Incel Invented AI

[Knajjd]

Photo of Hinton, Sutskever and Krizhevsky.

In 2012 Krizhevsky purchased two Nvidia GeForce GTX 580s. He combined the CUDA tech on the GPU's with neural network tech to develop the first artificial neural network.

Prior to this both Nvidia's CUDA and neural networks were considered failures. It was Krizhevsky who combined the speed of CUDA and the learning ability of neural nets to produce modern AI.

Of interest is that Nvidia added the cost of CUDA tech to the price of their GPUs so that in effect gamers funded the tech.

Also, please note that transformer tech also achieved a breakthrough by combining CUDA tech and neural nets, using Krizhevsky's original idea of combining both techs.

A Japanese documentary team made a film about AI and soon realised from interviews that a key person was missing. They tracked Krizhevsky down to a barely furnished appartment over a Vietnamese restaurant in San Franscisco. He no longer works in AI and is cooking up something else on his laptop.



Wiki link to Krizhevsky:

Alex Krizhevsky - Wikipedia

en.wikipedia.org

Wiki link to AlexNet:

AlexNet - Wikipedia

en.wikipedia.org

Wiki link to CUDA tech:

CUDA - Wikipedia

en.wikipedia.org

 

[F̶e̶d̶-̶L̶i̶n̶k̶]

Sad, atleast the Ethereum guy got some fame

 

[SlayerSlayer]

you know you're a loser when you provide immense value to the world and no one cares while they all profit like robber barons off the research you did in the first place. He may not be the ugliest guy, but man, to come up with this idea and you're still a loner living in apartments you should just shoot yourself

 

[Friezacel]

you know you're a loser when you provide immense value to the world and no one cares while they all profit like robber barons off the research you did in the first place. He may not be the ugliest guy, but man, to come up with this idea and you're still a loner living in apartments you should just shoot yourself

Being rich and nobody giving a shit about you doesnt sound too bad

 

[SlayerSlayer]

Being rich and nobody giving a shit about you doesnt sound too bad

hes not rich though, did you see the shithole apartment he lives in?

 

[Friezacel]

hes not rich though, did you see the shithole apartment he lives in?

Brutal. Cant find any informations on how much money he has

 

[GeckoBus]

They tracked Krizhevsky down to a barely furnished appartment over a Vietnamese restaurant in San Franscisco. He no longer works in AI and is cooking up something else on his laptop.

nicely written. Sounds good. Just wanted to say that

 

[Knajjd]

nicely written. Sounds good. Just wanted to say that

 

[Knajjd]

you know you're a loser when you provide immense value to the world and no one cares while they all profit like robber barons off the research you did in the first place. He may not be the ugliest guy, but man, to come up with this idea and you're still a loner living in apartments you should just shoot yourself

He has an email address: [email protected]

Maybe ask him to join the site where he will be welcomed.

Source:

Alex Krizhevsky

 

[Adolf Kitler]

He has an email address: [email protected]

Maybe ask him to join the site where he will be welcomed.

Source:

Alex Krizhevsky

Imagine if he does join

 

[stalkerKiller]

hes not rich though, did you see the shithole apartment he lives in?

That's why you never do more than you should. Never sacrifice yourself for society because it'll always abuse you

 

[Failed Pullout]

you know you're a loser when you provide immense value to the world and no one cares while they all profit like robber barons off the research you did in the first place. He may not be the ugliest guy, but man, to come up with this idea and you're still a loner living in apartments you should just shoot yourself

It's 6'2 chad CEOs playing with his ideas and the company they inherited from boomer daddy.

 

[LOLI BREEDING]

Brutal. Cant find any informations on how much money he has

just because he lives in a random apartment doesn't mean he's poor. I live with my mom but I could easily afford several houses

still, it's very possible if he didn't involve himself further in the field especially in the early 20s that his cred went unpaid

 

[Knajjd]

Jensen Huang during his keynote address at the GTC 2024 says that AlexNet was "AI first contact".

AlexNet is the neural network developed by Alex Krizhevsky using his two GeForce GPUs ($500 each) that he ran in his bedroom at his parents house. It was the first interaction between CUDA and neural nets.

Video of Huang when he makes the comment:

 

[Knajjd]

That's why you never do more than you should. Never sacrifice yourself for society because it'll always abuse you

Sometimes a high IQ person will take pride in his work or hone his skills, no matter what they are, as a cope.

 

[Knajjd]

Brutal. Cant find any informations on how much money he has

Hinton created a start up called DNNResearch with Hinton, Sutskever and Krizhevsky each holding one third of the shares. The company had no assets other than the three men.

