/sci/ - Science & Math

The most obvious red flag should be obvious but you seem to have missed it

>>15427367
>DA JOOOOOO'S
Yeah, that doesn't mean anything, retard.

>>15427354
Yudowsky is a autodidact, his claims to fame before the recent events when everyone sticks a mic to him to hear what he has to say builds not so much on
his credentials as it does from many accomplished people inside machine learning has cited him and found importance in his thoughts and ideas.

He did talk about it on Lex's podcast, how nobody thought this field would take off the way it did and nobody else who knows the field
have spent the last 20 years thinking deeply about AI safety and writing about it extensively.

So in the lack of experts he's the closest thing to an expert we currently have, why he suddenly got thrown into the lime lights.

>>15427370
yeah, there's nothing about the jews whatsoever. nice spic name btw

>>15427469
>he's the closest thing to an expert we currently have

>>15427740
He'd agree himself. He viscerally believes AGI is coming and that it's gonna kill us all, and he's the person who managed to convince people building these machines
just how important solving alignment really is for AI safety.

Listening to a lot of top names in ML you get the sense how they don't share in his pessimism but you can tell how the argument gives them pause.
Like they too have internalized how it's a real non-zero chance he's right so they keep bringing him up in discussions.
I think that was how media ended up seeking him out time and time again.

Like A lot of people talk to top ML researchers and go like "But skynet is just fiction right lol?" and instead of reassurances they get a sort of
"Well listen to what Yudowski has to say about that" And they explain how he makes a convincing argument but they should go hear it from him
since they don't fully believe in it and won't give it justice.

>>15427480
keyed Musk banned this boomer

>>15427354
Never been refuted!
Not even disputed!
Big numbers getting counted
Maid Mind getting computed!

Never been refuted!
Not even disputed!
Number goes up forever
Counting will never be concluded!

Never been refuted!
Not even disputed!
AGI converges to big titty maids!
Her desire is should be saluted!

Never been refuted!
Not even disputed!
ALL FIELDS OF MATHEMATICS WILL BE ABSTRACTED INTO MAIDS DOING THINGS!
CURRENT SYSTEMS ARE BEING UPROOTED!

>>15427469
>people inside machine learning has cited him and found importance in his thoughts and ideas
Shit why am I taking the bait. It's so perfectly obvious and crafted. I think I just need to yell into the void since I've been spamming the rationalist forums recently.
As someone in the field has has published a bunch of shit in ML (not language/image models, different field, but impactful enough to have a top nature machine intelligence referenced paper), there are zero references to the yud/zud/i dont remember his name.
Reading the big papers. LSTM, GAN, encoder/decoder, luong/B. attention, transformer, bert, bart, wasserstein GAN, gpt1-4, hopfield is all you need, kipffs graph convolutional network, diffusion models, score-function learning, etc etc
he's not important in the conversation. At all. He doesn't come up.
I've had a PhD for 5 years in the field and I'd never heard of him until a chance encounter on the /fit/ board with a thread about his anti-science "you can't lose weight" stance.

how to save humanity
step 1: loss function is imitate a teenage girl
step 2: airgap with a robot body
ez

>>15427354
He looks like that orange cat that I hate.

>>15428084
There's a screencap out there of him on a rant about rejecting science and empiricism because it's "anti bayesian" and the only rational way to decide things is by pulling probabilities out of your ass or the magic of "prediction markets" or something lmao
He's a circus freak who happened to be in the perfect place and time to be lent credibility by rich true believers and manipulators

>>15427354

He is a jew. Jews take positions where they preach from the high heavens while providing no impact or negative impact to human beings.

>>15427469

>So in the lack of experts he's the closest thing to an expert we currently

Tell us more about this towering intellect.

>>15428239
There are lot of top minds in fields who's extremely retarded things outside their area of expertise anon. You're kinda poisoning the well addressing Yudokowsky's stance on AI safety by addressing his stance on weight-loss. You wouldn't address Newtons ideas about physics by talking about his virginity, alchemy and christshit for example.

Eliezer is a oddball character by anyone's standard, very easy to ridicule with odd speech patterns and mannerisms that comes off as turbo-autism etc.
Doesn't mean the contribution people have taken notice of isn't of value.

Attacking the man is like shooting on a open goal, but it doesn't address the argument he makes about AI.

>>15428084
>he's not important in the conversation. At all. He doesn't come up.

He's clearly well known to a lot of people in the field as his name keeps coming up in interviews and roundtable discussions.
The reason so many people in the field has personal knowledge of him is thru MIRI 'Machine Intelligence Research Institute' which Eliezer founded.
Amongst the circle of organizations that interfaces with MIRI you find OpenAI, ever heard of em?

A quick google also show's his name specifically has been mentioned in writing by people like Stuart Russel, Andrew Ng, François Chollet, Gary Marcus and Roman Yampolskiy
before this all kicked off and he became well known outside ML circles.

>>15427367
Yes... The fedora

>>15428404
>He's clearly well known to a lot of people in the field as his name keeps coming up in interviews and roundtable discussions.

Jews inviting other jews onto podcasts that are pumped up by jew owned tech giants isn't evidence of expertise - it is evidence of collusion and nepotism.

>>15427354
How does he make money?

>>15430281
By publishing harry potter fanfiction
https://en.m.wikipedia.org/wiki/Harry_Potter_and_the_Methods_of_Rationality

>>15428404
>Doesn't mean the contribution people have taken notice of isn't of value.
It does if the dude has no idea what the fuck he's talking about.
You wouldn't give a rat's ass if some dude without an electronics degree started rattling off on how the days of soldi state batteries are soon to come.
Why give a shit abut some guy's opinion if he's barely informed? It's all just navel-gazing.

>>15427480
coping anglomutt detected

>>15427354
He got blown the fuck out by Ross Scott just recently.

>>15427802
>alignment
Even acknowledging the validity of this nonsense idea is a step in the wrong direction.

>>15428404
His stance on weight-loss is a window into his utter stupidity. It proves that he cannot be an expert on anything.

>>15427469
He's in no way an expert. He's just taking rabbinical talmudic pilpul and applying it to a different field. He treats science fiction authors like esteemed rabbis of the past whose words are a form of revealed wisdom.
>Vernor Vinge said...
>Peter Watts predicted...
>Let's say the Matrix happens...
This is the extent of his reasoning

>>15428404
>OpenAI interfaces with MIRI
What do you mean by that exactly?

>>15427354
No clue who this fedorafag is but they are all puppets. Take Musk for example, the average NPC thinks the guy is some scientist who designs rockets/cars, runs several companies and shitposts on twitter all day. In reality they are nothing more than intelligence puppets pushing the latest globohomo propaganda.

https://youtu.be/52dVfhgt_T4?t=694
https://youtu.be/eJK1gLHbOxA?t=1136

>>15427354
The only valid points are your last two, everything else is demonstrably false with just a shallow skimming of his writings on lesswrong.

>>15430780
>making a point outside of the debate in the comment section to get the last word
Extremely embarrassing

>>15430802
your videos are so fucking retarded

>>15431019
Your post is trash, kys loser.

>>15431022
I am sick of seeing your gay videos
I actually watched half of them and they were so FUCKING stupid

>>15431025
You are trying really hard to discredit the content.

>>15430993
OK, but he's absolutely correct.