It’s really good at making us feel like it’s intelligent, but that’s no more real than a good VR headset convincing us to walk into a physical wall.
It’s a meta version of VR.
(Meta meta, if you will.)
The biggest issue with AI as it currently exists with LLMs and such as I see it is that there is a pretty big gulf between what AI is today and what the average person has been taught AI is by TV/Movies/Books/Games their entire lives.
And OpenAI, Google, Nvidia, et al are heavily marking the former as if it is the latter.
The big players are marketing the expectations creating by science fiction, not the reality of their products/services.
A tic tac toe opponent algorithm is also considered Artificial Intelligence. People never had a problem with it.
Why? We already have a specific subcategory for it: Large Language Model. Artificial Intelligence and Artificial General Intelligence aren’t synonymous. Just because LLMs aren’t generally intelligent doesn’t mean they’re not AI. That’s like saying we should stop calling strawberries “plants” and start calling them “fake candy” instead. Call them whatever you want, they’re still plants.
Bruh you just said that AI isn’t “I”. That’s the entire point of the OP
They said not generally intelligent, which is a specific and important property of AGI, not AI. In the tic tac toe example, the AI is intelligent (can play tic tac toe), but this intelligence cannot be generalised to playing chess, appreciating art, whatever the general measures may be.
No I didn’t.
The term “Artificial Intelligence” is actually a perfectly cromulent word to be using for stuff like LLMs. This is one of those rare situations where a technical term of art is being used in pop culture in the correct way.
The term “Artificial Intelligence” is an umbrella term for a wide range of algorithms and techniques that has been in use by the scientific and engineering communities for over half a century. The term was brought into use by the Dartmouth workshop in 1956.
But it’s not simulated intelligence. It’s literally just word association on steroids. There are no thoughts it brings to the table, just words that mathematically fit following the prompts.
It’s not just statistics. To produce a somewhat coherent sentence in English you need a model of the English language AND a world model.
If you ask a question like “an apple is on a glass, what happens if I remove the glass”, the correct answer (“the apple will fall”) is not a statistical property of the English language, but an emergent property of the world model.
Where do you draw the line for intelligence? Why would the capacity to auto complete tokens based on learned probabilities not qualify as intelligence?
This capacity may be part of human intelligence too.when it can come up with a solution it hasn’t seen before.
that’s the threshold.
that’s the threshold for creative problem solving, which isn’t all there is to intelligence, but i think it’s fair to say it’s the most crucial part for a machine intelligence.
It can come up with a brand new sentence that hasn’t been written before. Does that count?
Maybe you mean a solution to a textbook math/physics problem, it most likely would be able to solve that too with tool use.
Or maybe you mean solving something like the Riemann Hypothesis?
no, none of those are what i mean, that’s way too specific to be useful.
a system exhibits intelligence when it can use existing insights to build entirely new insights.
a popular example is that no current “AI” can extrapolate from basic mathematical stipulations to more advanced ones.
(there’s tons of example you could put here, but this is the one i like)
here’s the example:
teach an LLM/DNN/etc. basic addition, subtraction, multiplication, and division.
give it some arbitrary, but large, number of problems to solve.
it will eventually encounter a division that isn’t possible, but is not a divide-by-zero (which should be covered by the rules it was given).
then it will either:
- throw an error
- have an aneurysm
- admit it can’t do that (proving the point)
- or lie through it’s teeth, giving wrong answers (also proving the point)
…but what it will definitely NEVER do, is simply create a placeholder for that operation and give it a name: square root (or whatever ot calls it, that part isn’t important).
it simply can’t, because that would be a new insight, and that’s something these systems aren’t capable of.
a human (or a lot of them) would encounter these impossible divisions and eventually see a pattern in them and draw the proper conclusion: that this is a new bit of math that was just discovered! with new rules, and new applications!
even if it takes a hundred years and scores of them, humans will always, eventually, figure it out.
…but what we currently call “artificial intelligence” will simply never understand that. the machine won’t do that, no matter how many machines you throw at the problem.
because it’s not a matter of quantity, but of quality.
and that qualitative difference is intelligence!
(note: solving this particular math problem is a first step. it’s unlikely that it will immediately lead to an AGI, but it is an excellent proof-of-concept)
this is also why LLMs aren’t really getting any better; it’s a structural problem that can’t be solved with bigger data sets.
it’s a fundamental design flaw we haven’t yet solved.
current "AI"s are probably a part of the solution, but they are, definitely, not THE solution.
we’ve come closer to an AI, but we’re not there.
This.
I have taught highschool teens about AI between 2018 and 2020.
The issue is we are somewhere between getting better at gambling (statistics, Markov chains, etc.) and human brain simulation (deep neural networks, genetic algorithms).
For many people it’s important how we frame it. Is it random word generator with a good hit rate or is it a very stupid child?