In 2012 Google paid $44 million so Krixhevsky ended up with close to $15 million.

Source:
The Thinking Machine by Stephen Witt (chapter 11 AlexNet)

 

[ArabAndFramecel]

Brutal

 

[Asgard]

Transformer architecture's paper was presented by a bunch of normies and curries back in 2017. No incels in sight judging from the pictures I've encountered on Google Image Search but most normal guys in general aren't that far from meeting the same fate as incels (which is both long term sexlessness and undesirability from women).

 

[ArabAndFramecel]

Transformer architecture's paper was presented by a bunch of normies and curries. No incels in sight but most guys aren't that far from meeting the same fate as incels (which is both long term sexlessness and undesirability from women).

Brutal

 

[Asgard]

Brutal

Just be born with an IQ three standard deviations above the mean, literally and figuratively. An IQ of 130 is simply not enough for actual research and novel major contributions in science, but it is enough for the most cognitive demanding STEM related curricula.

 

[Maikowski]

Some rich Jews literally stole his tech and made billions while he lives in some rundown apartment in Chinatown.

Meritocracy my ass.

 

[ArabAndFramecel]

Just be born with an IQ three standard deviations above the mean.

Brutal

 

[ArabAndFramecel]

Some rich Jews literally stole his tech and made billions while he lives in some rundown apartment in Chinatown.

Meritocracy my ass.

Because at the end looks >>>>> IQ/personnality

 

[Asgard]

Some rich Jews literally stole his tech and made billions while he lives in some rundown apartment in Chinatown.

Meritocracy my ass.

Shocking, Ashkenazi Jews being smarter on average (thanks to in-tribe enfoced eugenics over many generations).

 

[PureImagination]

The way he nervously stays in the back looking away from the camera

 

[PureImagination]

Because at the end looks

He has a gifted mind but was emotionally an idiot that’s how they got him.

 

[Asgard]

Because at the end looks >>>>> IQ/personnality

An IQ of 130 is simply not enough for actual research and novel major contributions in science, but it is high enough enough for the most cognitive demanding STEM related curricula. You're looking at the 0.3% here, actual geniuses.

 

[Asgard]

The way he nervously stays in the back looking away from the camera

He's not only virgin looking (like in the meme itself) but he's also more richer than half of this forum, if not the entire user base.

 

[ArabAndFramecel]

He has a gifted mind but was emotionally an idiot that’s how they got him.

Look at his incel frame, his glasses, jfl of course he was weak emotionnaly

Take the physiognomy pill bro

 

[ArabAndFramecel]

An IQ of 130 is simply not enough for actual research and novel major contributions in science, but it is high enough enough for the most cognitive demanding STEM related curricula. You're looking at the 0.3% here, actual geniuses.

Doesnt matter you can have 200000 IQ if you have an incel look than its over you will rot until you die

Foids dont care about IQ/petsonnality

 

[Asgard]

Doesnt matter you can have 200000 IQ if you have an incel look than its over you will rot until you die

Foids dont care about IQ/petsonnality

No. The only thing that virgin looking guy lacks is looks.

 

[PureImagination]

He's not only virgin looking (like in the meme itself) but he's also more richer than half of this forum, if not the entire user base.

Material wealth is massive cope, Poor Chad lives a more fulfilling life than he ever will be unfortunately.

 

[Asgard]

Material wealth is massive cope, Poor Chad lives a more fulfilling life than he ever will be unfortunately.

Well, assuming he's not in the verge of bankruptcy due to poor life choices (debatable due to higher intelligence primarily), I could say that's better than being one paycheck away from homelessness like your average lower middle class Joe, don't you think?

 

[PureImagination]

Well, assuming he's not in the verge of bankruptcy due to poor life choices (debatable due to higher intelligence primarily), I could say that's better than being one paycheck away from homelessness like your average lower middle class Joe, don't you think?

Yeah absolutely, money is the most important thing on this world after all. But poor Chad would mog richcel any day of the week

 

[Chromium]

Are you sure he's incel and not oofy-doofy?

 

[Suigin Trismegistus]

Incel discovered AI

 

[ArabAndFramecel]

Bump

 

[UglyDumbass]

it shows that we are intelligent but yet those damn girls are still chad only chat

 

[Knajjd]

Transformer architecture's paper was presented by a bunch of normies and curries back in 2017. No incels in sight judging from the pictures I've encountered on Google Image Search but most normal guys in general aren't that far from meeting the same fate as incels (which is both long term sexlessness and undesirability from women).

