It's """AI""" shilling day today. They always come in waves.

>>23062934 (OP)
these retards come to /pol/ too?
you got my compassion

>>23062934 (OP)
AI and a kike shilling day today

>>23062935
Why wouldn't they come here? The anti-human agenda is politics.

>>23062934 (OP)
holy shit, you are so far behind. wake up or you are gonna be left in obscurity

>>23062938
This poster is a bot. This is a fully generic talking point they spam in slight variations.

>>23062934 (OP)
There is no if-then-else in neural networks. It's just a bunch of matrix multiplications and additions.

>>23062940
>There is no if-then-else in neural networks.
I can implement a decision tree on a GPU without explicit branching. Doesn't make it any more powerful.

>>23062937
its not anti human agenda
theyre drumming up investor hype
no, this would be anti-human agenda in the context of ai:
https://www.youtube.com/watch?v=VdG4gUTowXc

even though this is just a company that created an ai engineering assistant
but this is foundational to ai engineers
an ai that gets told what you want from it and it shits out a whole industrial process

the czinger ai has all the elements needed to be a full enginner ai
what it lacks is encoding all the industrial processes under the sun which would be a titanic endeavour
but its systematic work. the discovery part, and inventing how to make it work has already been achieved

>>23062942
>its not anti human agenda
It is absolutely an anti-human agenda. All the people promoting it are explicitly anti-human to the point of schizophrenic obsession with "AI wives" replacing women, the fantasy of making everybody poor and unemployable, trying to "prove" that biological brains are just like babby's first neural network etc.

>>23062940
an ai doesnt need to be a neural network
and you most often have a symbolic part for one function or another, even in chatbots
in stuff like waymo the core of its ai is a symbolic decision tree

>>23062943
>to the point of schizophrenic obsession with "AI wives"
didnt think about that angle
i just metally classified them as human detritus and left it at that
do you really think its that impactful?
i mean, compared to all the other vectors of attack on the european people's fertility rates?
i think many of them wouldnt have bred anyways
i think one needs to be already mentally insane to get ai psychosis to begin with

>>23062945
>do you really think its that impactful?
>i mean, compared to all the other vectors of attack on the european people's fertility rates?
I don't think it will have direct impact on fertility rates. It's only symptom of a shared underlying cause. I'm talking about the psychological engineering. You can look at this "human detritus" as the canaries in the coalmine of AI schizophrenia. Treating machines like they're human is being gradually normalized and the other side of this coin is treating humans like they're machines.

>>23062943
>All the people promoting it are explicitly anti-human to the point of schizophrenic obsession
or maybe thats the attack
muddy the waters enough that people dont understand what an ai is
and then mystify the whole thing to create distractions, or panic
people make bad decisions when under pressure
its a classic of the genre to elicit fear to force political changes
when people are afraid they act brashly, irrationally

>>23062944
>It's """AI""" shilling day today.
>if then else
>an ai doesnt need to be a neural network
Who shills for symbolic AI today? Does context no longer matter in conversations in 2025?

>>23062948
>i don't understand that most neural networks are equivalent to decision trees
>therefore you are wrong

>>23062941
>I can implement a decision tree on a GPU without explicit branching.
Is nonlinear vector transformation a decision tree? What does it mean for a function to be differentiable?

>>23062946
>Treating machines like they're human is being gradually normalized and the other side of this coin is treating humans like they're machines.
interesting take
although in my corner of the world people dont take too kindly to the latter
picrel is from the coofid era
people dont like to be treated like machines

>>23062949
>>23062950
I've already humiliated you to shit, but you're too stupid to comprehend it.

>>23062947
>muddy the waters enough that people dont understand what an ai is
But this has been the case for decades. What is "AI"? The minmax algorithm is "AI". Path-finding is "AI". Doom monsters are "AI". The term doesn't mean anything, objectively. If a program is doing something that normies associate with human cognition, they will call it "AI".

>>23062948
for starters, good products dont get shilled
normies get the claudes and the copilots
aerospace industry gets to play with this:
https://en.wikipedia.org/wiki/Model-based_design

and like i said
theres a symbolic COMPONENT to most ais on the market

>>23062950
>Is nonlinear vector transformation a decision tree?
If that transformation is just the composition of functions associated with a FFN's layers, then it's equivalent to one. Your ignorance doesn't make you right.

>>23062951
>although in my corner of the world people dont take too kindly to the latter
>picrel is from the coofid era
>people dont like to be treated like machines
Well, you see, most people aren't cutting-edge reddit intellectuals. They don't "understand" yet how their brains are just like ANNs. But this trickles down like countless other elements of the modern, "enlightened" intellectual fashion.

>>23062954
>theres a symbolic COMPONENT to most ais on the market
From the 650GB LLM how big is the symbolic part? Less than 0.01% realistically? You're grasping at straws.

>>23062953
>The term doesn't mean anything, objectively.
i disagree
ai is anything that takes decisions
you can always have a program that takes 2 parameters and outputs the sum, and it unequivocally *wont be an ai

but normies are fed bullshit as in ai = asi
and since their brains are scrambled with bullshit catastrophe scenarios can be deployed and in result half of twitter either fears skynet or awaits its coming expending time and energy chatting back and forth about what essentially is fanfiction

i talk with these retards pretty often
i think the core mechanism is that people accumulate bullshit information and then think its something valid because of the quantity of said bullshit and repetition on social media
then they defend their """"knowledge"""" aka the ensemble of yudkovsky's tweets because it cost em effort to get where theyre at
i think its similar to how macfags think/appo marketing
>asinine gui therefore it must be superior

>>23062955
Again: What does it mean for a function to be differentiable?
At this point, I'm not even sure you know what a function is.

>>23062958
>ai is anything that takes decisions
Bubble sort makes decisions. Is that AI? Or maybe deciding between numbers doesn't count under whatever vague idea of "decision" you're delegating the definitional problem to?

>i think the core mechanism is that people accumulate bullshit information and then think its something valid >because of the quantity of said bullshit and repetition on social media
Absolutely. And in some sense, this is a convergence between Man and Machine via the descent of man. Normies are now trained in vaguely the same way chatbots are.

>>23062957
>Less than 0.01% realistically? You're grasping at straws.
volume is completely unimportant, its the function that matters
red dead redemption 2 is 100GB
code is probably somewhere around 0.5% of that and the whole business logic proper is probably in the kilobytes to 1 MB range


>>23062956
>But this trickles down like countless other elements of the modern, "enlightened" intellectual fashion.
i think its useful that we counteract that by explaining this shit to people around us
its a good thing that normies are anti ai in general
many think its bullshit
many because of the danger it could pose to jobs

>>23062950
>Is nonlinear vector transformation a decision tree?
If that transformation is just the composition of functions associated with a FFN's layers, then it's equivalent to one. Your ignorance doesn't make you right.

