MercuryCobra
Attorney. He/him. Anti-fascist.
- this is just head in the sand shit man, I'm sorry. like it is unambiguously useful for many things, which is why it's such a potential problem. we wouldn't be concerned about it's impact on teaching and learning if it couldn't spit out a decent approximation of undergrad writing.
-
View full threadAgain, I would not describe “a machine that makes it easy to do crime and cheat” as a “useful” machine *even if it is objectively useful to criminals and cheaters.* That cheaters and criminals have flocked to LLMs in order to cheat and do crime is not evidence of its utility
- But more fundamentally where is this debate taking us? We agree about what LLMs can do, what they can’t, why they’re bad and why they’re (in my view very narrowly) good. We disagree about whether it’s appropriate to call them “useful.” Would resolving this dispute get us anywhere?
- Or, since you like Bentham: “according to the tendency it appears to have to augment or diminish the happiness of the party whose interest is in question”
- People improperly evaluate the utility of tools in their lives all the time. People are taken in by new toys that promise the world and then find they don’t work as advertised or just aren’t that helpful. But even more importantly, a lot of those people are using LLMs to do bad things!
- Yes, your argument is a simple set of assertions that are at their absolute strongest, just a statement that you don’t like AI & wish to generalize your intuition in order to give it more weight. But it’s worthless to anyone who disagrees—& plenty do, by their actions.
- Again, a lot of ten dollar words for “I disagree because I think AI is useful. And the reason I think it’s useful is because a lot of people use it.” I think you’re smart enough to see the problem with that argument when phrased straightforwardly.
- Buddy, the goalposts haven’t moved just because you forgot where we started, & no, utilitarians don’t generalize a moral valence like that—that’s not even wrong, & thinking a line from Wikipedia refutes that shows you don’t understand how to apply Bentham’s p of utility—he’s explicitly individualist
- You: “People do not generally intrinsically link moral valence with descriptive statements of utility” Me: “the entire utilitarian tradition does.” You: “if an object has even one be beneficial use case it is intrinsically useful!” Lots of ten dollar words to cover up a simple goalpost move
- My argument is pretty simple: “AI has more harmful uses then beneficial uses, and its beneficial uses aren’t very useful, so I would not colloquially describe it as an overall useful tool.” We don’t need to get all Philosophy 201 about this.
- lol no People do not generally intrinsically link moral valence with descriptive statements of utility. Guns are useful for a handful of use, even some bad uses Lockpicks are useful if you want to pick a lock. Stop trying to deontologize “utility.”
- yeah I think a really signal distorting thing is that the people really into crypto and NFTs, which *were* basically just scams/total bullshit are also really into AI. if you're not someone who works in a profession with direct exposure to it's use, totally understandable to use that as heuristic
- [Not loaded yet]
- Ok
- [Not loaded yet]
- Because it’s a neat toy, they’re being propagandized into believing it can do more than it can, and/or they find it very helpful in facilitating bad but lucrative behavior like scams or cheating on schoolwork. And maybe sometimes because it does something actually useful like write an email for you.