Of course the brain is more advanced - it has way more neurons than an AI model has nodes, it works faster and we have years of “training data”. Also, we can use specific parts of our brains to think, and some things are so innate we don’t even have to think about it, we call them reflexes and they bypass the normal thinking process.
BUT: we’re at the stage where we could technically emulate chunks of a human brain through AI models however primitive they are currently. And in it’s basic function, brains are not really much more advanced than what our AI models already do. Although we do have a specific part for our brain just for languages, which means we get a little cheat code for writing text in comparison to AI, and similar other parts for creative tasks and so on.
So where do you draw the line? Do you need all different parts of a brain perfectly emulated to satisfy the definition of intelligence? Is artificial intelligence a word awarded to less intelligent models or constructs, or is it just as intelligent as human intelligence?
Imo AI sufficiently passes the vibe check on intelligence. Sure it’s not nearly on the scale of a human brain and is missing it’s biological arrangements and some clever evolutionary tricks, but it’s similar enough.
However, I think that’s neither scary nor awesome. It’s just a different potential tool that should help everyone of us. Every time big new discoveries shape our understanding of the world and become a core part of our lives, there’s so much drama. But it’s just a bigger change, nothing more nothing less. A pile of new laws, some cultural shifts and some upgrades for our everyday life. It’s neither heaven nor hell, just the same chunk of rock floating in space soup for another century.
I dunno, the power requirements would seem to be an ecological catastrophe in the making, except it’s already happening.
Well, if we are not looking at all the disaster the hype is doing on so many levels (which is fine in the sense that technology and fools are different things), I draw the line at… intelligence, not simulation of hardware. I care lot less if something before me runs on carbon, metal or, say, sulfur than I care if it is intelligent
And as someone has already pointed out, even defining intelligence is damn hard, and different intelligence works differently (someone who is great at moving their body, like dancers or martial artists, is definitely more intelligent than me in quite a few areas, even if I know math or computers better than them). So… “artificial intelligence” as a bunch of algorithms (including LLM) etc - no problem with me, “artificial intelligence” as “this thing is thinking” or “this thing is just as good as a human artist/doctor/lawyer” - nah, bullshit
A simulation doesn’t have to be the actual thing. It implies it literally isn’t the true thing, which is kind of what you’re saying.
Simulated Intelligence is certainly more accurate and honest than Artificial Intelligence. If you have a better term, what is it?
If you have a better term, what is it?
Large Language Model.
Doesn’t seem to be catching on…
Professor Hotpants’ Astounding Rhetorical Thingamajig
My dog can do calculus but struggles with word association beyond treat, walk, vet and bath. Intelligence is hard to define.
I mean to friends and family – people who have accepted it as smart.
I don’t know about you, but when I try to explain the concept of LLMs to people not in the tech field, their eyes glaze over. I’ve gotten several family members into VR, though. It’s an easier concept to understand.
words that mathematically fit following the prompts
if only we had a word for applying math to data to give the appearance of a complex process we don’t really understand.
yeah i call it an english simulator
It’s not even simulating intelligence.
Personally been a fan of shoggoth with a smiley face mask
In Mass Effect, it’s VI (Virtual Intelligence), while actual AI is banned in the galaxy.
The information kiosk VIs on The Citadel are literally LLMs and explain themselves as such. Unlike AI/AGI they aren’t able to plan, make decisions, or self-improve, they’re just a simple protocol on a large foundational model. They just algorithmic.
Simulated Intelligence is okay, but virtual implies it mimics intelligence, while simulated implies it is a substitute and actually does intelligence.
AI is a parent category and AGI and LLM are subcategories of it. Just because AGI and LLM couldn’t be more different, it doesn’t mean they’re not AI.
I don’t at all agree with this graph, and I think you’re sort of missing the point of the original post.
What do you not agree with the graph?
Yeah this graph doesn’t make sense to me either. Where did this come from? Who is teaching this?
Does this help?
Yes! When I started looking deeper into LLMs after GPT blew up, I thought “this all sounds familiar.”
This is perfect. I’m definitely going to shoehorn this into any discussion that even tangentially applies to SI.
I prefer simulated human intelligence type or SHIT for the people who like acronyms
high five
Artificial Simulated Sentience
You could also go with CRAP or Complicated Reasoning And Processing.
I have been referring to LLMs and image generators as “Plagiarism Engines” for some time. Even SI seems too generous.
@LillyPip
PseudoIntelligence.sudo intelligence
This incident will be reported.
Reporting the incident has been reported
deleted by creator
Sudo not recognized.
… BSD, eh?
I prefer VI (virtual intelligence) from Mass Effect
It’s not even that. It is just a PwaD (Parrot with a Dictionary).
Parrots are way smarter than LLMs.
That solid but I prefer to call it a “synthetic text extruding machine” or better yet call it “a racist pile of linear algebra”