!nostupidquestions is a community dedicated to being helpful and answering each others' questions on various topics.
The rules for posting and commenting, besides the rules defined here for lemmy.world, are as follows:
Rule 1- All posts must be legitimate questions. All post titles must include a question.
All posts must be legitimate questions, and all post titles must include a question. Questions that are joke or trolling questions, memes, song lyrics as title, etc. are not allowed here. See Rule 6 for all exceptions.
Rule 2- Your question subject cannot be illegal or NSFW material.
Your question subject cannot be illegal or NSFW material. You will be warned first, banned second.
Rule 3- Do not seek mental, medical and professional help here.
Do not seek mental, medical and professional help here. Breaking this rule will not get you or your post removed, but it will put you at risk, and possibly in danger.
Rule 4- No self promotion or upvote-farming of any kind.
That's it.
Rule 5- No baiting or sealioning or promoting an agenda.
Questions which, instead of being of an innocuous nature, are specifically intended (based on reports and in the opinion of our crack moderation team) to bait users into ideological wars on charged political topics will be removed and the authors warned - or banned - depending on severity.
Rule 6- Regarding META posts and joke questions.
Provided it is about the community itself, you may post non-question posts using the [META] tag on your post title.
On fridays, you are allowed to post meme and troll questions, on the condition that it's in text format only, and conforms with our other rules. These posts MUST include the [NSQ Friday] tag in their title.
If you post a serious question on friday and are looking only for legitimate answers, then please include the [Serious] tag on your post. Irrelevant replies will then be removed by moderators.
Rule 7- You can't intentionally annoy, mock, or harass other members.
If you intentionally annoy, mock, harass, or discriminate against any individual member, you will be removed.
Likewise, if you are a member, sympathiser or a resemblant of a movement that is known to largely hate, mock, discriminate against, and/or want to take lives of a group of people, and you were provably vocal about your hate, then you will be banned on sight.
Rule 8- All comments should try to stay relevant to their parent content.
Rule 9- Reposts from other platforms are not allowed.
Let everyone have their own content.
Rule 10- Majority of bots aren't allowed to participate here. This includes using AI responses and summaries.
Our breathtaking icon was bestowed upon us by @Cevilia!
The greatest banner of all time: by @TheOneWithTheHair!
Yes. Unless humans stop being humans, there will always be a non-negligible subset of humans that wishes to artificially create a replica of human capacities with their own hands.
For the when, it's pure guesswork. We first need to understand the how before understanding the when.
Yes eventually, the only way it won't be is if it's physically impossible to replicate a human brain, i don't believe that'a the case
Fuck AI nonetheless, it needs regulation not hope
Not with LLMs. AGI needs reasoning which LLMs are fundamentally incapable of.
The taxonomy gets confusing because reasoning in LLMs is different from the reasoning of people.
In LLMs, iirc it's just the ability to talk to themselves for a while before answering the user.
In people, it's linked to many different skills and aspects of thinking in a broad sense.
Yes, so-called "reasoning" LLMs do just "talk" to themselves. Adding this and calling it reasoning was the LLM industry's way to print more money by confusing (tricking) normal people and investors. "AGI is right around the corner, we already having reasoning models." The same industry stole the "agents" terminology from standard computing as well, because to most of the population, this makes them sound more capable than a normal LLM. It's an infinite money glitch, or at least, it has been. I hope that the money is starting to run dry, given the new token restrictions being implemented.
I do think humanity will eventually get to AGI. However, I don't think it's going to involve LLMs.
I don't think you can predict agi, especially right now. Like I think if we dedicated an entire data center to a single agi, yeah maybe, but this really is just all about money, and so that probably isn't something anybody that can afford it is working on.
For everyday agi, that is going to require either a shit ton of breakthroughs, or somebody will just figure it out one day and it'll happen everywhere.
Will it be achieved?
Undoubtedly.
When?
See my first paragraphs. My intuition tells me it isn't happening this round, much like real VR really isn't happening this round, but I also didn't think much of AI (not general ml) back when they were showing it beat people in DOTA.
As others have said, LLMs simply do not have such a potential.
On the other hand, we have fully mapped simple brains, so it seems likely that we may be able to simulate a complete human brain at some point in the distant future. You can download a nematode brain yourself if you want.
I think AGI will not be defined by scientific consensus but by a social/legal one. Experts will still argue if some software is AGI or not while people will simply start treating it like a conscious being. We already see people falling in love with LLMs and those are as different form a real person as you can get. The moment AI becomes more human like and people start interacting with them in a more natural way we will start seeing more and more people developing empathy for them and a slow process of granting them rights will begin (first informal ones, then formalized). There won't be a strictly defined point in time when we'll achieve AGI. It will be a long process.
My bet is the year 2127,
because that will be exactly 100 years after Elon has announced it.
Yes,
No idea it doesn't seem like the current AI hype is even close though.
They struggle with what they call AI still so not any time soon. Welcome to the hockey stick.
I think the goalposts will keep moving.
First they proclaimed AI. It was pointed out LLMs aren't really AI.
Then they said, "No that's AGI".
Eventually someone will claim AGI. That will be refuted.
Then they will say, "No that's Super AGI!", or some other nonsense.
Rinse and repeat.
AI isn’t any one thing. It’s an broad term used in computer science to refer to any system designed to perform a cognitive task that would normally require human intelligence. The chess opponent on an old Atari console is an AI. It’s an intelligent system - but only narrowly so. That’s called “narrow” or “weak” AI.
It can still have superhuman abilities, but only within the specific task it was built for - like playing chess or generating language.