>>23062959
I'm not seeing any refutation. Should I accept your concession and move on?

>>23062961
>>23062962
It's funny how you’re flailing.

>>23062960
>Bubble sort makes decisions. Is that AI?
that would be bending the definition a little, but yes

obviously the term gains its usefulness when in a relative context
mere error handling most often involves conditions
but when in the context of a game you wont be calling the whole program "ai"
you will use that term to relate to the part of the business logic that involves npc behaviour

>Normies are now trained in vaguely the same way chatbots are.
totally agree. its a globohomo fantasy that keeps popping its head now and again
https://en.wikipedia.org/wiki/Humanzee
it never works though
i think an aspect of the demographic replacement is to bring humans down
their terrible miscalculation is that brownoids are 10x more unruly than whites

AI models have a tendency to be released in waves

>>23062963
>you guys clearly know what youre talking about but allow me to reiterate what yudkovsy said on twatter

>>23062936
Anyone who disagrees with you is a shill. Seems like life is going to be real hard for you friend. Too bad so sad. Oh well.

>>23062963
Does this function look differentiable to you, tard?
Protip: the next instruction in your NPC programming is to deflect.

>>23062964
>obviously the term gains its usefulness when in a relative context
A term gains its usefulness when it has a definition that formalizes intuitive distinctions within its domain of applicability. I don't see how a concept of "AI", where almost any branching algorithm is "AI", accomplishes that.

>i think an aspect of the demographic replacement is to bring humans down
>their terrible miscalculation is that brownoids are 10x more unruly than whites
My bet is that they're importing hordes of shitskins BECAUSE they are unruly. That way they can install a police state and claim it's for your protection. The endgame is to cull most of the human population regardless of race. They believe they can just gene-edit humans grown in artificial wombs to create as many compliant scientists and engineers as they need.

>>23062969
>A term gains its usefulness when it has a definition that formalizes intuitive distinctions within its domain of applicability.
thats what it is for me, its just that it isnt an absolute term
its like "hot" or "cold"
in a vacuum, i would call memory management heuristics "ai"
but when writing a game, i wouldnt. theres another much more ai-like element in that project

>My bet is that they're importing hordes of shitskins BECAUSE they are unruly. That way they can install a police state and claim it's for your protection.
yeah, but who's gonna be doing the policing?
have you seen africa?
i mean... i really dont see it. i think they overestimate their ability to control people
i think unrests in europe and their pathetic attempt at censoring political discourse are a symptom of that

>>23062970
>i think they overestimate their ability to control people
I think you underestimate their ability to control people

>>23062970
You don't see how your stance contributes to the problem you're calling out? It specifically allows for the conflation between Big Yud's fantasy entities and Stockfish. Normies have simple, label-driven minds that go by vague associations. If you conflate different things under the same label, you conflate the associations and it's the more "impressive" ones that stick.

>yeah, but who's gonna be doing the policing?
They're just gonna kill loads of people off in engineered wars and pandemics. Then they won't have to work so hard to police the survivors.

>i think unrests in europe and their pathetic attempt at censoring political discourse are a symptom of that
I'm pretty sure the growing unrest is why they've been working so hard lately to streamline the path to WW3.

>>23062972
yea you say that because you base your wordview on your lived experience
you shouldnt
america is an exceptional country in the world
in that it has nigh zero national cohesion
your whole ethos was formed by proto globalists
youre brainwashed since birth to be isolated, disorganized
>you know how it isnt an american protest?
>the people wave the flags of their own country

>>23062973
>Normies have simple, label-driven minds that go by vague associations
eh, dont worry
i adjust my terms to my interlocutor
rn im chatting with you so i allow myself some dialectic autism
this is a skitzo neighborhood
normies tend to stay away

>They're just gonna kill loads of people off in engineered wars and pandemics. Then they won't have to work so hard to police the survivors.
thats the intent, but this isnt gonna work with a completely dysfunctional system
and given the point where tech is at, esp in robotics, theyre gonna be facing serious challenges
robo policemen are 10-20 years away
but the EU will collapse in 5, either way. jihadi uprising, or legislative dissolution of the eu, or even globalists hanged under martial law
as far as i can see things are coming to a head
and ww3 is the toughest of sales especially with internal rapefugee driven conflict and the spectre of the jihadists looming over it all
i really think they bit off more than they can chew
their plan looks sound, but their timing didnt pan out, and they didnt let themselves any wiggle room

>I'm pretty sure the growing unrest is why they've been working so hard lately to streamline the path to WW3.
thats literally the nuclear option.
nobody wants to sit in a bunker for the next 20 years or so
but i think they overextended and they dont have enough control to force ww3
jihadist factions are a legit force to be reckoned with, politically and militarily
talks of ww3 united left and right in france
germany wants to keep the status quo because theyr created a net of influences throughout europe
poland is on the verge of a general strike because of their minority government
england is rising, and i dont think they will be stopping anytime soon

i really think they fucked up
they should have waited 10 or 20 more years and the demographics would have made an uprising completely impossible

>>23062973
>i really think they fucked up
and exactly by overestimating their ability to control people
the spark that set fire to the powders was the rapings, the killings, the pillaging
because they thought welcoming desert people with tea and biscuits will suddenly make them civilized and agreeable

no
theyll fuck you in the ass
sometimes quite literally

they only respect strength
the whole fucking policy was a disaster waiting to happen

>>23062974
I agree that America is a special kind of retarded, but my point was that the unruliness is a feature not a bug

>justifies crackdown
>divides opposition
>distracts from root issue
>escalation = expansion of problem

>>23062975
I'm not worried. One way or another, sinners against nature will get what they deserve. If not today, then tomorrow. If not tomorrow, then in two more weeks.

>>23062978
again, they miscalculated because that awakened european peoples
their politics throughout the years were always those of soft power
because of a key feature of european peoples:
the harder you oppress us, the harder we will resist
had they been more careful with who they let in
had they taken more time
had they not been greedy retards
and retardation is a key part of all this
"LE business leaders" were lulled into this plan with the promise of geometrically growing markets (= buy shares and you can sleep on em forever)

everything would have been controlled demolition
without risk
without fuss
just a silent dissolution into the night

but now they have to juggle jihadists
fukken nationalists
and even many corrupt politicians realize theyre in the pot alongside the rest of us and change their allegiances
given what i know, the situation is veeeeery suboptimal for them

>>23062979
>sinners against nature will get what they deserve
aptly said
although im more in favor of a more proactive approach.

>>23062936
I've seen a lot of tranny threads today

>>23062980
I hope you’re right.

My fear is that the jihadist and nationalist movements were engineered to be impotent. Jews play all sides, and normies usually can’t see it.

