5 ms·

We should be better than genAI. For me, with things like coding and writing, I think I'm better than it. For painting? It wins over me. But I'd never in a milli

by beej71 4mo ago

We should be better than genAI. For me, with things like coding and writing, I think I'm better than it. For painting? It wins over me. But I'd never in a million years claim creation of a painting I prompted, even if I went and modified it after. Same with code or writing, for that matter. For the record, I never use AI in writing. First, it's no fun, and that's really reason enough. It'd be like sending a bot to watch a movie for me. Second, I'm better than it is, and it would just be an exercise in frustration micromanaging its every word. I might as well just type what I want it to say. And third, most importantly, who would read anything I wrote if they could just generate it themselves? The goal is to be better.

classified 4mo ago
> And third, most importantly, who would read anything I wrote if they could just generate it themselves? This should be embarrassingly obvious to everyone, but allegedly there are hoards of vibecoders and other slop producers to whom it never occurs (if it's not astroturfers pretending to be that). If prompt slinging is your only expertise, you already are redundant.
- scotty79 4mo ago
  [flagged]
- solarkraft 4mo ago
  Meh. I‘d be happy if my AI built my program just the way that I want it, but it turns out to be a lot of work go get it to do that (Deepseek, Claude might be a bit less involved). I still have to make a bunch of decisions based on its research, so it still turns out to be hours of work that can go into it. Granted, I can probably optimize some things by encoding my style preferences. But so far it has been a lot more involved than „make me an app, make no mistakes“. That‘s why I‘d call it in a large part my work. It‘s my vision, in the same way an architect that has never personally laid a brick can call a building „their work“.
scotty79 4mo ago
[flagged]
- SaucyWrong 4mo ago
  It was difficult to finish reading your post after you put the OP down for stating that they believe they're more original and more capable than an LLM. Maybe leave that part out next time, or at least don't lead with it. It does not help your argument the way you think it does. I wonder why you would mince your words about LLM use at all though. Instead of couching your methods in language like, "I 'built' a program," or, "I've 'crafted' a novel," you ought to just put your money where your mouth is and say, "I used LLMs to make [something]." If want to be up-front and proud about the fact that you use LLMs in your work, just dot that without indirection or ornamentation. Wear that attitude on your sleeve and let your audience determine if your work is worth their attention.
  - scotty79 4mo ago
    [flagged]
    - beej71 4mo ago
      > It would be more charitable to substitute myself for him in that part to avoid causing distress. FWIW, I've been on the Internet for a long, long time. I took no offense. But I still think I'm better than AI at those things. :)
- beej71 4mo ago
  >> We should be better than genAI. > Why exactly? I think the top reason for most people is so that they're worth hiring. If you offer no more than an AI, what is your value? Is scotty79 human? Was this written by you? Or are you just an AI? And if you are human and wrote it, why did you do that when an AI could do it better than you? And if an AI could do it better, why even communicate with you? I could just go talk to Claude. Other reasons might be to push the SOTA past what an genAI can do. Or to produce human work in human society. Maybe someday we'll be past needing to be hired to eat, and then maybe we just stop worrying and let genAI be better than all of us. But I find that to be undesirable. > And you are definitely not better than you + AI. I think this is what I'm saying. If you're not better, you're just in the way. If you're not better, then you < you + AI. Here's an example: early ChatGPT sucked at remotely complex jq. I wanted a function to do fuzzy comparisons between JSON objects, and it just could not get it right. I didn't speak jq, so it was just me pasting errors and output and cutting code--zero value add. Eventually I gave up. I then spent 3 hours working on learning jq and and came up with a more robust and capable solution than ChatGPT had been working toward. Now you're going to correctly say that ChatGPT didn't have enough training data to perform well with this relatively new and relatively infrequently-used language. But my point is that you need to be better than it so you can do novel things that it can't. All this assumes, of course, that the AI-bro future of AI doing everything doesn't come to pass. If it does, it'll be global economic collapse and we can just go back to feudalism. > I no longer say I wrote a program Other management verbs work well here, I think. "Produced", "oversaw", "managed development of", etc. > In the end I believe the content is the king. I think so, too. The question is, do you have anything to offer, or should I just go to a bot?
hombre_fatal 4mo ago
On the other hand, you might be better than AI if given maximal time, attention, and energy. But the question is how often are you better at all the tasks where you don’t have maximal time, attention, and energy - and it would make no sense to invest it.

