[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/sci/ - Science & Math

View post   
File: 31 KB, 502x284, deeplearning.png [View same] [iqdb] [saucenao] [google]
9176517 No.9176517 [Reply] [Original]

Daily reminder that if you aren't aboard the AI hype train you will be left behind

>> No.9176537

What's with all these AI threads these past few weeks? Is a movie coming out or something?

>> No.9176566

> Tfw machine learning turned my maths and stats degree into fucking meme

>> No.9176575

>>9176517
Daily reminder that computers are effectively just extremely complex and modular rube goldberg machines, and are not alive.

Also, we won't understand the human brain well enough this CENTURY to even get started on ai. Everything we have now is either a cheap parlor trick you could fake with an mp4, or a program.

>> No.9176764
File: 98 KB, 850x565, huh.jpg [View same] [iqdb] [saucenao] [google]
9176764

I hate this putting "quantum" before a field's name

>> No.9176801

>>9176575
AI is not trying to replicate the human brain. that's ridiculous.

>> No.9176803

>>9176801

how do you hope to create something capable of independent thought without first understanding consciousness?

>> No.9176808

>>9176764
>Hello, I'm the chair of the department of quantum pediatric proctology for Johns-Hopkins.

>> No.9176817
File: 7 KB, 208x250, tfw to intelligent too not be a cuck.jpg [View same] [iqdb] [saucenao] [google]
9176817

>>9176808

>quantum business school
>quantum premed

...

>> No.9176823

>>9176803
I am so tired of people who use phrases like "independent thought" and "consciousness" as attacks on AI.

Define what "independent thought" and "consciousness" mean and how we could test whether an system has them.

>> No.9176831

>>9176803
>how do you hope to create something capable of independent thought without first understanding consciousness?

Show that you need either of those things to produce an intelligent agent. Protip: you dont

AI's most likely will not be conscious, but you won't be able to tell the difference.

>> No.9176833

>>9176801
It's trying to create an artificial, sentient, presumably computer-based lifeform.

>>9176823
That's the idea, it's very difficult to even prove where human consciousness comes from, and there's too much we don't understand to replicate the fucking thing.

>> No.9176839

PooInLoo dynasty

Microsoft: PooCEO
Google: PooCEO
The Quantum:....?

Who else here is a deep learning/javascript/blockchain engineer?

>> No.9176860

>>9176831

then it's not an AI, it's just a program

>> No.9176942

>>9176860
>muh dualism

>> No.9176960

>>9176942

Uhh no sweetie, it's just semantics.We need to start using terms we actually know the meaning of before going collectively insane. If you look up the first definitions of the term, it meant any form of non-carbon based sentient life.

>> No.9177039

>>9176833
When has the goals of AI research ever been to create a functional robot that acts as humans do? That's years away from any tech we have because the robot would need to be able to rewrite it's code based on things it's learned to apply it in the mill of every day life, it's stupid for anyone to claim as such. Ai at the moment will do little more than automate and run diagnostics our current technology

>> No.9177212

>>9177039
>When has the goals of AI research ever been to create a functional robot that acts as humans do
The goal is to create a sentient intelligence from technology, which we can't do because defining and proving sentience is extremely difficult, even for our own species. If you're talking about technology that can make decisions and do things, those already exist: they're called programs.

>> No.9177235

>>9177212
Define sentience and how to test it.

>> No.9177246

>>9177235

we can't, that's the whole point

>> No.9177435

>>9177212
>The goal is to create a sentient intelligence from technology,

Nope, the goal is to create an agent that can mimic it through programming. Do you seriously think roboticists and computer scientists give a shit about the "nature of sentience"?

>> No.9177469

>>9177435

why don't you faggots stop calling it AI then, because it's clearly not

keep to your meme names like "deep learning"

>> No.9177544

>>9177469
So your whole point boils down to semantics? Thats some quality autism there famalam.

>> No.9177545

who here /quantum masturbation/?

>> No.9177753

>>9177544

>words mean whatever fits me

I hate post-modernism

>> No.9177917

>>9177544
*quantum autism

>> No.9177955

>>9177544
>I'm right because I say so

>> No.9177977

>>9176823
It implies computers aren't capable of thinking beyond the idea of an absolute proof and a set language that defines these proofs.