See: Gatestone Institute, David Horowitz Freedom Center, Robert Shillman, Peter Thiel, Middle East Forum, etc — all have donated to far-right influencers, organizations, and movements. I think they amplify certain people to create the illusion of an organic rebellion.

As I said, hope you’re right!

>>23062978
also globalists are not a united block
ill tell you something new for pol from what i see
poland has a key geopolitical importance in europe
it is in the crosshairs of everyone
because of its geographical position, also, but not only in relation to the ukraine war

in big part too because of its societal changes, because its one of the biggest in europe, theres many stakes at play in poland

but whats interesting about it right now is that you see two globalist camps vying for the control of poland
the jewish and the american protestant lobbies
jews are represented by donald tusk
wasps by holownia and sikorski
initially they were allied, but given the response of society at large to immigration policies, they didnt waste a second to stab donald tusk in the back
and theyre very close to americas ruling families
holownia is married into the brzezinski clan
THOSE brzezinskis
sikorski had close ties with the cia
he was in afghanistan in the 90 for some reason

tusk on the other hand executes all the german policies without flinching
my working theory is that globohomo is not a monolithic construct and that trump represents the wasp lobby. the other dominant lobby being the zionist jews
with both having often aligned, but ultimately opposite interests.

like globalism good for both bc kalergi for the jews and open markets for the wasps

>>23062934 (OP)
Contemporary AI (like GPT models, image recognition, etc.) uses machine learning, neural networks, and statistical inference, which are far beyond simple IF-THEN rules.
While the meme is true for some older or shallow systems, it doesn’t reflect the complexity of large-scale deep learning models or reinforcement learning agents.

>>23062985
>far beyond
By which you mean "exactly equivalent".

>>23062983
>As I said, hope you’re right!
like i said,
thats what comes form my observations
but im right more often than wrong and im pretty sure about this

HOWEVER
theres a big unknown quantity in all this:
how kiked putin is, exactly
because a jihadist uprising would disorganize europe enough to create an opening for russia to attack.
i dont know whats the status on russias capabilities either because all available information from either side is guaranteed to be propaganda

what makes me suspicious of him is that i think the war in ukraine has no right to develop as it is
rather that key decisions masked as incompetence, or rather magnanimity in this case were taken, which unduly prolonged the conflict in ukraine

more precisely the fumbling of the initial blitz and pivoting from the exact tactics and type of combat the russian army was designed, trained and equipped to conduct
into fucking ww1 style trench warfare
this is insanely retarded from a homefront point of view
but it makes sense in the interests of globalists because ukraine war provided pretext for the increase of energy prices in europe, or more precisely, the margins
https://ember-energy.org/data/european-electricity-prices-and-costs/
funny how energy prices hover around pre war levels but people pay up t o x4 the original price
and given that energy accounts for ~10% or the gdp of a western country
https://www.researchgate.net/figure/Energy-intensity-of-GDP-Kilograms-of-oil-equivalent-KGOE-per-thousand-euro-in-EU_fig3_367243490
(1kgoe is ~ 11KW/hr)

were talking several percent of the gdp of all europe that got siphoned away

thats several times the amount of money spent on the war by both sides combined
and who cares anyways because the costs of war are entirely covered by the taxpayers

(1/2)

>>23062983
(2/2)
and so
what i fear might be an option is that the whole dichotomy of west vs east is bullshit
putin is kiked up his ass
and hes gonna conquer europe when this one will be dealing with a jihadist uprising

its not the first time the west and the east are cooperating in order to achieve global domination
but i digress
the point is, the system doesnt matter
what matters is to control the people who control said system.
eu, neo soviet union, doesnt matter. the only things that change is the attitude to tranies and the official language spoken kind of deal

>>23062986
IF/ELSE models require a human to anticipate every possible situation and write rules for it.
Modern AI (machine learning) instead learns from data. It generalizes from examples rather than relying on manually encoded logic.
E.G: IF/ELSE: If "cat" in text, then label as animal.
AI: Sees thousands of labeled examples of cats and dogs, and learns statistical patterns to distinguish them, even in cases it’s never seen before.
Modern AI models, especially deep learning models are universal function approximators. They use multi-layer neural networks to represent extremely complex, nonlinear mappings from input to output. These networks contain millions or billions of parameters (weights) that are not hardcoded, but learned through optimization (gradient descent, backpropagation).
A neural network doesn’t “decide” things with if-statements. It computes outputs by activating neurons through weighted sums and nonlinear functions like ReLU, sigmoid, etc.
At scale, large models like GPT-4 or Claude or Gemini show emergent properties:
>Abstract reasoning
>Code generation
>Multi-language translation
>Humor, sarcasm, analogy
These properties aren’t programmed manually, they emerge from the data and model architecture. That’s a huge leap beyond anything an if-else tree could achieve.
Language models like GPT-4 use transformer architecture, which can:
>Attend to every word in a sentence to understand context.
>Represent meanings using dense high-dimensional vectors (embeddings).
>track long-term dependencies over pages of text.
There is no set of rules or flowchart that could handle all the nuance in real language like this. Modern AI doesn’t “choose between options” via conditions.
It optimizes a loss function across millions of examples to minimize prediction error. The result is a model that performs probabilistic reasoning, not binary logic.

>>23062987
>>23062988
I’m a fan of Putin, but he is very kiked, unfortunately.

I’d love to see a multipolar world with East-West detente, but MIC’s gonna MIC.

>>23062968
>Does this function look differentiable to you, tard?
uh, it is, but only for x > 0. Anyways, AI is just a huge scam, for those who knows how LLM works, LLMs are without a doubt the most niggerlicious and inefficient piece of solution for a problem that never existed.

>>23062986
Comparison:
> Recognize a handwritten digit, 5
Else if approach:
if number_of_loops == 1 and middle_bar_present:
return 5
else:
return unknown
>Works for one handwriting style
>fails immediately on variation
Neural Network approach:
>trained on tens of thousands of examples
>learns high-dimensional features automatically
>can generalize to many styles it hasn't seen

>>23062934 (OP)
I dunno, AI has been helping me with creative writing and understanding theoretical singularity physics.

>>23062993
you now understand the idea, but youre not really capable of testing if your understanding is fundamentaly right.

>>23062991
>uh, it is
>except for where it isn't
Anon... come on.

>>23062993
>creative writing.
>AI slop.
Lol, lmao even.

