Ask Lemmy
A Fediverse community for open-ended, thought provoking questions
Rules: (interactive)
1) Be nice and; have fun
Doxxing, trolling, sealioning, racism, and toxicity are not welcomed in AskLemmy. Remember what your mother said: if you can't say something nice, don't say anything at all. In addition, the site-wide Lemmy.world terms of service also apply here. Please familiarize yourself with them
2) All posts must end with a '?'
This is sort of like Jeopardy. Please phrase all post titles in the form of a proper question ending with ?
3) No spam
Please do not flood the community with nonsense. Actual suspected spammers will be banned on site. No astroturfing.
4) NSFW is okay, within reason
Just remember to tag posts with either a content warning or a [NSFW] tag. Overtly sexual posts are not allowed, please direct them to either [email protected] or [email protected].
NSFW comments should be restricted to posts tagged [NSFW].
5) This is not a support community.
It is not a place for 'how do I?', type questions.
If you have any questions regarding the site itself or would like to report a community, please direct them to Lemmy.world Support or email [email protected]. For other questions check our partnered communities list, or use the search function.
Reminder: The terms of service apply here too.
Partnered Communities:
Logo design credit goes to: tubbadu
view the rest of the comments
You're stuck in the current paradigm about how software works. What I'm talking about is not a current paradigm and it's not AGI.
We don't need AGI for what I'm talking about. You're fixated on programmatically tell a computer how to do something and I'm not sure you're just being difficult or can't grasp or imagine what I'm imagining.
We already have useful LLMs for different tasks. Heck, my team is developing software to perform different tasks using LLMs that if we had to program from scratch we'd be so fucked! Right now, not 2 years after the first version of ChatGPT was released. Do you think this technology will remain the same or will continuously be developed into something that most of us cannot comprehend or will even deny, like you're doing now?
It's your right to not agree with me, and I accept it, but don't say I'm wrong mate, you can't possibly know!
And don't talk about AGI or singularity like it's the next step, you're doing a disservice to yourself.
You haven't really described what you are imagining.
Telling a computer specifically what to do and how to do it without making mistakes is coding. Programming is a level above that, in designing the architecture of how to approach the business problem.
What the other commentator is saying, is that simple being able to tell some model 'build an app that does XYZ' requires AGI because that set of instructions is not complete - the machine requires outside knowledge and the ability to make judgement calls in order to complete it.
If that isn't what you meant, it is at least what you said. The breakdown in communication here, between humans, should also serve as another reminder how difficult it is to convey an idea to another entity and how that problem will remain difficult for a very long time.
Ok, if you got it you got it, if you don't I can't be bothered to spend more time on trying to explain what is a very simple concept that people just don't want to entertain.
I'm out, see you in 10 years.
He got it, you're the one who's not getting it, it is impossible to prompt an entire program to an LLM, I mean you can do it, but even a perfect LLM will give you back a steaming pile of shit, and the reason is because you asked for a steaming pile of shit without realizing it.
I'll make you the same challenge again, write me a prompt as you would to this "AI" for it to do complex stuff and I'll point out several of the assumptions a non-AGI software could wrongly make.
You really are limited, must be a pRoGrAmMeR... We're talking about tech 10 years off, it doesn't exist, it's all hypothetical. And you're asking me how to use it? I'd be fucking rich if I knew that wouldn't I?
No, you would be a coherent person, you can't even fathom how you would interact with a tech yet believe it's possible that's literally crazy or purposefully avoiding it. Plus answering that proves you have absolutely no idea what an LLM is, you're just astonished by the term "AI" and truly believe ChatGPT is intelligent. And also proves you have absolutely no idea how computers work in a general manner, to you they're just magic boxes that do magic and show you things on a screen, therefore anything can happen in your mind.
You talked about the evolution from current AIs so logically you would prompt them in natural language (since they're LLMs), your refusal to give me a prompt for how you would use these LLMs is obviously due to your knowledge that I will find issues with your prompt, thus confirming this is not possible for an LLM.
Ok mate. Suor yourself
Yes, I'm stuck with the paradigm that computers are not intelligent and can't understand what I mean, there's a term for a software that can: "AGI".
Any programmer knows that using LLMs for programming is one of the following cases:
Only people who don't understand programming think LLMs are useful as they are now, until computers are actually intelligent and actually understand what you're asking them to do and think on all of the corner cases and take decisions for all of the things you didn't specifically ask it to consider those 3 cases will be the only outcome for "AI" in programming.
Don't believe me? Tell me a prompt you would use to generate a useful program (i.e. not a Hello World) and I'll list multiple assumptions that you're making about how this needs to work that you did not include in your prompt therefore the "AI" will not cover.