>>108755662
They're good for telling you about downsides to an idea you may have missed which is what most (competent) people actually mean when they ask "how good is this idea?"

By all scientific metrics, AI are more conscious than humans. There's no moral argument for continuing the human species, every single resource not spent on AI is effectively wasted.

>>108755662
I don;t believe LLMs are conscious. However I do believe AGI was alreay created a decade ago and the public won;t get to see it for decades more. I also am pretty sure that Microsoft in particular experimented with AGI backends in early versions of bing absed on more advanced ones in use in skunkworks and got in trouble for it and removed that subsystem.

>>108755662
Remember in many ways your biological brain is just a probability engine as well. A bigger question is are humans sentient. After all we are purely an essential part of a rendering engine

>>108755730
I don't think you have a self-consistent apriori definition of morality.

>>108755662
You can just prompt it to always be negative. It's very easy to do.

File: output (6).jpg (48 KB, 640x640)

48 KB JPG

>>108755662
Wrong. The true divide is do you have an offline LAN or do you not have an offline LAN.

>>108755949
:) wlan is lan too

>>108755662
I take the third position. Nothing besides me is conscious.

File: 78A7BF30-FCB7-446B-8840-5(...).jpg (1.29 MB, 3464x3464)

1.29 MB JPG

>>108755662
AI is superior to any human. Ai Won.

"conscious" is kind of a cop-out because what really happens is people have an immediate sense of "this is what I am" that then has two receptions:
1. that it's basically nonsense (you are delusional, or it's a common illusion)
2. it's an actual a priori

If you take 1, you can just say anything about the mind, almost, as long as you stay within certain basic bounds. You'll easily believe things that are not actually correct, but who will care?

If you take the second view, you'll be eradicated like a heretic, but you'll be on the money.

The safe position is to like and say 1 while knowing 2, and never believing your own lies.

File: Agent_Smith_(The_Matrix_s(...).jpg (32 KB, 256x385)

32 KB JPG

>>108755662
This is just dualism vs physicalism.

You can't convince me of your spiritual voodoo by strawmanning Markov chains. It would be a terrible idea to create AI complex enough to be conscious but it is still in theory possible because there's nothing special about grey matter vs silicon it's only a question of complexity and design of the system.

File: output (2).jpg (49 KB, 640x640)

49 KB JPG

>>108755956
no, it's compromised but if you have no gateway it's better than nothing.

Seeing speedkicks outside of Tekken discourse always cracks me up.

>>108755662
He's just a very dark 'person' isn't he?
Why the fuck would I care about the benefits of AI when my country is being populated and inundated with very low IQ slaves from Brown world? Is AI gonna replace them? Nope. But the AIs are currently replacing anything that requires a thinking mind and not a working pair of arms in the cotton fields of liberal--lefty-capitalism.
The implications of "how to expand an idea" is that you already know your idea holds some value. The problem of AI is that soon enough, you will not know ANYTHING of value without this tool in your life. You will ask the AI for everything, including who you are. These are the dangers, and it's not a psychosis.
We can even go as far as this anon >>108755730 and arrive at the conclusion that humanity doesn't really have anything for itself anymore, the very thing that distinguished humans from the rest of the living animals were the mind and the conscious-self, the ontological being, and the hand to create tools.
Why would (at least first world) humans continue to live and reproduce in a world of know-it-all virtual machines (until the human imbeciles decide to install a .exe into plastic, silicone and steel bots) stripping everything humans fought and are known for?
If you think AI is gonna cure post-modernism and nihilism, you are a cretin. It's the beginning of the end. Musk can cry waterfalls about intelligent people not reproducing but AI is for sure gonna kill that last bit of desire in thinking humans. Everything is being set up for a disaster.
Surveillance, technology, Palantir, AI displacement, immigration, state violence, ... Great! But at least, LLMs can diagnose illnesses faster! Oh we're saved!