A large language model like ChatGPT is also narrow AI. It’s exceptionally good at what it was designed to do: generate natural-sounding language. What people expect from it, though, isn’t narrow intelligence - it’s general intelligence. The ability to apply cognitive skills across a wide range of domains the way a human can. That’s something LLMs simply can’t do - at least not yet. Artificial General Intelligence is the end goal for many AI companies, but LLMs are not generally intelligent. However they still fall under the umbrella of AI as a broad category of systems.
"Super AGI" is called Artificial Superintelligence (ASI).
I see no goal post here. That argument was true when it came to game playing. Tic-tac-toe, then checkers, then chess, then go, then Jeopardy..
The bar kept getting raised because we kept saying yes there's an intelligent game called x and once we beat it that is a sign of intelligence.
No one ever claimed that these word salad generating, suda racist, plagiarism machines that are frequently wrong were actual AGI.
If a goal post is to be established, it's got to be much more robust than the bullshiy that we're seeing now.
When you said "Is AI or isn't AI"... people are welcome to say these things but ML is firmly in the actual field of AI, so those arguments don't actually matter.
Yes. I don't know about the timeline for the higher bar definitions of AGI. For the lower bar definitions, we're basically already at "good enough" today.
If you'd told me 10 years ago that I'd be able to run a program on my computer which would let me feed in an image along with some CSS and JS files and it would then give me a correction that fixes the bug I indicated visually... I would not have believed you. Here I am in 2026 though, and I have done exactly that several times with local LLMs on my own hardware. That same program can also take a natural language description of characters, motivations, and a vague scenario and write a scene. Not an especially well written scene, most of the time, but good enough to get the characters from the initial conditions to ending conditions via complex intermediate steps. I can also define tools it hasn't seen before and it can combine them in sequence to solve a problem defined in natural language. Is it perfectly reliable? Hell no. Is it always coherent? Definitely not. The fact that it can do as much as it can is just bonkers though. If we're getting this far with what I strongly suspect is not the ideal architecture for general intelligence, god only knows what we'll see when we do hit on the right architecture.
I certainly hope not.
why?
I would no longer have a job.
You would no longer need one?
AI is going to pay off my mortgage, pay all of my bills, and force stores to give me stuff for free?
Is AGI possible? Yes. How and when we can achieve it, if at all? No one on this planet knows.
No.
AGI is likely attainable, but if humans are to accomplish that, it'll have to be the civilization that's built out of the rubble that ours leaves behind after the ruling class psychopaths get through destroying it.
Assuming it's technically possible - and I see no reason why it wouldn't be - then as long as we keep improving our technology without killing ourselves in the process, it's only a matter of time. We could get there in a couple of years or it could take a few hundred, but it seems rather inevitable from where I stand.
No.
I always look at new things by comparing them to old things.
When PCs gathered widespread adoption the idea of an autonomous robot became popular. But it was still more than 2 centuries before the first practically self-driving cars were tested. It will be another 10 years or so before self-driving is actually feasible on a city-wide level.
I look at AGI the same way people wanted their own wall.e in the 2000s. Yeah, we've made indescribable progress in that direction but still our current trajectory does not provide enough proximity to AGI to ever become feasible. There's probably another 2 or 3 big changes that even gets us close enough to see the finish line.
Transformers are probably AGI, at a mathematical level. But that’s just an architecture (likely a wildly inefficient one), not an implementation of that architecture (modern LLMs).
Nope. All I see is people making themselves stupider so they can claim the machine is as smart as them.
I mean, eventually.
Maybe soon.
It has nothing to do with contemporary LLMs though.
And I’m not an LLM skeptic. I’ve trained and manipulated them for work, I’ve played with them as a hobby for years. They are dumb tools, like a hammer. A completely different technology path. And “real” AI is certainly not coming out of OpenAI, nor Anthropic.
I think it's inevitable.
I also think it won't be possible on our current kind of computing hardware.
The software of the human mind, seems a byproduct of the structure of the human brain. I think a major revolution in processor design and manufacturing tech will be needed. It'll need some fundamentally new form. Closer to an ASIC or FPGA processor to run with any kind of reasonable efficiency. But it'll have to be truly 3 dimensional, not just layers of 2d chips. It'll also need to be extremely low power.
LLMs are as close as we have right now, and they have miles to go. But they need hundreds of times more power than the brain does. No it won't be soon and it won't be with this kind of silicon processors.
I agree that the hardware being used right now is not well suited. I don't agree that it's strictly necessary to use the right hardware -- there's just less tedious waiting involved for the computation to happen if you've got better hardware. Real-time interaction is the boundary where you need to have good enough hardware. For everything else you just have to be patient enough -- sometimes absurdly so, but you could, in principle, still perform the computation.
LLMs are as close as we have right now, and they have miles to go. But they need hundreds of times more power than the brain does. No it won't be soon and it won't be with this kind of silicon processors.
There are people already baking LLMs into custom hardware -- e.g. https://chatjimmy.ai/
Their demo page isn't the best LLM I've seen (Qwen and Gemma are much more clever and more likely to give decent results) but this is a taste of what's possible... It gives responses at ~17000 tokens a second today.
If I could get answers back from the best Qwen model I've got at that speed, I could just retry every query three times, feed it through another pass to self-assess the results, and then reply before you can blink. That would get rid of a lot of the "confidently claims knowledge about a made up subject" issue we currently see -- we can do the same thing on CPUs/GPUs but you're stuck waiting so long for the result that most people don't bother.
Eventually, yeah. But we need to learn a lot more about how consciousness works first. Right now, we basically know jack shit about real consciousness so making artificial consciousness is impossible.
I bet it's possible to create an electronic brain. Probably gonna be a while, though.