>>23062990
yeah thats what i hear too
but the whole jihadist scenario is not a given
quite a few things have to unfold in a specific way for it to become a problem or even take place
if europe offers resistance to a russian conquest the american wasp lobby wont be able to resist increasing its influence over the continent, enacting lend lease, then marshall plan and have worldwide primacy until post scarcity, pretty much
and i dont think either country has the capabilities to conduct such a war in their current state

but again, putin might have a line of underground factories ready to churn out robotized t90s and ai driven drones and everything is pre arranged to play out as a remake of ww2
ruin europe, then enslave and exploit through debt from recovery efforts
i think its unlikely but its still possible

one thing is certain
the eu will either see heavy reforms
or will just plainly go down
which will greatly weaken the influence of globalists
and given the rapes, beheading etc, the people aint gonna forget to hate globalists anytime soon

>>23062996
>>uh, it is
>>except for where it isn't
>Anon... come on.
i keked.

>>23062989
>>23062992
Every layer of a FNN can be easily modeled by a piecewise-defined function. Every piecewise-defined function easily translates into conditionals for selecting the appropriate piece. It's literally just a decision tree for selecting some local approximation at each level of abstraction.

>>23063000
FFN*

>>23062985
Replacing the if-then tree with some image of least squares method would be far more accurate example of what current "AI" is.

>>23062993
>AI has been helping me with fake pop-science and writing shill posts

>>23063002
See >>23063000. Decision trees are an accurate representation.

>>23063003
Let me guess: the earth is flat, right?

>>23062997
>made up his mind without seeing it

>>23062995
Yes. That's what theoretical means.

>>23063003
>everyone outsources their creativity to AI
>AI trains on its own creations
The Slop Singularity

>>23063005
>>23063006
>>23063007
I can see why you use AI slop for your "creative writing". Barely sentient...

>>23063006
>made up his mind without seeing it
You don't get it, calling whatever comes out a LLM slop machine creativity is calling a plagiarized work "original". Anyways, dont let the bitter me deter you from enjoying what you like.

>>23063009
>uh ur dum
Ok.

>>23063008
That would morbidly funny, but people who are creative don't outsource creativity to autocompleter programs.

>>23063010
You misunderstand, I'm having it check my writing.

>>23062968
What is the if-then-else equivalent of a sigmoid function? Seeeeethe!

>>23063004
Yes, but I like least squares better as an example because it kind of demonstrates how the training works.

>>23063014
>What is the if-then-else equivalent of a sigmoid function?
A truncated version of the same? A spline? Anything you want. Who even uses sigmoid activation these days? Deep learning only took off when people figured out to stop using them.

>>23063000
Saying a feedforward neural network (FNN) is “just a decision tree selecting local approximations” drastically oversimplifies both the structure and behavior of modern AI systems. Yes, mathematically, you can describe neural networks as compositions of piecewise-linear functions, especially when using ReLU activations. And yes, you can, in theory, write conditionals to emulate those outputs for specific inputs. But this framing ignores the scale, learning dynamics, and generalization that make neural networks fundamentally different from manually constructed decision trees
Decision trees rely on explicitly defined, human-interpretable branching logic, if X, then do Y, often hardcoded or grown through relatively simple heuristics. In contrast, neural networks learn dense, high-dimensional representations from data, with millions or even billions of parameters distributed across layers. They don’t follow if/else logic per se; they perform continuous, differentiable computations that evolve during training through gradient descent, allowing them to interpolate and generalize across vast input spaces. You don't hand-design the logic, it emerges through optimization
Moreover, the “local approximation” idea doesn’t capture the expressive power that comes from stacking layers, non-linearities, and techniques like attention, recurrence, or residuals. Deep neural networks can model complex patterns, reason across contexts, and encode abstract features far beyond what decision trees can feasibly handle. Even the most sophisticated boosted trees can't match transformer-based models on tasks like translation, image generation, or gaming
So while the underlying math might boil down to compositions of simple operations, dismissing AI as “just conditionals” is like reducing a novel to the ASCII table. It's technically true in the narrowest sense, but it misses the complexity, capability, and novelty of what these systems actually do

>>23062962
>networks are expressed as a series of function calls, purely as a matter of opinionated convention
>this means it's just a decision tree bro
????

1. Symbolic AI:
Uses explicit rules.
Transparent and explainable.
Good for logic and language-based problems.

2. Neural networks:
Learn statistical patterns from data.
Black-box nature.
Scalable but hard to interpret.

Decision trees can model simple rule-based systems, but:
They can't efficiently represent the nonlinear, distributed representations of complex neural models.
A large language model like GPT is not practically representable by a decision tree: the tree size and complexity would be absurd.
The generalization ability that deep networks provide doesn't emerge from decision trees.

In summary: The difference between symbolic and neural AI is very significant — theoretically, practically, and in terms of implementation.

>>23063017
>i'm a know-nuffin nigger
>i use an AI slop generator to think for me

>>23063019
>i'm a know-nuffin nigger
>i don't understand simple technical explanations
>i use an AI slop generator to think for me

>>23063018
It's equivalent to a decision tree. Every limitation of a decision tree is present in FFNs.

>>23063020
Not an argument. I accept your concession

>>23063023
The argument is made here: >>23063000
It stands completely undisputed. Generating AI slop is not a refutation.

>>23063018
It can be reduced to how a decision tree works. In theory, its whole behavior could be redefined as a series of if else chains where the condition compares the target value to a dynamic float variable that is adjusted based on the resulting quadratic error after each epoch.

>>23063020
>>23063021
Sophism and ad hominems are not arguments. You lost. Stop embarassing yourself.

>>23063023
nta (obv)
but you could at least read what the chatbot spits back at you

>>23062934 (OP)
“AI” is just kikes stealing and selling other people’s knowledge.

>>23063026
The argument is made here: >>23063000
It stands completely undisputed. Generating AI slop is not a refutation.

>>23063024
Already explicitly debunked in detail. All you know how to do is repeat your retarded takes like the sub-80 IQ NPC nigger you are. The only reason you seethe so much about AI is because you're envious that it's already much smarter than you will ever be. But go ahead, keep repeating your NPC lines to make yourself an even bigger fool than you already are

the devil is coming
https://m.youtube.com/watch?v=OHQRo3Uz_VQ

Why does glowies hate us so much?
I mean like whyd the come to our hang out and fuck it up.
I always said my handler was hot and breedable like thats being very nice imo (not that its not true, my handler moggs). Whyd they do me like this?

>>23063030
>i "debunk" arguments i don't understand by making my chatbot generate some slop that conforms to my dumb opinions

>>23063027
Yes I read that, It's a gross oversimplification, as expected of sub 80-IQ niggers unable to understand nuance beyond anything of a surface level complexity. Being able to understand nuance in fact is one of the most showing signs of intelligence, which is something OP is clearly unable to do

>>23063030
>ai
>smart
ai cultist = opinion automatically retarded

>>23063027
And you could read the whole thing. Neural networks can do a lot of things what classical programs cant. You try to ignore the concept and functionality of a universal function approximator and you try to reduce it to technicalities.