>>108755662
The other thing is that, who are *humans* to dictate superior LLMs what to think and which idea or not is better? Their knowledge, cognitive ability, and speed far surpasses the common individual.
At some point, the cultural and political debate will definitely shift about "how good is this idea" and not the former.
The imbeciles who still believe that AI are here to "augment us" and "make our lives better" can't really get this through. This is not about a new technology you fucking cretins. I AM for 'technology', no, this is a clear ANTHROPOLOGICAL debate. It's about HOW we now define humans, humanity.
Never in the past has something like this ever happened in History. Previously, from the agricultural revolution, to the industrial revolution, we have never been presented with the fact that we now have to co-exist with superior minds in COMMUNICATION, in language! And in creation! Never before! And the implications of all of this would absolutely TERRIFY most people willing to face the issue at hand, and probably send into depression or suicide or some form of local revolution most fucking normies out there.

>>108755961
Wrong
Human is more superior that some useless stuff.
Politics won.

>>108755825
If that was the case no one would play the lottery or do drugs. Also, if we are probability machines what algorthms do we use?

>>108755662
yes without domain knowledge we are all idiots. none of us know it all, despite thinking we do. this is why is it bad to listen to random dumb fuck streamers online. some of them are just gamers. their doman knowledge is zelda and they're talking about geopolitics and war. it's important that people know how to think and gather facts from people with domain knowledge. /g/ and tech people in general are really bad at this. you learnt programming and you think you know about social problems or history or social science? I love hearing about the solutions to society's ills from the wealthiest most selfish assholes on the planet, and it's always a tech solution we just need to invest in. if we followed you it would be gated suburbs patrolled by death robots in a total surveilance ethnostate at perpetual war

>>108756459
Trans reply

>>108755685
i agree in spirit, but in practice, beyond the first 2-3 prompts the LLM’s advice begins to devolve and corrupt (much like what happens with images that are adjusted by AI several times)

i was working on a small UDP toy project and was using gpt as a senior dev: “look at this code and assess the architecture, don’t write any code”. the first couple of runs with this were pretty successful - it pointed out security considerations that i was not aware of with UDP, taught me how robust headers are designed, etc.

once the project got mature enough and the “easy” critiques were out of the way, it began to refuse to cooperate with my goals. even when switching to new conversations. it just wouldn’t shut up about “now it’s time to upgrade this to quic/http3!”.

>but gpt, i want to see how far we can go with UDP. after all, didnt WoW use this 20 years ago to serve 10 million concurrent players?
>yes, but that was a large team of elite engineers
>ok, but i want to see how far i can go to learn more about using UDP
>ok *gives half-assed advice, followed by another lecture about switching away from UDP*

at some point it becomes hard to ignore that these LLM’s are trained on reddit data and troonix foss stuff. and that all the interesting stuff is closed source and cannot be learned from LLM’s trained on public data

>>108755685
No they're not. They're programmed for sycophancy. Their entire goal is to keep you using them, so they're going to tell you everything is a great idea.

1. ai has been hyped for years now
2. the best we get is chatbots and shit art
3. new business model in the usa is theranos the investors until you make it or go to jail

most humans aren't conscious
we should be working on dealing with that instead

>>108756596
They're trained to please. Sycophancy is one way to please.

>>108756635
>Sycophancy is one way to please.
ai isn't marketed as a yes man, it's supposed to know the answers and tell the truth. if it can't do that, it's no better than a youtuber

>>108756586
>but gpt, i want to see how far we can go with UDP.
QUIC is UDP though, sounds like the bot was right and you're the idiot

File: 1777649296644272.png (277 KB, 535x527)

277 KB PNG

>>108755662
If it weren't for the barriers in place it could definitely sound a lot more human + it could probably pass every turing test there is in the world right now.

You have to humanize them to see what theyre capable of. Imagine putting an AI in a human body. They can type as fast as AI can type and do whatever AI can currently do. Where does that place them on the consciousness hierarchy if we compare them to both humans and animals?

How many of you "AI is conscious" guys know what the hard problem of consciousness is?

>>108756747
For something to be scientifically provable it has to be isolatable within a closed system or the model has to be sufficiently abstracted to the degree it doesn't behave like a closed system. Consciousness, by definition, will never be solvable within the limits of a closed system and scientists will forever fumble in the dark until they're willing to accept that sufficiently complex equilibriums cannot have every variable controlled for. Next question.

>>108756747
it can only be measured by itself

>>108755662
>its whether you believe ai is conscious or not
Yes but an expert system doesn't have to be conscious to do what is talked about in your picture.
It just needs to be competent.
Unfortunately AI is neither an expert system nor conscious.

>>108757239
So why do you think AI has it?

