show me a so-called ai that doesn’t fuck up all the goddamn time and maybe I’ll use it for something simple. except it fails the simplest things all the time. does it so much that they have cutesy names for it. it’s not libel, it’s hallucination. it’s not murder, it’s mortality manifestation. fuck ai. get back to making tools that actually work.
why do i have a feeling if i asked you to tell me what hallucinations are in a technical sense i would get a regurgitated answer from google?
being blind to the obvious doesn’t help anyone, man. anyone who has genuinely worked on or even just with these tools knows that they are capable of producing quality outputs. sometimes they mess up, sure, but it also can work 1000000x faster than you can. the energy problem in turn is a valid discussion but this is just being oblivious to the obvious.
why do you guys all mistake the climate of early tech adoption as an indicator of the technology itself being bad? were you not alive for the rise of the internet or something? i think you guys all just hate corporatism, not AI, but for some reason can’t take the logical step to that conclusion.
I don’t know why you have that feeling because you definitely wouldn’t get a regurgitated answer from google, since I don’t give a shit what it is in a technical sense. guess what, if I buy a phone that might catch fire every once in a while, I don’t need to know how or why it does that in a technical sense to confidently say that it is shit and not worth my money or time.
“sometimes they mess up” is not good enough, and no, the output is not “quality”.
i was alive for the rise of the internet and the analogy doesn’t work. llms are fundamentally useless for 90% of what they’re currently being used for, which is mostly generic assistance. assistance needs knowledge and actual skills not a glorified autocomplete for everything.
That’s entirely on you for using it for what its bad at and then claiming its bad at everything. I use it an LLM literally every day for work and it’s a time saver. I had to learn what its good for and what it’s not though. I also use the better available versions, not the publically available ones. Asking it questions about vague and subjective things isn’t where its best. Asking it to make an excel formula that does a thing without needing to even know a function exists to do that? Priceless.
using things for what they’re advertised for is not “entirely on me”. you can’t sell me a phone and blame me when I say it’s shit because it makes a great doorstop.
Yes it is. The creators don’t fully know what their own products are capable of or how best to use them. If you are dependent on others to tell you how to use this new tool, you will be behind the curve.
show me a so-called ai that doesn’t fuck up all the goddamn time and maybe I’ll use it for something simple. except it fails the simplest things all the time. does it so much that they have cutesy names for it. it’s not libel, it’s hallucination. it’s not murder, it’s mortality manifestation. fuck ai. get back to making tools that actually work.
why do i have a feeling if i asked you to tell me what hallucinations are in a technical sense i would get a regurgitated answer from google?
being blind to the obvious doesn’t help anyone, man. anyone who has genuinely worked on or even just with these tools knows that they are capable of producing quality outputs. sometimes they mess up, sure, but it also can work 1000000x faster than you can. the energy problem in turn is a valid discussion but this is just being oblivious to the obvious.
why do you guys all mistake the climate of early tech adoption as an indicator of the technology itself being bad? were you not alive for the rise of the internet or something? i think you guys all just hate corporatism, not AI, but for some reason can’t take the logical step to that conclusion.
I don’t know why you have that feeling because you definitely wouldn’t get a regurgitated answer from google, since I don’t give a shit what it is in a technical sense. guess what, if I buy a phone that might catch fire every once in a while, I don’t need to know how or why it does that in a technical sense to confidently say that it is shit and not worth my money or time.
“sometimes they mess up” is not good enough, and no, the output is not “quality”.
i was alive for the rise of the internet and the analogy doesn’t work. llms are fundamentally useless for 90% of what they’re currently being used for, which is mostly generic assistance. assistance needs knowledge and actual skills not a glorified autocomplete for everything.
That’s entirely on you for using it for what its bad at and then claiming its bad at everything. I use it an LLM literally every day for work and it’s a time saver. I had to learn what its good for and what it’s not though. I also use the better available versions, not the publically available ones. Asking it questions about vague and subjective things isn’t where its best. Asking it to make an excel formula that does a thing without needing to even know a function exists to do that? Priceless.
using things for what they’re advertised for is not “entirely on me”. you can’t sell me a phone and blame me when I say it’s shit because it makes a great doorstop.
Yes it is. The creators don’t fully know what their own products are capable of or how best to use them. If you are dependent on others to tell you how to use this new tool, you will be behind the curve.
lol I don’t consider being able to construct my own sentences as behind the curve but hey.
But you can’t construct sentences that produce value from AI while others can.