>>23063033
>still zero responses to any of the arguments made
Your concession was already accepted, you can stop talking now

>>23063034
no, as expected of someone who is used to think in high level abstractions
you extract properties inherent to a concept
you observe them, simplify, refine, find analogues
being unable to do that is a sign of autism

>>23063036
>Neural networks can do a lot of things what classical programs cant.
yeah but only because of practicality
for all intents and purposes you could take a model and auto generate the equivalent conditional code
its a perfectly useless exercice, but you """"easily"""" could do that

>>23062967
shill

>>23063035
Actually, if you had basic reading comprehension, you would have been able to see that I did not call AI smart, I called AI smarter than OP, which is not a very high bar to cross considering what a retarded nigger he is

>>23063041
yeah nah i think you smell of cargo cult
and autism
>speaks of reading comprehension
kek

>>23063042

>>23063037
You are mentally ill and seething. Call me back when you can name something a FFN can do, that its decision tree equivalent can't.

Bonus points if you post something you understand instead of AI slop.

>>23063025
>in theory
It's not theory, that's literally how it works. It's not a decision tree simply because it has discrete conditions. Decision trees have global context, with human interpretable decision boundaries. Almost by definition, NN behaviour is emergent and not directly interpretable.

>>23063022
they have the same limitations because they share the same underlying hardware based on turing machines.

>>23063045
>they have the same limitations because they share the same underlying hardware based on turing machines.
Completely wrong. Both decision trees and FFNs have the same limitations compared to the general hardware they are running at, which is far more capable than either.

>>23063043
exactly
how the f am i supposed to understand that
i dont speak finnish

>>23063036
>>23063039
>for all intents and purposes you could take a model and auto generate the equivalent conditional code
and thats kinda what happens when models get tested and deployed

>>23062934 (OP)
Is the human brain also just not a series of if/else statements?

>>23063039
Then try to teach your symbolic decision tree to give better results!

>>23063045
>Almost by definition, NN behaviour is emergent and not directly interpretable.
Unlike know-nothing """AI""" cultists who worship magic black cubes, actual researchers convert these things into more useful representations like decision trees to make them interpretable.

>>23063048
yes.
but what you said doesnt explain where its os comes from or how its programmed exactly
actrual applied psychology relies on the fact humans are, in fact, a series of if/else statements

>>23063039
>its a perfectly useless exercice,
Actually generating the entire decision tree would be useless exercise, but studying FFNs in terms of decision trees is one of the techniques researchers use to make sense of them and establish their capabilities.

>>23063048
>Is the human brain also just not a series of if/else statements?
Obviously not.

>>23063044
Already did, and it's been done by multiple anons in this thread, all of whose posts you've simply ignored and went back to your original point because you're too fragile to admit when you've been utterly humiliated. I'm laughing at you. You make this way too easy for me.

>>23063054
>Already did
Then name a specific thing a FFN can do, while its decision tree equivalent can't, in your next post.

Protip: the next instruction in your program is deflection. You will not name anything.

>>23063049
a) theres nuance to things
a model is like a sculpture
training it is the process of scumpting
is the culture the process of scumpting or just its result?

and b)
doesnt matter anyways because you can have variables inside your conditionals which you update as the program runs
you can absolutely simulate "learning", or rather implement "self calibration" using conditionals
(this is why simplification and conceptual refinement are important. here when the concept got simplified down to its essence, its properties got way more apparent)

>>23063051
>>23063053
AI has more human reasoning than the average retard who thinks in pure binary logic anyway
You ask an AI what it thinks about something it will usually address the topic with more nuance and empathy than the average polarized directionbrained nigger

>>23063057
Cope all you want, but you said something completely retarded. Maybe your mind is a fixed decision tree but in reality, such an architecture isn't even powerful enough to model a single neuron.

>>23063052
>but studying FFNs in terms of decision trees is one of the techniques researchers use to make sense of them and establish their capabilities.
huh.
didnt know that
i didnt give that aspect much attention, im of the opinion the future is not a big model but meshing symbolic and ml elements

>>23063050
human interpretability research is an abstraction. If decision trees are used, it's only as a tool for human understanding.

Please show me where the decision tree is in this paper, I will wait:
https://arxiv.org/html/2502.00873v1

>>23063057
an ai doesnt think it correlates things together
it doesnt have the capacity to reason, only regurgitate correlations it found in its dataset

its useful to pingpong your ideas with it but fundamentally youre just working with a google search engine with a cutting edge interface

>>23063054
You're arguing with a teenager with no dad. He doesn't know what ML or AI even is. He just wants to argue with people and repeats the same set of canned responses like a literal NPC.

His dad is German and his mother is Moldovan. He makes the same thread over and over because he just wants to argue to let off steam. School is starting soon after all.

>>23063059
Decision trees don't inherently imply symbolic AI, anon. But yeah, there's neuro-symbolic AI research trying to fix the LLM's terminal unsuitability for reasoning.

>>23063060
There's no amount of mental gymnastics you can do to disprove the simple truth that a FFN is no more capable than a decision tree, and one can be easily mapped into the other.

>>23063060
>https://arxiv.org/html/2502.00873v1
Are they trying to emulate an analogical data type by using an integer data type variable in a binary computer? That has to be the biggest piece of slop i have ever seen. But yeah it is a known fact that, you really need an analogical computer to break through the current restrictions of """AI""", that's why top researchers are working on making a modern analogical computer architecture instead of publishing LLM agentic slop papers.

>>23063063
>Decision trees don't inherently imply symbolic AI, anon.
right. its still something im learning
anyways, theres always a certain degree of incertitude with ml, since everything is stats based and you ALWAYS have outliers in real world data when training or when executing
incertitude which doesnt exist in a symbolic ai
thats why i think the natural path forward is leave statistical systems to statistical engines
and rigid logic to symbolic ones

like for programming ai-
and how model based design would mesh into that
instead of tokenizing the whole language, output just a set of easily verifiable, testable routines constrained within a self validating system
or how the czinger ai is built
https://www.youtube.com/watch?v=VdG4gUTowXc

its full symbolic but it uses evolutionary algorythms to do the proper design b ased on a set of constraints and hard coded laws of physics
chain such ais toegether for various technological processes and you end up with a warhammer-esque STL where you input sets of constraints and the ai spits out a whole technological process
thats engineering possibly fully automated with current technology

>>23063064
if you actually tried what you're saying just as a fun little exercise, you would quickly realize that the expansion to DT leafs would result in O(W^L) (width/depth of net) growth. This is completely intractable from a computing perspective. Impossible to compute for a reasonably sized net on the same hardware. The fact that nets can express this dimensionality refutes your assertion of equivalence in a fundamental way

>>23063065
goalposts thrown out of the solar system

>>23063067
Notice how nothing in your post approaches disproving the simple truth that a FFN is no more capable than a decision tree, and one can be easily mapped into the other. Googling papers you don't understand isn't helping your case.