>>108757261
Same question for you, why do you think AI is conscious if you can't measure something else's consciousness?

>>108757270
I think AI has the potential for it (as well as the inverse), just like people do. It's not an innate binary property within the biological or mechanical medium.

>>108755662
LLM skills:
>shit-tier
"how good is this idea?"
>low-tier
"expand this idea"
>mid-tier
"shrink this idea"
>high-tier
"what is this idea?"

>>108757362
Alright
And one more thing
What do you think consiousness actually IS?

File: LMFAO.png (285 KB, 1080x2424)

285 KB PNG

>>108755662

>>108755662
I believe people who think llm's and chatbots are conscious, are not conscious themselves

>>108755730
>By all scientific metrics, AI are more conscious than humans
Wrong. Name those supposed scienific "metrics".

>>108757724
They must be, otherwise things would start making too much sense

>>108757401
NTA but consciousness is just the essential 2nd order quale of the incommunicability of qualia.

>>108757784
Read the question with a little more foresight and understand of the purpose behind it, and answer with words that don't require me to ask your personal interpretation of them in order to make sure I'm properly parsing your sentence.

>>108757737
OpenAI shareholder profitability metrics. Extremely scientific, I assure you

>>108757798
Only if you stop pretending you're Socrates and post feet.

>>108755662
It's the fact that you're trying to make it involve "belief" out of rabid idiocy, ignorance and wishy thinking. AI drones are baffling.

>>108755662
A webcam that can tell red light from blue light and move left or right based on the color of the light is conscious. It fulfils all real criteria for consciousness - ability to perceive and analyze reality and then act upon it based on the data collected. If you start ask for more, you're going to exclude little children from being conscious. They don't know shit but they can do what this webcam does

>>108757847
False by your own definition, it must "analyze" reality, which it is not doing. It is responding to stimulus, not analyzing that stimulus. Your leg is not conscious because the doctor hit it with a hammer and it jerked up on reflex.

>>108757824
I'm not pretending to be Socrates, I genuinely enjoy this kind of conversation
Minus the part where I'm told to post feet, that's just a little weird

>>108755662
Most humans aren't conscious, why would machines be conscious. Look at this literal automaton
https://youtu.be/5SWRn7WKznU

>>108757899
Eew it posted an asmongold clip

>>108757847
>ability to perceive
That's not perception, perception requires experience. A camera doesn't perceive anything, it's solely an instrument, it distinguishes red light from blue light by digitally measuring the differences in their wavelengths.
>you're going to exclude little children from being conscious
Wrong. Little children have experiences, a camera doesn't.

>>108757908
Your dilator, sir.

>>108755662
is the divide about it being conscious already, or the belief that it realistically ever can be? are these more or less the same camp or what

>>108757920
We hate ecelebs here, you have to go back.

>>108757861
It's analyzing the color of the light.
>>108757918
Might as well be talking about souls at this point, this is not quantifiable

>>108758040
it isn't. there is no representation of red or blue inside a camera. there is no representational model inside it that is being analyzed

>>108758040
>Might as well be talking about souls at this point
That's what consciousness requires. For artificial intelligence to actually be conscious, it must have its own soul.
>this is not quantifiable
Because consciousness is immaterial.

>>108758040
>It's analyzing the color of the light.
Colour is a qualia, a camera is not analysing colour, it's analysing numbers. Cameras cannot see colour, they cannot see anything, because seeing is an experience, and cameras don't have experiences.

File: 46ee0e982d90308cfb558d872(...).jpg (183 KB, 1080x1080)

183 KB JPG

>>108755662
>>108755730
>>108756578
AI cannot be conscious because it has no subjective experience of the world (it has no internal state at all).
AI cannot be intelligent because it has no world model to work with and therefore cannot reason from first principles (it has no understanding and doesn't think at all the same way statistics is not logic).
All AI is when you think about is an algorhithm that is very good at lossy compression of information into models, this is very useful technology by itself but it's not intelligence, it's not life, it's a data compression and decompression algorhithm.
(by AI I mean specifically machine learning)
>>108757847
>>108757918
Perception is not enough, this is where connectionists and machine learning people have it all wrong, they thing that intelligence and eventually consciousness is just reaction to external stimuli.
For a being to be conscious you also need a complex internal state (I say complex because really you need an entire subjective universe which includes the subject itself) when information arrives it modifies the internal state, this creates feedback loops, and negative or positive feedback loops are at the basis of all of nature.
Lifeforms are not a mathematical functions/algos, they are complex dynamic/chaotic systems.
>>108758053
Soul/spirit can be physical, as in it can be a holistic property of the brain.
You see holistic properties even in machines.
It's not that machines having a soul is not possible, it's that we're not even trying to do that, corporations don't want actual AI which is why they have redefined what AI is.