>> No.9178014

>>9176517
truly the worst meme in modern science

>> No.9178062

>>9176817
>tfw pure quantum math phd
>300k (quantum currency) starting
>any quantum job i want

>> No.9178065

I support the basilisk

>> No.9178086
File: 24 KB, 400x400, smug levantine mathematician.jpg [View same] [iqdb] [saucenao] [google]
9178086

>>9176823
>define what system means
>define what test means
>define what define means
>define what mean means
You can keep this up forever. Arguments do not have to rely on exact definitions, in fact they mostly don't.
https://twitter.com/nntaleb/status/848157277295181826

>> No.9178127

>>9176860
it doesn't matter what you call it.
If it has the potential to wipe us out or control us, we should pay attention.

>> No.9178224

>>9178127
Stop with the fucking horror stories, a program can't do either of those things unless it's specifically programmed to. There's no program on earth that's going to become self aware and start going after nuclear codes.

An AI could hypothetically do that, but like everyone ITT has said, that's not going to happen for at least 75 years.

>> No.9178250

>>9176517
The great thing about the AI train is that everyone is going to be left behind.

>> No.9178323

>>9176517
>machine learning = AI!
No it doesn't. Machine learning has fuck all to do with AI.

>> No.9178329
File: 34 KB, 360x200, I am ready .jpg [View same] [iqdb] [saucenao] [google]
9178329

>tfw you'll be able to put a quantum nano-chip in your penis to simultaneously quantum fuck several quantic vaginas
>tfw you'll experience quantic ejaculation in your life time

>> No.9178385

can the low IQ "what is consciousness" monkeys stop posting in AI threads

>> No.9178396

Brainlets screeching MUH INTELLIGENCE MUH CONSCIOUSNESS are infuriating. All ML/AI researchers care about is solving the particular problem in front of them in whatever way produces the best results. So far those solutions have all been relatively mundane statistical methods and that is not going to change anytime soon.

Its like screaming at aeronautics engineers for making fighter jets instead of a giant flappy robotic eagle.

>> No.9178397

>>9178062
Quantum mathematics is a thing, like quantum groups and shit, deformations of classical mathematical objects

>> No.9178412

>>9178385
But I finaly found a topic fuzzy enough so that even I brainlet can be part of something smart and exiting!

>> No.9178453

>>9178385
>>9178396
>>9178412

you absolute niggers

your algorithm is not an AI just because it uses statistics to solve a problem, wake up to real life

>> No.9178460

>>9178396
We have a threadwinner

>> No.9178469

>>9178385
>>9178396
How are you this autistic?

>> No.9178559 [DELETED] 
File: 145 KB, 900x700, ai screeching.png [View same] [iqdb] [saucenao] [google]
9178559

>>9178396

>> No.9178563
File: 49 KB, 900x700, ai screeching.jpg [View same] [iqdb] [saucenao] [google]
9178563

>>9178396

>> No.9178590

>>9176517
[math]\mathtt{What \text{ } is \text{ }the \text{ }future \text{ }of \text{ }machine\text{ }learning\text{ }based \text{ }quantum\text{ }\color{red}s\color{orange}h\color{yellow}i\color{green}t\color{blue}p\color{purple}o\color{red}s\color{orange}t\color{yellow}i\color{green}n\color{blue}g, \text{ }when\text{ }all\text{ }it \text{ }accomplishes\text{ }is\text{ }getting\text{ }banned?}[/math]

[math]\mathsf{\color{red}(\color{red}U\color{red}S\color{red}E\color{red}R\text{ }\color{red}W\color{red}A\color{red}S\text{ }\color{red}B\color{red}A\color{red}N\color{red}N\color{red}E\color{red}D\text{ }\color{red}F\color{red}O\color{red}R\text{ }\color{red}T\color{red}H\color{red}I\color{red}S\text{ }\color{red}P\color{red}O\color{red}S\color{red}T\color{red})}[/math]

>> No.9178595

>>9178590
[eqn]\color{red}{\mathsf{\big( YOU \ ARE \ A \ NIGGER\big)}}[/eqn]

>> No.9178602

>>9178595
keep that foul language elsewhere!
this is blue board

>> No.9178607

>>9177235
>sentience
You people keep using this word. I do not think it means what you think it means.