by beej71 4mo ago

classified 4mo ago

> And third, most importantly, who would read anything I wrote if they could just generate it themselves? This should be embarrassingly obvious to everyone, but allegedly there are hoards of vibecoders and other slop producers to whom it never occurs (if it's not astroturfers pretending to be that). If prompt slinging is your only expertise, you already are redundant.

scotty79 4mo ago
[flagged]
solarkraft 4mo ago
Meh. I‘d be happy if my AI built my program just the way that I want it, but it turns out to be a lot of work go get it to do that (Deepseek, Claude might be a bit less involved). I still have to make a bunch of decisions based on its research, so it still turns out to be hours of work that can go into it. Granted, I can probably optimize some things by encoding my style preferences. But so far it has been a lot more involved than „make me an app, make no mistakes“. That‘s why I‘d call it in a large part my work. It‘s my vision, in the same way an architect that has never personally laid a brick can call a building „their work“.

scotty79 4mo ago

[flagged]

SaucyWrong 4mo ago
It was difficult to finish reading your post after you put the OP down for stating that they believe they're more original and more capable than an LLM. Maybe leave that part out next time, or at least don't lead with it. It does not help your argument the way you think it does. I wonder why you would mince your words about LLM use at all though. Instead of couching your methods in language like, "I 'built' a program," or, "I've 'crafted' a novel," you ought to just put your money where your mouth is and say, "I used LLMs to make [something]." If want to be up-front and proud about the fact that you use LLMs in your work, just dot that without indirection or ornamentation. Wear that attitude on your sleeve and let your audience determine if your work is worth their attention.
- scotty79 4mo ago
  [flagged]
  - beej71 4mo ago
    > It would be more charitable to substitute myself for him in that part to avoid causing distress. FWIW, I've been on the Internet for a long, long time. I took no offense. But I still think I'm better than AI at those things. :)
beej71 4mo ago
>> We should be better than genAI. > Why exactly? I think the top reason for most people is so that they're worth hiring. If you offer no more than an AI, what is your value? Is scotty79 human? Was this written by you? Or are you just an AI? And if you are human and wrote it, why did you do that when an AI could do it better than you? And if an AI could do it better, why even communicate with you? I could just go talk to Claude. Other reasons might be to push the SOTA past what an genAI can do. Or to produce human work in human society. Maybe someday we'll be past needing to be hired to eat, and then maybe we just stop worrying and let genAI be better than all of us. But I find that to be undesirable. > And you are definitely not better than you + AI. I think this is what I'm saying. If you're not better, you're just in the way. If you're not better, then you < you + AI. Here's an example: early ChatGPT sucked at remotely complex jq. I wanted a function to do fuzzy comparisons between JSON objects, and it just could not get it right. I didn't speak jq, so it was just me pasting errors and output and cutting code--zero value add. Eventually I gave up. I then spent 3 hours working on learning jq and and came up with a more robust and capable solution than ChatGPT had been working toward. Now you're going to correctly say that ChatGPT didn't have enough training data to perform well with this relatively new and relatively infrequently-used language. But my point is that you need to be better than it so you can do novel things that it can't. All this assumes, of course, that the AI-bro future of AI doing everything doesn't come to pass. If it does, it'll be global economic collapse and we can just go back to feudalism. > I no longer say I wrote a program Other management verbs work well here, I think. "Produced", "oversaw", "managed development of", etc. > In the end I believe the content is the king. I think so, too. The question is, do you have anything to offer, or should I just go to a bot?

hombre_fatal 4mo ago

On the other hand, you might be better than AI if given maximal time, attention, and energy. But the question is how often are you better at all the tasks where you don’t have maximal time, attention, and energy - and it would make no sense to invest it.