Link to Transformer research paper:

Attention Is All You Need

The dominant sequence transduction models are based on complex recurrent or convolutional neural networks in an encoder-decoder configuration. The best performing models also connect the encoder and decoder through an attention mechanism. We propose a new simple network architecture, the...

arxiv.org

Wiki page on the paper:

Attention Is All You Need - Wikipedia

en.wikipedia.org

Photos of 2 of the authors of the landmark research paper:

Transformer code below from:

GitHub - tunz/transformer-pytorch: Transformer implementation in PyTorch.

Transformer implementation in PyTorch. Contribute to tunz/transformer-pytorch development by creating an account on GitHub.

github.com

class Transformer(nn.Module):
def __init__(self, i_vocab_size, t_vocab_size,
n_layers=6,
hidden_size=512,
filter_size=2048,
dropout_rate=0.1,
share_target_embedding=True,
has_inputs=True,
src_pad_idx=None,
trg_pad_idx=None):
super(Transformer, self).__init__()

self.hidden_size = hidden_size
self.emb_scale = hidden_size ** 0.5
self.has_inputs = has_inputs
self.src_pad_idx = src_pad_idx
self.trg_pad_idx = trg_pad_idx

self.t_vocab_embedding = nn.Embedding(t_vocab_size, hidden_size)
nn.init.normal_(self.t_vocab_embedding.weight, mean=0,
std=hidden_size**-0.5)
self.t_emb_dropout = nn.Dropout(dropout_rate)
self.decoder = Decoder(hidden_size, filter_size,
dropout_rate, n_layers)

if has_inputs:
if not share_target_embedding:
self.i_vocab_embedding = nn.Embedding(i_vocab_size,
hidden_size)
nn.init.normal_(self.i_vocab_embedding.weight, mean=0,
std=hidden_size**-0.5)
else:
self.i_vocab_embedding = self.t_vocab_embedding

self.i_emb_dropout = nn.Dropout(dropout_rate)

self.encoder = Encoder(hidden_size, filter_size,
dropout_rate, n_layers)

# For positional encoding
num_timescales = self.hidden_size // 2
max_timescale = 10000.0
min_timescale = 1.0
log_timescale_increment = (
math.log(float(max_timescale) / float(min_timescale)) /
max(num_timescales - 1, 1))
inv_timescales = min_timescale * torch.exp(
torch.arange(num_timescales, dtype=torch.float32) *
-log_timescale_increment)
self.register_buffer('inv_timescales', inv_timescales)

def forward(self, inputs, targets):
enc_output, i_mask = None, None
if self.has_inputs:
i_mask = utils.create_pad_mask(inputs, self.src_pad_idx)
enc_output = self.encode(inputs, i_mask)

t_mask = utils.create_pad_mask(targets, self.trg_pad_idx)
target_size = targets.size()[1]
t_self_mask = utils.create_trg_self_mask(target_size,
device=targets.device)
return self.decode(targets, enc_output, i_mask, t_self_mask, t_mask)

def encode(self, inputs, i_mask):
# Input embedding
input_embedded = self.i_vocab_embedding(inputs)
input_embedded.masked_fill_(i_mask.squeeze(1).unsqueeze(-1), 0)
input_embedded *= self.emb_scale
input_embedded += self.get_position_encoding(inputs)
input_embedded = self.i_emb_dropout(input_embedded)

return self.encoder(input_embedded, i_mask)

def decode(self, targets, enc_output, i_mask, t_self_mask, t_mask,
cache=None):
# target embedding
target_embedded = self.t_vocab_embedding(targets)
target_embedded.masked_fill_(t_mask.squeeze(1).unsqueeze(-1), 0)

# Shifting
target_embedded = target_embedded[:, :-1]
target_embedded = F.pad(target_embedded, (0, 0, 1, 0))

target_embedded *= self.emb_scale
target_embedded += self.get_position_encoding(targets)
target_embedded = self.t_emb_dropout(target_embedded)

# decoder
decoder_output = self.decoder(target_embedded, enc_output, i_mask,
t_self_mask, cache)
# linear
output = torch.matmul(decoder_output,
self.t_vocab_embedding.weight.transpose(0, 1))

return output

def get_position_encoding(self, x):
max_length = x.size()[1]
position = torch.arange(max_length, dtype=torch.float32,
device=x.device)
scaled_time = position.unsqueeze(1) * self.inv_timescales.unsqueeze(0)
signal = torch.cat([torch.sin(scaled_time), torch.cos(scaled_time)],
dim=1)
signal = F.pad(signal, (0, 0, 0, self.hidden_size % 2))
signal = signal.view(1, max_length, self.hidden_size)
return signal