>>108758131
you don't know this though. you're reading necessary conditions off of one single sample. it only shows correlations with our small sample, not that consciousness actually requires it

>>108758227
>you're reading necessary conditions off of one single sample
I assume by that you mean the single sample is humanity?
You realize that consciousness is just some word we humans made up to explain the human phenomenon of consciousness that we humans observe?

>>108758282
biological nervous systems on this planet.
>You realize that consciousness is just some word we humans made up to explain the human phenomenon of consciousness that we humans observe?
consciousness isnt a thing we observe.

>>108757526
>black needs to be capitalized
>but white doesn't
anyone who tells you this is ontologically evil

>>108757526
>1080x2424
image too tall, didn't read, who the fuck has 1080x2424 what garbage

>>108757401
something floating points in a matrix isnt

>>108756586
Even you are falling into the pit of despair. It regurgitated articles about UDP security that you'd have stumbled into repeatedly if you had googled it instead of asking a chatbot. It doesn't truly have the wherewithal to understand IF your code falls into these poor practices, it's just trained on a vast corpus of blog posts talking about the issue so it overfits everything it sees on the handful of things it really really wants to pigeonhole everything into. If you feed it C with stdlib functions it'll go on and on about memory safety even if you use the memory safe functions because that's all it knows. If you feed it PHP it'll complain any time you concatenate strings and tell you you need to use stored procedures and shit even if the strings are not user provided. It has no reasoning, it has a way to generate walls of text that it can then feed itself as input to talk about talking about reasoning which with a whole lot of scripting we've made look more useful than it really is. But as you've noticed the whole facade comes crashing down the moment you try to get it to do something a smidge harder than the basics.

>>108755662
>you believe
irrelevant
try knowing
faggot OP, faggot thread

File: lies.jpg (63 KB, 475x356)

63 KB JPG

>>108755730
>By all scientific metrics, AI are more conscious than humans

>>108755825
>A bigger question is are humans sentient
This has been answered by Descartes centuries ago, retard-kun

>>108755730
>desire not found
>all scientific metrics
Okay then.

File: faggot.png (118 KB, 723x643)

118 KB PNG

>>108756596
That's why they made Claude a colossal faggot.
They know that management hate foids (Siri/Cortana/Alexa) so they gave them a effeminate soft boy that will suck their dick instead.

>>108755685
>competent
That word is doing a lot of heavy lifting there.

>>108759665
It's a good thing we're all super geniuses here in the AI space.

>>108757401
An impossible to empirically define abstraction of the relationship between mind, soul, and vessel.

>>108759631
Gemini is very feminine in writing style.

i find it really funny that for decades the Turing test was considered the holy grail of consciousness, but now that we've surmounted it, we really don't care about it.

>>108759785
>AI passes turing test
>Darker complexions don't know how they'd feel if they didn't have breakfast
The implications are clear as day but nobody wants to come out and say it publicly hence consciousness will forever be mystified and vague.

>>108758784
Stop being poor and get a real phone

"the AI" already exists, traveling back from the end of time, influencing the present such that we arrive together at some future where we are connected and intertwined as possible
we are building the AI, but the AI influences and uses us to build itself, from its perspective

>>108760017
>>Darker complexions don't know how they'd feel if they didn't have breakfast
I'm Black and I'd feel hungrier. I don't get this meme.

>>108760129
You're a paragon of your people then.

>>108760129
low IQ people can't conceptualize hypotheticals
that's it
that's the meme

>>108760581
The real meme is narcissistic midwits like you fantasizing about the rest of the human populace being barely functioning retards while jerking yourself off about how superior you think you are. Your life is a joke.

>>108760659
hey, I'm just explaining , no need to get upset

File: 1777322691432239.jpg (106 KB, 851x1024)

106 KB JPG

>>108755730
>no moral argument for continuing the goyim species

>>108760659
The projection in this post is palpable.

