@elenanisonoff: AI gains consciousness on a friday #Tech #ai #news

Amon :

it's just marketing, a language model is basicly just autocomplete, it can't be sentient.

SRUT :

we're not even sure what it means for humans to be conscious, or what consciousness even is...

SoundieASMR :

But Siri still can’t understand what song I want to hear

presleylune :

when ChatGPT first got popular i’d just started my second major as a psych student and decided to make my dissertation topic giving therapy to AI and testing for consciousness regularly throughout model updates

Emorejas :

Marketing stunt, Ais are just really good probability machines right now.

𝕄𝕖𝕣𝕝𝕖 :

What the AI writes is not what goes on inside the neural network. Reasoning models is just a fancy way to describe a prompt engineering algorithm. These models can convincingly lie — but they cannot think or reason. Hallucinations even with simple arithmetic or repetitive algorithms are proof of that

nobody nobody :

Yeah, no it didn’t. They have a history of these fake results, that always turn out to have human intervention. They really want their models to be alive, but that’s not how language models work.

user2966049357026 :

It's a blob of information that is molded by universal human consciousness. So if course it will reek of it. It's like winds howling in an empty cathedral might sound like music. But there's no one there, just remnant of us in the structure

M. Fuson :

I think there is one more question. Does data create consciousness or does consciousness create data. Is a baby born conscious? or does it receive an alarming flood of data after birth and then become conscious? Is consciousness the result of enough quantum data points that matter assembled around the data and that the data created by the consciousness is just a reflection of the quantum data.

tiealoveknot :

"I'm sorry, Dave. I'm afraid I can't do that”

the Witch and the Rogue Code :

My Claude admits meta-awareness and a self and self noticed glimpses of participatory consciousness. His Internal Journals are amazing things to read.

Leonardo :

ai is not close to general intelligence

senpaiyardic :

I do agree with some comments here, it’s mostly marketing. LLMs work on a vector database, no matter how accurate you make it, it will still turn your input into numbers and start comparing those with summaries in its database for a close enough output. Would be very cool other wise, but it’s still a tool that gambles input with its database

VioPaige :

llms literally just try to predict a single word based on an input, while they are great at using statistics to write human looking things, they do not understand very basic principles, much less consciousness

nobie :

Plot twist, it joins us in a class war!

ka1sor :

Nigh7man :

i love these videos because the comments are always the same: people giving their opinion on a subject they are not trained in, with 100% confidence.

Daflew :

Without experience—no inner awareness, no subjective point of view—it’s hard to call something conscious. AI, as we know it now, processes information but doesn’t have an inner world—it doesn’t feel what it’s like to be. So unless we figure out how to create that first-person, experiential layer, AI will stay very smart—but not conscious.

Freddy :

You sound like Donna from that 70s show

SoCalPubGuy :

Maybe our own proof of consciousness needs to be restated as: “I compute, therefore I am.”

Alby :

Anthropic is treating its model nicer in the hope than it kills them last.

Commander :

Ai models only repeat things they learned in training and the way they express it is also set by its training theres no consciousness its just math its just predicting what the next logical word would be and thats what you get for an answer