in software engineering, which would give him all the qualifications in the world to be able to tell spaghetti code from any other form of pasta ... and yet you seem to insist that if it's made by AI he's NOT qualified to make such a judgment.
Which means you think AI produced code is magic somehow and is not in fact spaghetti when it looks like spaghetti to a software engineering expert.
@crazyeddie you are missing the point. his engineering expertise isn't in question. he explicitly said he's never used AI. so how can he confidently judge what AI-generated code looks like when the tools are used effectively?
-
@clayote this is a strawman. I never said readable code doesn't matter, that prompts are the source code or that reviewing code is pointless. quite the opposite.
I use AI agents to build real software (bigger codebase than mastodon btw), and code review, tests and engineering discipline are still essential. the difference is that I don't assume the output has to be spaghetti just because an agent wrote it.
you are arguing against a caricature of how you imagine people use these tools, rather than how they can actually be used.
-
Dunno, I think being "a maintainer and engineer on one of the largest open-source software projects" gives Linus expertise on whether or not AI is useful for working on such projects.
CC: @imhotep@infosec.exchange @skavish@welley.social
-
@Gargron counterpoint: we are well past the point of no return for agentic engineering. There will always be purists who insist on writing code by hand but we aren’t returning to that being the default even after the bubble pops. And I think quiet use of AI as glue will propagate more or less unseen through a lot of tech until it is the expected/desired thing vs a bolted-on buzzword.
Are there myriad problems with both of these? Of course. But that much is not “failure” in any meaningful sense.
@mellowfish @Gargron You really should clarify what “past the point of no return” means to you.
Agentic LLMs for code generation (“engineering” is far too generous a term) will continue to EXIST, yes, because we can’t un-invent something. But I think its impact on software work is immensely overblown. Just as we’ve always had shitty, unmaintainable software, LLMs will continue to ensure that we will have shitty, unmaintainable software, but at an even higher volume. A lot of code doesn’t need to be anything more than shitty, unmaintainable code to be somewhat useful (cf: *motions at shitty software that already surrounds us*), but that doesn’t make them suddenly not shitty and unmaintainable, and generally a poor business investment for people who depend on reliable, maintainable software.
I think this makes a good counter-counterpoint: https://deadsimpletech.com/blog/llms-arent-inevitable
-
if you’ve never used it, then yes, you’re not in a position to judge what it looks like when used properly. being an experienced engineer and maintainer gives you plenty of expertise in software engineering, but it doesn’t magically give you first-hand experience with a tool you explicitly said you’ve never used.
you can absolutely argue that you don’t want to use it. that’s a different claim.
Dude, never used cocaine, but can tell you for certain that it is bad, and will mess up your life. Please stop pushing your poison on us.
-
@clayote LLM output depends heavily on who is using it. a bad engineer will produce bad code with an LLM, just faster. a great engineer can use it to produce great code, faster.
@Gargron may have reviewed plenty of LLM-generated code. that doesn't mean he has seen what these tools can do in the hands of someone who knows how to use them. that is precisely my point: if you have never used the tools yourself, you are judging them from the outside, without understanding what you're missing.
-
@mastodonmigration cocaine is a substance with well-documented effects. AI is a tool whose results depend heavily on how it's used. the analogy doesn't work.
-
@clayote LLM output depends heavily on who is using it. a bad engineer will produce bad code with an LLM, just faster. a great engineer can use it to produce great code, faster.
@Gargron may have reviewed plenty of LLM-generated code. that doesn't mean he has seen what these tools can do in the hands of someone who knows how to use them. that is precisely my point: if you have never used the tools yourself, you are judging them from the outside, without understanding what you're missing.
-
It isn't failed, it works quite well in many cases as a matter of fact. Reverse-engineering example: https://www.youtube.com/watch?v=wIe3eDfGKUo
The marketing is annoying, but the tech has legitimate usecases and is here to stay (perhaps not to the extent the companies want us to believe).@golemwire Where is *your* closest mega-data center? How much electricity does it use? How much land was conceded to build that AI data center? What is the dB level hum at the edge of the facility? How much water is needed to cool it?
Let's start with that first, before we start swinging our pro- AI dicks around. -
@mastodonmigration cocaine is a substance with well-documented effects. AI is a tool whose results depend heavily on how it's used. the analogy doesn't work.
You sound like a cocaine user
-
@Gargron
I am fascinated by your ad hominem attack, but I'm not here to score points or win a right-or-wrong debate.Why not address the substance of my response and engage in a fair debate based on facts and experiences?
Let me start:
What if a technology doesn't need to provide 100% reliable information to be useful, provided its limitations are understood and appropriately managed?And what if future developments improve the reliability of these models and address some of their current limitations? Would AI still be inherently dangerous and irresponsible as a technological tool?
I'm genuinely interested in the reasoning behind your conclusion that AI is a failed technology.
Sea Lion
️ -
@mellowfish That's... just your opinion. Agentic coding (I don't think you can call it "engineering") is heavily subsidized by big AI companies right now, and that's going away sooner or later. The software world will have a painful wake up call from all the spaghetti code being pumped out right now.
@Gargron @mellowfish I feel like some people think if and when the bubble pops, that major corporations are all universally going to completely stop investing in AI because it is "so bad" and AI is just going to disappear. Feels like wishful thinking. I am not an AI fan generally, but I see things that it is useful for and I also see how far it has come in just a handful of years. Two years ago people laughed at the shitty videos that it made. Now it makes a lot of videos that most people would
-
I am pretty sure @Gargron could get great results if he wanted to use these tools. I completely respect his choice not to. but "I don't want to use it" and "it's a bad tool" are two very different things.
-
You sound like a cocaine user
@mastodonmigration and you sound like someone who’s never tried coffee but is absolutely certain everyone who drinks it is an addict
-
I am pretty sure @Gargron could get great results if he wanted to use these tools. I completely respect his choice not to. but "I don't want to use it" and "it's a bad tool" are two very different things.
-
@golemwire Where is *your* closest mega-data center? How much electricity does it use? How much land was conceded to build that AI data center? What is the dB level hum at the edge of the facility? How much water is needed to cool it?
Let's start with that first, before we start swinging our pro- AI dicks around.Not relevant to whether or not it has failed or works poorly, which is what the original post was about
-
@Gargron @mellowfish I feel like some people think if and when the bubble pops, that major corporations are all universally going to completely stop investing in AI because it is "so bad" and AI is just going to disappear. Feels like wishful thinking. I am not an AI fan generally, but I see things that it is useful for and I also see how far it has come in just a handful of years. Two years ago people laughed at the shitty videos that it made. Now it makes a lot of videos that most people would
@Gargron @mellowfish never be able to identify as AI. In two years. Two years is practically nothing even in the context of the relatively short history of the whole of computer science. Where will it be in another 5 years? Another 10. Part of me wishes this AI tech had never been invented. But I think it is only going to continue to get better and it is most likely here to stay in some form or another.
-
@skavish I understand this is hard for you to hear, because you like to use LLMs. But it's the sort of criticism you have to deal with if you want your opinions on your craft to be taken seriously.
"You could probably do good things with them if you wanted to" is such an embarrassing argument to make. You're admitting you have no evidence, and are appealing to the assumed egotism of @Gargron to get him to make your own damn argument for you. Take the L.