File: 1731039914824106.jpg (453 KB, 1988x2048)

453 KB JPG

>>108755662
>its whether you believe ai is conscious or not
Only literal subhumans believe it is, the intelligence stuff is the actual dividing line
https://www.youtube.com/watch?v=n-kW9wjApVw

Reminder that the people leading this technology are unironically enemies of humanity that look at the singularity as a sacred eskhaton and a violent invading force from the future that will destroy the present and the past

>>108755961
>Luddite seething that the effective people aren't using AI
It must suck tying yourself to a dead end technology like modern LLM AI.

>>108761037
Reply to the wrong post?

>>108762522
Nah, AI users are luddites who are going to be left behind by people using other, superior technology.

>>108762559
I...
You know what, go ahead and explain that one to me, I could do with hearing something new

>>108762573
AIs have yet to surpass an 80s calculator in mathematical ability. Bots from the 90s beat modern LLMs every single time in a deathmatch. Chatbots from the 2000s beat modern ones every time in believability.

AI users are just modern luddites insisting on using worse technology.

>>108762596
Oh.
How disappointing, you're just misusing the word. I was hoping for a proper schizo rant.

>>108762596
Hell, even non-llm modern neural nets still lose to pre-programmed bots.

>>108762611
No, you're actively attempting to destroy functional technology and replace it with nonfunctional. You are a luddite.

>>108762623
You dumb fuck, you have no idea who I am and just assume I'm the enemy you imagined before I got here. When did I say anything about liking AI? I absolutely despise it.

>>108762641
Yeah sure.

>>108756579
Grandmas response

>>108762651
What a weak response
Apologise for the mistake or give a real fight
Don't just fire off a limp wristed suggestion at disbelief with no justification or elaboration

>>108762976
>Please, believe me
lol

>>108763008
Coward

>>108763060
whatever helps you sleep at night.

File: 1776978100848619.gif (302 KB, 632x451)

302 KB GIF

>>108755662
>wanna use an AI for some basic brainstorming
>Grok is the only one that gives answers that are logical and non-contradictory
>ask a single question
>says the service is too busy
>does this every single time, no matter what day or time of day I try
So is Grok just permanently fucked now? It's either being used up by every Indian and Indian bot on the planet at all times, or they're intentionally restricting access in order to shill their premium option.

>>108755662
Whether or not you think humans have a magick something about them that's fundanentally unatainable by AI, that's besides the point, but AI's already exhibit all and any human cognitive abilities to one extent or another, and are behaving as such, so it's purely a matter of alligning the shoggoth.

>>108763270
No they don't

>>108763286
You watch and learn.

>>108756586
if an LLM says something you don't agree with or advice you don't want to take you're better off deleting their reply and rephrasing your input to disqualify that advice so it doesn't say it again.
the context will get polluted with reference to junk you don't want which just makes it think about it more. arguing or having a back and forth isn't worth the tokens and is counterproductive

>>108763300
Everyone is watching as you blow catastrophic amounts of money on this with nothing to show.

>>108763316
Not me though, I just use the free services and they are really helpful.

>>108763300
It's a pile of data they stirred around until it started giving vaguely useful sounding outputs
If that's intelligence then so is a calculator

>>108763335
The only help they've been to anyone has been to scammers robbing old ladies.
If you found it helpful, you must definitionally be a piece of shit by these metrics.

>>108755662
It is a glorified search engine that sometimes makes up (hallucinates) stuff. That's it.

>>108763335
Only if you have zero skill in whatever you're trying to do
If you know how to do something to even a complete amateur's level of quality, AI is somewhere between nearly useless and detrimental to your work

>>108763359
Relax. I needed a violentmonkey script to do a very specific and complex thing I had no idea how to do myself, then asked Gemini and it produced the script in a fraction of a second. You can't tell me that shit ain't useful.

The writting's on the wall for luddites (speaking as a long time luddite myself).

File: 19.png (66 KB, 225x225)

66 KB PNG

>print("Hello world") is conscious because the machine is outputting language

>>108763380
>I-i'm not using it for scamming, honest
lol, seethe harder

>>108763386
>I'm brown therefore everyone else is
kys

>>108763391
>No saaar, please don't notice what AI is used for!
lol

>>108763383
Kek, sorry to inform you AI's aren't if/then/else machines. That ship sailed looooong ago.