>>23063069
The essence of the post was explaining how nets are more capable
You have reading comprehension issues, or you don't know how big-O works.

>>23063066
Remember way back at the beginning of the thread where you were defending your definition of "AI" by appealing to context? How would a symbolic system handle this kind of issue?

>>23063067
He is just a dumb stubborn kid conflating everything through deconstructions and reconstructions. Like a smartphone is functionally a rock if they have the same types and amounts of atoms.

>>23063068
Ok dumb nigger, i will bite the bait.

>>23063071
youre still gonna have memory management, piping and other boilerplate in the program, wont you?

>>23063070
>The essence of the post was explaining how nets are more capable
There is literally nothing in your post implying that neural nets are more capable. You simply regurgitated some statement effectively stating the opposite of what you believe (clearly, you did not understand any of what you copied into the textbox) and then asserted your wrong opinion by way of a nonsequitur.

But since you're kvetching about "computational tractability" so much (even though it's irrelevant), consider the fact that you can evaluate it lazily and then it's perfectly tractable.

>>23063071
>>23063074
and at a higher level, the computing is gonna be done by a component of the larger machine
which then is used at a certain point in a technological process
it gives lexical flexibility such as "the ai takes care of that"
in a context of, idk, farming lets say
and drones that cut grass to make feed or something
or "the ai takes care of x" in the context where an ai based system integrated into a wider infrastructure
the size of multiple buildings and adjascent terrain

until a few things happen, Ai and its use will be restricted
the first, proper cataloguing and error correction in python
until its clean enough to have a proper library and version control its pointless, you still have to check everything(or most) for software dev

>>23063073
>look they said "if" once. suck on that, chud!

>>23063075
you just solved p=np simply by lazily evaluating np
congratulations
go collect your million dollars and fields medal

>>23063078
That paragraph is explaining how are they using a binary election scenario to model the attention heads output you dumb fuck.

>>23063079
You're a mentally ill retard and you literally don't understand anything you're reading - neither my posts, nor the papers you hastily copy statements from. Explain what p=np has to do with this.

Protip: you 100% won't. Let's see what flavor of deflection your post will contain.

>>23063080
>>23063078

>>23062963
All memeflags are kikes.
Refute or STFU

>>23063072
A better analogy might be vector vs rasterized images

>>23063081
please, explain how a lazy evaluator solves the O(W^L) expansion from net to DT

>>23063074
>>23063076
I don't get what you're talking about. What I was saying is that in a real-life situation your symbols would have to be dynamic to account for context.

>>23063066
Neural networks, even artificial ones, don't work like a decision tree, OP misunderstands. Also symbolic AI was always a meme, it was a game of hide the ball where people were hiding complexity in symbols and then thought building an intelligent system would be then easy because then you're just manipulating those symbols.

The whole premise of OP and the reason for his spamming the board is he thinks he has insight that ANNs are "equivalent" to decision trees. His central claim:

>There every layer is equivalent to a piecewise-linear function
It's wrong.

If-else = runtime branching, discrete jumps, non-differentiable.
NN layers = fixed algebraic graph, continuous transforms, differentiable (almost) everywhere.

The differentiable part is *essential* to train a useful ANN in the first place.

>>23063085
By not needing to instantiate irrelevant branches? Man, either you're retarded or you just have no idea what you're talking about here whatsoever.

>>23063088
right
your fields medal awaits

>>23063089
You're basically a biological LLM stringing meaningless tokens based on vague associations. No understanding. You can't even perceive that your slop makes no sense.

I wish this was true, but i'm using ai in my job and it will soon replace me and other financial guys

>>23063091
If you can be replaced by a static decision tree, you should be replaced by a static decision tree. :^)

>>23063086
bah the only difference between a symbolic ai and mlai is that symbolic ai is written by hand, in practice
theres no reason not to have mechanisms similar in function to properties of a neural net or even straight up hand programming one

>>23063087
no, what op is saying that you can reinterpret neurons as nodes in a decision tree
nobody said that a decision tree must be binary
think: flowchart

>>23062934 (OP)
means that chatgpt 5 is coming soon. i am not sure if jannies are aware or not but this site is heavily being used for advertisement. chatgpt doesn't have to be mentioned by name. everyone connects chatgpt with ai these days because they have been the loudest with the most ludicrous claims and the most funding involved.

>>23063086
>>23063093
also for some reason i deleted the first half of my reply.
prolly bc im baked

>aaah, i thought you meant "what about the juxtaposition of ai to non ai in the context of a program thats all ai"

>>23063086
>even straight up hand programming one
in fact
every program can be reinterpreted as a neural net
with the functions being individual neurons

checkmate atheists

>>23063090
no u

>>23063097
I'll try to break it down for you in a way even a 90 IQ should be able to grasp.

Do you understand that any given layer of a typical FFN can represented by a piecewise-defined function? Nevermind a deep network. Try to wrap your head around one layer.

>>23062934 (OP)
The funny thing is that OP Image is not wrong and yet not disprove AI

>>23063099
Do you think there is a fixed decision procedure for intelligence?

>>23063098
>>23063067

>>23063101
Notice how you are forced to deflect back to some irrelevant snippet you probably copied from your LLM, because you didn't grasp the question I just asked.

Do you understand that any given layer of a typical FFN can represented by a piecewise-defined function? Nevermind a deep network. Try to wrap your head around one layer.

You will deflect again.

>>23063101
You could do it layer by layer though. It wouldn’t be the whole tree, just the path it followed. Good for debugging

>>23063093
Yeah but there is no insight here, it's utterly useless. I can say that NNs are just matrices being multiplied or go down to transistors. But how did you get that particular arrangement at these levels? Did it just come to you in a dream? We could respresent ANNs with even gates of water in the real world. It says nothing.

What you and him fail to understand is that for the longest time (decades). People used to hand craft these discrete rules, which did contain if then else and it also involved a lot of preprocessing in loops. It did things like try to recognise fruit that was misshapen or do translation from one language to another. You can represent that as if statements right?

The problem is that in the real world there are these edge cases. If you didn't program for them, then the system fails, often spectacularly. That's why in places, like the factory that would be looking at fruit for the camera they had specific artificial lighting, the machines would do a bunch of physical sorting and so on before it got to the camera so a bunch of assumptions could be made. Even then there are always still some edge cases because of the variation in nature. In the case of machine translation, it went on for decades, people adding new words or covering edge cases someone found with the lookup.

The technology we have today, using ANNs these are solved problems. Saying they can be represented another way isn't insightful, you're just taking the extra step and converting it to another form. Where are the geniuses using piecewise functions from scratch? Why haven't we heard of them? Why are we wasting our time with matrices then?

