News for hackers. Stuff that matters.
Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by Eisenstein on Tuesday December 16, 2025 @09:33PM (#46297567)
> They seem smart, but they are not; they are really just good at appearing to be smart.

Can you give an example of the difference between these two things?

Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by this_user on Tuesday December 16, 2025 @09:46PM (#46297660)
Imagine an actor who is playing a character speaking a language that they actor does not speak. Due to a lack of time, the actor decides against actually learning the language and instead opts to just memorise and train how to speak their lines without actually understanding the content. Let's assume they are doing a pretty convincing job too. Now, the audience watching these scenes may think that the actor is actually speaking the language, but in reality they are just mimicking.

This is what an LLM essentially is. It is good at mimicking, reproducing and recombining the things it was trained on. But it has no creativity to go beyond this, and it doesn't even possess true reasoning, which is how it will end up making mistakes that are just immediately obvious to a human observer, yet the LLM is unable to see them, because it just mimicking.

Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by retsibsi on Tuesday December 16, 2025 @10:02PM (#46297762)
> Imagine an actor who is playing a character speaking a language that they actor does not speak. Due to a lack of time, the actor decides against actually learning the language and instead opts to just memorise and train how to speak their lines without actually understanding the content.

Now imagine that, during the interval, you approach the actor backstage and initiate a conversation in that language. His responses are always grammatical, always relevant to what you said modulo ambiguity, largely coherent, and accurate more often than not. You'll quickly realise that 'actor who merely memorized lines in a language he doesn't speak' does not describe this person.

Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by Salgat on Wednesday December 17, 2025 @12:00AM (#46298441)
You've missed the point of the example, of course it's not the exact same thing. With regard to LLM, the biggest difference is that it's a regression against the world's knowledge, like an actor who memorized every question that happens to have an answer written down in history. If you give him a novel question, he'll look at similar questions and just hallucinate a mashup of the answers hoping it makes sense, even though he has no idea what he's telling you. That's why LLMs do things like make up nonsensical API calls when writing code that seem right but have no basis in reality. It has no idea what it's doing, it's just trying to regress code in its knowledge base to match your query.
Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by retsibsi on Wednesday December 17, 2025 @01:23AM (#46298842)
I don't think I missed the point; my point is that LLMs do something more complex and far more effective than memorise->regurgitate, and so the original analogy doesn't shed any light. This actor has read billions of plays and learned many of the underlying patterns, which allows him to come up with novel and (often) sensible responses when he is forced to improvise.
Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by captain_coffee on Wednesday December 17, 2025 @03:09AM (#46299417)
> LLMs do something more complex and far more effective than memorise-regurgitate

They literally do not, what are you talking about?

Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by esafak on Wednesday December 17, 2025 @11:55AM (#46302020)
What kind of training data do you suppose contains an answer to "how to build a submarine out of spaghetti on Mars" ? What do you think memorization means?

https://chatgpt.com/s/t_6942e03a42b481919092d4751e3d808e

Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by mlmonkey on Tuesday December 16, 2025 @10:48PM (#46298014)
You are describing Searle's "Chinese Room argument"[1] to some extent.

It's been discussed a lot recently, but anyone who has interacted with LLMs at a deeper level will tell you that there is something there; not sure if you'd call it "intelligence" or what. There is plenty of evidence to the contrary too. I guess this is a long-winded way of saying "we don't really know what's going on"...

[1] https://plato.stanford.edu/entries/chinese-room/

Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by x______________ on Tuesday December 16, 2025 @11:26PM (#46298247)
If an LLM was intelligent, wouldn't it get bored?
Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by jimbokun on Tuesday December 16, 2025 @11:37PM (#46298314)
Why should it?
Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by Eisenstein on Tuesday December 16, 2025 @09:53PM (#46297708)
1. I would argue that an actor performing in this way does actually understand what his character means

2. Why doesn't this apply to you from my perspective?

Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by heavyset_go on Tuesday December 16, 2025 @10:21PM (#46297881)
Wisdom vs knowledge, where the word "knowledge" is doing a lot of work. LLMs don't "know" anything, they predict the next token that has the aesthetics of a response the prompter wants.
Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by Eisenstein on Tuesday December 16, 2025 @11:57PM (#46298431)
It doesn't seem obvious to me that predicting a token that is the answer to a question someone asked would require anything less than coming up with that answer via another method.
Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by subb on Wednesday December 17, 2025 @12:31AM (#46298624)
I suspect a lot of people but especially nerdy folks might mix up knowledge and intelligence, because they've been told "you know so much stuff, you are very smart!"

And so when they interact with a bot that knows everything, they associate it with smart.

Plus we anthropomorphise a lot.

Is Wikipedia "smart"?

Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by Eisenstein on Wednesday December 17, 2025 @02:48AM (#46299300)
What is the definition of intelligence?
Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by anon84873628 on Wednesday December 17, 2025 @07:25AM (#46301185)
Ability to create an internal model of the world and run simulations/predictions on it in order to optimize the actions that lead to a goal. Bigger, more detailed models and more accurate prediction power are more intelligent.
Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by Eisenstein on Wednesday December 17, 2025 @07:20PM (#46307466)
How do you know if something is creating an internal model of the world?
Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by anon84873628 on Friday December 19, 2025 @11:24AM (#46327527)
Look at the physical implementation of how it computes.
Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by Eisenstein on Friday December 19, 2025 @06:33PM (#46332224)
So you are making the determination based on the method, not on the outcome.
Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by anon84873628 on Friday December 19, 2025 @08:34PM (#46332940)
Did I ever promise otherwise? Intelligence is inherently computational, and needs a physical substrate. You can understand it both by interacting with the black box and opening up the box.
Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by subb on Wednesday December 17, 2025 @08:09PM (#46307825)
Definitely not _only_ knowledge.
Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by Eisenstein on Wednesday December 17, 2025 @11:20PM (#46308862)
Right, so a dictionary isn't intelligent. Is a dog intelligent?
Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by g947o on Tuesday December 16, 2025 @09:50PM (#46297693)
Hallucinating things that never exist?
Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by mycall on Tuesday December 16, 2025 @09:58PM (#46297741)
Imagination?
Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by g947o on Tuesday December 16, 2025 @10:00PM (#46297757)
I think these are clearly two different words that mean different things.
Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by mycall on Saturday December 20, 2025 @04:56PM (#46340010)
Yet they are correlated and confused in part.
Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by anon84873628 on Tuesday December 16, 2025 @10:00PM (#46297753)
For fun I asked Gemini about this. It wrote a lot, including:

Appears smart: pattern matching. Actually smart: first principles understanding.

Is that specific enough?

It also made reference to stochastic parrots vs emergent reasoning, the bat and ball problem, the library vs the librarian, and the Chinese room.

It ended by asking if I would like it to solve a logic puzzle I made up on the spot to see if it relies on patterns or reasoning.

Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by mrits on Tuesday December 16, 2025 @09:42PM (#46297626)
Being able to learn to play Moonlight Sonata vs. being able to create it. Being able to write a video game vs being able to write a video game that sells. Being able to tell you newtons equations vs being able to discover the acceleration of gravity on earth
Re:Meta's New A.I. Superstars Are Chafing Against the Rest o... by Eisenstein on Tuesday December 16, 2025 @09:50PM (#46297688)
So if an LLM could do any of those things you would consider it very smart?