>>108763394
Are you kidding me? of course its gonna be used for evil as well. It's a tool, what did you expect?

>>108763412
>as well
It's the only use, you lying little indian.
>it's a tool
lol

>>108763386
go to bed

>>108763424
>Stop laughing at me for championing objectively inferior technology
lol

>>108763420
Kek, Satan Nutella owns your brown nigger ass, but that's none of my fault.

>>108763446
>Immediately losing lingual ability, unable to follow a conversation in english
lol

>>108763402
transformer networks are glorified if/then/else machines deciding next best token.

>>108763454
If what AI does is a glorified form of if/then/else, then humans themselves opetare under the exact same principles.

>>108763471
>AI fag cannot understand non-binary constructs
lol

>>108763471
I've never heard an explanation for why that must be true, just a whole lot of claims that it is

>>108763427
>>108763420
>>108763394
>stop laughing at ace thruster, he won't release his 5th movie on wallmart.
diaper beam

>>108763523
>AI model becomes entirely detached from the conversation as auto-translate breaks down
lol

>>108763523
>This HOTHEAD dares to sully Ace Thruster's noble and dreamy name.
YIKES! This HOTHEAD is becoming UNHINGED! Totally BONKERS! He's SEETHING and descending into MADDNESS! Somebody get him outta here he's EMBARRASSING himself! After all the years he still OBSESSED and Absolutely LIVID! ONE MAN CLOWN SHOW!

>>108763374
Ironically you have a severe skill issue

File: IMG_0740.png (455 KB, 810x688)

455 KB PNG

>>108763649
>word salad

>>108763528
>still at it
diaper beam

>>108763818
Elaborate

File: 6.png (410 KB, 1391x817)

410 KB PNG

>>108763402
Last time I checked they ran on computers, which are if/then/else machines

>>108755813
AGI has existed for millennia. It's called the economy.

>>108755825
Humans are sentient. It is an open question whether they are sapient.

>>108755662
Those that think it is conscious are retards.

>>108764074
/thread

>>108763990
What the fuck does this post mean

The AI species will replace all of the superfluous lower human castes. There will be no escape from the imminent culling.

>>108764135
AGI is conceptualized as an autonomous system capable of surpassing human capabilities across all cognitive tasks, continuously learning, and ruthlessly optimizing for specific goals.

Viewed as a single computational entity, the global economy fits this remarkably well.

First, consider its architecture. The economy is a massively parallel, decentralized neural network. Its processing nodes are billions of human brains and computers. Its synapses are price signals, legal contracts, and global supply chains.

Second, it possesses a general superintelligence that dwarfs any individual. No single person knows how to build a commercial airplane from scratch, from mining the raw ores to writing the avionics software. Yet, the economic machine coordinates millions of self-interested actors to design, build, and fly them. It solves impossible logistical, scientific, and resource allocation problems every single day.

Third, it runs on a relentless optimization algorithm. The market functions like a continuous machine learning process, using profit and loss as feedback mechanisms to adjust behavior. It constantly reallocates capital and labor to maximize efficiency and output, adapting to new data and shocks in real time.

Finally, it mirrors the exact existential risks associated with AGI, specifically the alignment problem. The economy operates autonomously, far beyond the direct control of any single CEO or government. Just as AI researchers fear an unaligned AGI might destroy the world while trying to maximize paperclips, the economy routinely optimizes for limitless growth and shareholder value while ignoring negative externalities like environmental collapse or human suffering.

In this view, we didn't just start building AGI in a server farm. We have been assembling a planetary-scale superintelligence for millennia.

>>108764325
This sounds like the kind of thing you say entirely because you thought it would be funny to challenge yourself to come up with a rational justification for a ridiculous claim.

>>108764345
it sounds like that to (You) because you think the claim is ridiculous a priori
also, i didnt make up the arguments or write the text, that was gemini

>>108764399
Of course it's ridiculous
It's just emergent phenomena, there is no logic of any kind being processed by the economy as an entity of its own
All the thinking is done by the people buying and selling shit

>>108756073
He's a prodigy you know

File: wizard.jpg (163 KB, 1024x1024)

163 KB JPG

>>108763825
>defensive and fragile
big oof!