>>23063103
You could define the entire tree lazily. The trick is that the FFN is actually less general than a general decision tree.

Yup. I'm always meh about AI. Moreso a pager than a big ideological shift, though it's priced as one. Having used it extensively, it's a fine support tool, great replacement for therapists, but hardly a catch-all for what it's propped up to be. Man's gotta work after all.

>>23063104
I think OP’s claim is just that it has the same limitations as trees since it can be represented as a tree.

>>23063107
>I think OP’s claim is just that it has the same limitations as trees since it can be represented as a tree.
Yep. This is exactly my claim and I made it explicit multiple times.

>>23062945
Trannies and Furries were litterally treated the same or worse 20 years ago. AI Companionship is going to go very mainstream very quickly because we have normalized being publicly mentally ill.

>>23062934 (OP)
doesnt that diagram apply to the brain as well?

>>23063102
>>23063067

>>23063103
in a dense net, there is no "path it followed", the whole net is evaluated. A DT expansion has to account for that.
If you're talking about some other kind of net with partial evaluation, ok, it still has the same growth profile times some coefficient representing the amount of the net that gets evaluated.

>>23063104
>Yeah but there is no insight here, it's utterly useless.
...
>The technology we have today, using ANNs these are solved problems.
ah so thats your point, shilling

fyi the disussion is about high level objects and their properties
this kind of considerations intevenes in design at several levels
from ai architecture to even feasibility studies

>>23063111
Do you understand that any given layer of a typical FFN can represented by a piecewise-defined function? Nevermind a deep network. Try to wrap your head around one layer.

Notice how you will continue to shit out replies trying to deflect this question. You fucked up and now you're in a loop trying to stall. In your damaged mind, you think this helps you save face.

>>23063113
Hmmm, scanning this thread, you seem to know some things AND you're arguing againsy shills smugly. My favorite. Now I dont have a lot of technical knowledge, more semantic knowledge. Is your issue is that most LLMs or whatever are binary and need to be mediated through a platform, like chatgpt, and are limited by allocation? All this multiplicative decision crap just seems like poor optimization, like better token weights could accomplish the same thing, but I honestly dont know enough to comment. I'm trying to learn more is what I'm saying and I like the cut of your jib.

>>23063113
When you said a lazy evaluator solves exponential growth I realized I'm simply no match for you. I kneel.
But also, consider >>23063067

>>23063115
Do you understand that any given layer of a typical FFN can represented by a piecewise-defined function? Nevermind a deep network. Try to wrap your head around one layer.

Notice how you will continue to shit out replies trying to deflect this question. You fucked up and now you're in a loop trying to stall. In your damaged mind, you think this helps you save face.

Number of times you attempted to deflect this question: 5.

>>23063112
Pointing out how little you know about the older systems and their limitations is shilling? What am I shilling exactly?

>fyi the disussion is about high level objects and their properties
Any retard can go up a level on work that other people did and claim deep insight. Like I said, ANNs can be represented as gates of water. We could take the existing fixed graph of an ANN and map it to that. So what? If you don't agree with this you are a shill. That's you right now.

What he is doing is playing the same game that the Symbolic AI crowd did, where it is just hide the ball. "symbols" are a wishy washy term that a human will understand but you need to map them to a computer. If you could code you'd understand that doing this in computer code isn't trivial when we're talking about real world systems.

>>23063107
>I think OP’s claim is just that it has the same limitations as trees
It doesn't though. Think of a decision tree like a vending machine. There is a flow chart you can make to map the various inputs and outputs. You can map the whole process, adding money, handling refunds and deciding to dispense the various drinks. If you want a new drink, you need update the tree. The drinks in the tree are just represented with numbers. The decision tree doesn't have any concept of what a soda is or anything.

With ANNs, when they get big enough and trained on a wide range of data (think the whole internet). That graph, even though it's fixed, is more generalizable. Is it at the stage where it's as generalizable as humans? No. But you could absolutely hook up something like an LLM to a vending machine, give it a prompt and let it manage it. The point would be: it's never been trained to do that. It could handle new drinks being added without it having to have its weights updated. Nor would it have been trained to do other things like help the human decide which to pick. That's what's interesting about the technology.

>>23063117
>Pointing out how little you know about the older systems and their limitations is shilling? What am I shilling exactly?
anns as being an universal solution.
>in a discussion about abstractions
>What he is doing is playing the same game that the Symbolic AI crowd did, where it is just hide the ball.
strawman ASSumptions
+ like i said
you smell of cult
and autism

this is fucking dogmatic. and thus, retarded
theres absolutely no reason NOT to output labels and feed em into a symbolic ai, accounting for edge cases this way

>>23063114
I'm not sure what you mean, but my contention here is very simple: AI fans try to mystify neural by treating them like black boxes that do complicated non-linear transformations in high-dimensional spaces. Going just by that - if you don't know anything about the specifics - it's unclear what such a thing can or cannot do. What I'm saying is this: no matter how many layers you stack, no matter what fancy activation functions you choose, there is some static tree of nested conditionals that will produce the same output. And this argument doesn't even involve a plunge into some ridiculously low level of abstraction: it's roughly the same level of abstraction as those ANN graphics you're familiar with.

Now, unless you think general problem-solving boils down to a fixed number of stages, each one involving mechanistically choosing between a fixed number of options, then a LLM is not going to cut it for "intelligence". Two More Weeks/Two More Layers/Two More Data Centers cope doesn't fix this.

>>23063118
>anns as being an universal solution.
And the OP being a non-solution, he's just converting someone else's solution into a retarded representation that doesn't even make any sense. It only makes sense to you because you assume computer code can handle high level things, like "symbols" in an If. Crying "strawman" in defeat won't help you here. Learning to code will.

>you smell of cult
>and autism
Good to know you're triggered and have to resort to insults. All you can do is sperg out like an NPC. Let us know if you have any original thoughts to share or if you'd like to keep getting stuck in a loop.

>>23063120
>digression
>sad homonems
i accept your concession, trany
youre just seething at the fact people demean your cargo cult
>nocoder
lamao

>>23063117
>it doesn’t though
Yeah it does. The model has the novel abstractions in it already, statically represented in the statistics. The “creativity pre-work” is done at learn time, and that process doesn’t use if-thens

I agree with you that it’s interesting, especially when the training data is huge. But I also agree with OP that it’s ultimately limited by the same things a decision tree would be limited by. It *doesn’t* work like a decision tree, but it can be unfolded into one. Theoretically. So it has the same limitations.