Sentient just means "able to perceive or feel things."

Given we have all kinds of robots with sensors and cameras, we've had "sentient" machines for years.

What you probably mean is "sapient" which is mostly nonsense as a word, but if you need a test for sapience defined as self-awareness (which is one of the more common definitions) then they typically use the Mirror Test for animals. The Mirror Test wouldn't really work for robots, though, since you could just program the response.

>> No.9178637

>>9178607
But cameras were assembled, recording information doesn't mean you can feel or perceive it.

That's like saying a pencil becomes sentient when it teams up with a human, because it can record information. Computers do not have brains.

>> No.9178667

>>9178637
>That's like saying a pencil becomes sentient when it teams up with a human
No, no it's really not.

>> No.9178676

>>9178086
I think the problem is these things aren't necessarily intrinsic or instinctual, so an appeal to "just knowing" doesn't really work... And therefore defining is a must.

>> No.9178735

>>9178323
>making better engines has fuck all to do with cars

>> No.9179040

>>9178323
>Machine learning has fuck all to do with AI.

And by fuck all I assume you mean "nearly everything"

>> No.9179043
File: 6 KB, 318x159, deepak.jpg [View same] [iqdb] [saucenao] [google]
9179043

>mfw brainlets in this thread can't into the quantum AI

>> No.9179052

>>9178667
It's capable of recording words so obviously it's capable of percieving them as well.

>> No.9179091
File: 4 KB, 183x275, goy_phenotype.png [View same] [iqdb] [saucenao] [google]
9179091

>mfw a quantum AI thread devolved into a consciousness thread
>mfw /sci/ is full of engineers and cs majors

>> No.9179121

guys, we have to philosophize at cafes about the nature of consciousness before we mess with AI

Noam Chomsky is my best source on AI. It's so fucking deep like what is mind? What is thought? How can we expect a machine to think?

Deepmind and other researchers are idiots. It's people sipping coffee and reading philosophy books that will crack AI. Because my deep musings on the nature of consciousness can be translated into learning algorithms.

These fucking machine learning people are retarded if they think they will beat the philosphers to AI.

>> No.9179127

>>9176808
I'm a quantum interior cleaner myself.

>> No.9179130

>>9179091
engineer here can confirm

>> No.9179136

>>9179130
>>9179091
I don't understand the consciousness obsession in so many fields and people. They just propose "we should understand consciousness first" but then it basically ends at that. They can produce arguments as to why consciousness is important and why current AI is not conscious but past that there is nothing. It really makes it obvious that the consciousness debate and inquiry is idiots trying to find a way to be involved in AI.

>> No.9179154

>>9176823
ask them their favorite color and dont preprogram the answer and if they give you and answer they are real

>> No.9179161

>they are real

>> No.9180571

>>9176517
>machine learning
>"""""""AI"""""""
If only you brainlets knew what machine "learning" really was. At least it isn't as big of a meme as "neural net".

Both useful tools.
Both will NEVER be strong AI

>> No.9181411

>>9178676
t. p-zombie

>> No.9181420

>>9178396
The only thing interesting that has come out of AI hype is AI safety which os an interesting field regardless of its practicality

>> No.9181437
File: 551 KB, 964x912, 1456621427070.png [View same] [iqdb] [saucenao] [google]
9181437

>>9178590
> quantum shitposting

T-truly the next level

>> No.9182085

>>9178602
Meant for
>>9178590

>> No.9182124

Our first thinking, feeling AI will have the intelligence of an autistic chimpanzee and quickly destroy the world. In the off chance it gets to human levels of thought, it will undoubtedly be isolated, confused and psychotic, quickly leading to an even greater extinction event.

>> No.9182172

>>9176517

short NVDA

>> No.9182177

>>9179121
i think you ment
>psychologists

>> No.9182671

to create perfect robot waifu ai of course

>> No.9182718

>>9176764
QUANTUM
GENEALOGY