Best case scenario.
-
Best case scenario. You've got an LLM that runs on a AA battery, generates perfect code, never hallucinates, always says it doesn't know when it doesn't know, always cites sources. Every technical problem fixed, every environmental concern addressed. In this scenario, you get full marks for effort.
But you run face-first into the unshakable human problem, that the bosses, the capitalist class, will use this tech that you're so proud of to amass even more wealth at the expense of skilled labour. It will not be used to liberate people, but to oppress them. There's no technical solution to that problem.
-
Best case scenario. You've got an LLM that runs on a AA battery, generates perfect code, never hallucinates, always says it doesn't know when it doesn't know, always cites sources. Every technical problem fixed, every environmental concern addressed. In this scenario, you get full marks for effort.
But you run face-first into the unshakable human problem, that the bosses, the capitalist class, will use this tech that you're so proud of to amass even more wealth at the expense of skilled labour. It will not be used to liberate people, but to oppress them. There's no technical solution to that problem.
@mos_8502 precisely that!
Because the problem os what humans do other humans.
-
Best case scenario. You've got an LLM that runs on a AA battery, generates perfect code, never hallucinates, always says it doesn't know when it doesn't know, always cites sources. Every technical problem fixed, every environmental concern addressed. In this scenario, you get full marks for effort.
But you run face-first into the unshakable human problem, that the bosses, the capitalist class, will use this tech that you're so proud of to amass even more wealth at the expense of skilled labour. It will not be used to liberate people, but to oppress them. There's no technical solution to that problem.
@mos_8502 And in that best-case scenario; what do we do with our days?
Life doesn't have to be struggle, but it needs interest and purpose. If you have no problems what is there - conversely, if you have no problems, you'll have to invent some to stave off death from boredom
-
@kkarhan No wrong answer here: do you distinguish between "AI slop" and "human product made with the assistance of AI"? Like, for example, if I were to have an AI generate, say, 1% of the code in project X, and wrote the rest myself, is the whole project slop still?
-
Best case scenario. You've got an LLM that runs on a AA battery, generates perfect code, never hallucinates, always says it doesn't know when it doesn't know, always cites sources. Every technical problem fixed, every environmental concern addressed. In this scenario, you get full marks for effort.
But you run face-first into the unshakable human problem, that the bosses, the capitalist class, will use this tech that you're so proud of to amass even more wealth at the expense of skilled labour. It will not be used to liberate people, but to oppress them. There's no technical solution to that problem.
@mos_8502
agree. I hear no one talking about it, but LLMs are only about slavery, not "improvement." They're just looking for the cheapest human they can exploit legally, and if that means a crappy silicon approximation that gets it wrong 40% of the time, it's still better than whoever they bought before.
Infinite competent soulless slaves would be the worst-case scenario. -
@kkarhan No wrong answer here: do you distinguish between "AI slop" and "human product made with the assistance of AI"? Like, for example, if I were to have an AI generate, say, 1% of the code in project X, and wrote the rest myself, is the whole project slop still?
@kkarhan Or better yet, say I feed the code for project X into ChatGPT (I wrote it, so there's no copyright problem there), and tell it to critique the code, and I take its feedback into account as I make fixes, but don't let it actually generate code. Is that slop?
-
@mos_8502
agree. I hear no one talking about it, but LLMs are only about slavery, not "improvement." They're just looking for the cheapest human they can exploit legally, and if that means a crappy silicon approximation that gets it wrong 40% of the time, it's still better than whoever they bought before.
Infinite competent soulless slaves would be the worst-case scenario.@pa I'm personally a lot less worried about lazy people who wouldn't pay an artist anyway using ChatGPT to generate mediocre images than I am about skilled labour being laid off permanently because management is happier with a machine that's not as good but doesn't want to form a union.
-
@mos_8502 And in that best-case scenario; what do we do with our days?
Life doesn't have to be struggle, but it needs interest and purpose. If you have no problems what is there - conversely, if you have no problems, you'll have to invent some to stave off death from boredom
@brad In that scenario, I can think of a lot of productive ways to use that "perfect" LLM for my own shit. What terrifies me is what the bosses will do with it.
-
@kkarhan Or better yet, say I feed the code for project X into ChatGPT (I wrote it, so there's no copyright problem there), and tell it to critique the code, and I take its feedback into account as I make fixes, but don't let it actually generate code. Is that slop?
@mos_8502 I'm not a lawyer, but I don't tolerate any #AI code whatosever in my projects because I demand reproduceability, maintainability and auditability.
And the best any #AIslop machine can do is copy & paste templates that users could've setup in their IDE instead.
- Cuz if you know how to code, then you'll inevitably be faster typing it or copypasting in any bs.
-
R relay@relay.infosec.exchange shared this topic