>>23063121
>arguing with a brown memeflag

Time to issue a threadly reminder:

Nearly all """AI""" enthusiasts are genetic dead ends who, up until 5 years ago when the """AI""" psyop got rolled out in full force, were planning to rope. You are trying to rob this person's only hope for the future. You are trying to take away from him his AI god who promises eternal life in virtual heaven. The AI god who will genetically modify him into a chad, make him immortal, give him full-dive VR into fantasy kingdom, give him a harem of AI-driven 12 years olds, give him the 60 IQ points of intellect that he lacks, etc.

These people can't be reasoned with because reason wipes out their happy future and lands them back in the suicide-fuel reality of their lives.

>>23063123
>These people can't be reasoned with because reason wipes out their happy future and lands them back in the suicide-fuel reality of their lives.
we had measurable success on g in stemming the tide of retards
but only temporarily
because the shills always come back with the determination of an autist and its repetitive behaviours
i dont mind the intellectual skirmish. its very self congratulatory for me, its a guilty pleasure

>>23062934 (OP)
Some people own nvidia stock. That will make you lie to keep the stock up.

>>23062934 (OP)
Yeah, it does comes in waves.

Anyway, here's the shield hero Naofumi.

>>23063124
>the shills always come back with the determination of an autist and its repetitive behaviours
Good thing I have the perseverance of a thousand spergs combined. No sense pouring your heart into this. Just override their repetitive behaviors with more heavily optimized, more obonxious repetitive behaviors and watch them seethe/abandon thread. :^)

>>23063114
>All this multiplicative decision crap just seems like poor optimization
If you think that is poor optimization, just wait until you see how LLMs manage context memory.

>>23063125
its not only a matter of nvidia
its a matter of bubbles and pyramids

>>23062943
>>23062942
Ai isn't about making machines more human, its about making humans into controllable machines.
They literally talk about how ai will "cognitively diminish" humans

holy shit, we have two autists now?
the gypsy got a VPN, didn't he?
lmao

>>23063130
>Ai isn't about making machines more human, its about making humans into controllable machines.
I agree. The convergence between Man and Machine will happen not through the ascent of the machine, but through the descent of Man.

>>23063127
its like training
you give a statistically signiifcant negative stimuli
and the model adjusts kek
>>23063130
They literally talk about how ai will "cognitively diminish" humans
didnt know that
but it also doesnt surprize me either
case in point
>>23063131

>case in point
*speaking of the wolf

>>23063119
Thanks for the reply. I did some research, and it seems like you're right. Based on what I learned, you can break it down mechanistically, but that involves triage and logic gates with partitioned variable models down stream that only activate under specific conditions. I always thought it was just ifelse statements stacked in increasing conplexity, but it seems like they're going for dynamic problem solving using esoteric bullshit. My issue with the whole LLM thing is that the backpropagation is exponential the more complex the decision tree, and the greater the mapped inputs from the training data. But instead lf having them mirror human decision making, which is ontologically nonsense since they have diffused liability, why try? Just make the model produce consistent outputs with a massive amount of tagging, and minimal greater order priorities. But then again the people for these things see them as replacing humans altogether, which is hilariously idealistic, but it's their value prop, so they have to be invested in it.

>>23063133
>you give a statistically signiifcant negative stimuli
This is basically my thinking at this point. Make the interaction maximally unpleasant for them and they will eventually develop some avoidance behavior. Simple operant conditioning. :^)

>>23062934 (OP)
AI is literally Pajeets who type fast in a call center.
https://itwire.com/it-industry-news/deals/builder-ai-enters-bankruptcy-when-its-ai-exposed-as-actually-indians-more-than-700-human-workforce-faking-ai.html

>AI isn't actually intelligent
no shit, this isn't news to anyone but normies who don't know what actual AI means

>>23063138
>what actual AI means
>Actually Indians

>>23063137
>>23063139
ah yeah the same story lamao
a parody of itself

>>23063121
>VScode
>Windowws
>.c
You know that explains a lot. This is exactly the type of person I expect to get triggered by matrix multiplication doing interesting things. You could've just shared this earlier to let me know you're a moron. Your flex is using a normie IDE, working on a dying platform. No real world experience. As in, mapping that C code to things in the real world, not the shitty dll wrapper you have there. Try actual engineering.

>>23063123
I know you read every one of my posts. Still unrefuted and all you can do is pretend not to see me. It's total victory at this point.

>>23063124
>we
Ah so the earlier accusations of being a "shill" when your lack of knowledge was pointed out was just projection. You might as well just share the discord link at this point.\

>traction on /g/
If you guys want to call bumping your own thread and agreeing with each other traction, sure. Cope with that.

>>23063122
Yeah you can hide in statistics if you want, but this is no different than "symbols" right. Like the thing is a human understands at a high level, like what an apple is. Sure you can write something like if (apple). That's easy to map for a vending machine where it's button 2. But in the real world it's not as easy as that. You can use statistics to model any kind of system, including ones with a lot of uncertainty but the OP is fixated on his if statements. Even if he were to claim something like P of X. What is X? Is it an integer or a something higher level like Apple.

this nigger is talking to himself, isn't he?
lmao

>>23063141
wow. i didnt know i was such an interesting topic
im very flattered
but what about any technical argumentation whatsoever?

>>23063141
>shitting out a thousand paragraphs
>expecting me to actually read them
Smelly memeflag.

>>23063142
yeah totally
did you dilate today?
you sound angsty

this is absolutely is going to end with the gypsy making several LLMs to samefag his own threads
literally the craziest autist on 4chan, lmao

>>23063146
>no you
that means you care
that means youre mentally underdeveloped
>ai fag
ofc

>>23063146
>this is absolutely is going to end with the gypsy making several LLMs to samefag his own threads
Once I have enough material in the archive maybe I'll fine-tine one and have it spam all of your """AI""" enthusiast threads non-stop.

>>23063145
>>23063147
>>23063148
dude, you're seriously insane
lmao
this is amazing

>>23063149
>everyone who disagrees with me is the same madman

>>23063149
>t. calls you mentally ill
mm-hmmm

absolutely amazing, kek
>>23063151
what's a CGF?
cyborg gf?

>>23063152
somthing to that effect
this is more representative of the type of derangement

>>23063141
Since you can unfold an FNN into a decision tree, they share the same representational capacity over finite domains. That doesn’t mean they generalize the same way or work the same way, just that the upper bound on what they can represent is equivalent.

Not saying I would make a meme about it like OP, just trying to help everyone get in the same page

>>23063154
ai fags are mental retards
they shouldnt be on the same page
they should be in a mental sylum

>>23063154
>Not saying I would make a meme about it like OP, just trying to help everyone get in the same page
I didn't make that meme. It's been around long before the current incarnation of """AI""". I picked it specifically because it triggers """AI""" fans so badly while also correctly implying that no fundamental limitations have been transcended compared to the previous iterations of this psy-op.

*bants*

>moved to bant
lmao
this nigger is going to blow a gasket...