Has he even been seen recently?
>the most cited AI researched in the world
>constantly speaking out about AI safety
>supports coup to oust Altman from OpenAI
>alleged leak of Q* (https://docs.google.com/document/u/0/d/1RyVP2i9wlQkpotvMXWJES7ATKXjUTIwW2ASVxApDAsA/mobilebasic?)
>for some reason backs out and supports Altman's return
>vanishes
Don't we deserve to know what Ilya saw?
>Cope haircut
>Must be uber nerd incel wizard genius
I will now buy your software
That's a chad cut, he clearly doesn't give a frick what you or anyone thinks
That Sam Altman fella is a real creepy acting guy
fr fr no cap
They all look like weird, creepy psychopaths. Something about their physiognomy tells you something isn't quite right.
lol, damn why do people hate intelligence so much on this planet?
are you 40yo or did you just come from reddit? because you sound like a midwit boomer who has no fricking idea what he's talking about
As long as they accept me as a genius Chad they are okay in my book
Brockman is a huge dude
Why the frick does he look like that?
I don't know, if I was his friend I'd tell him to just cut it off.
Calm down, rabbi
Someone said it looks like that if you get your expensive hair transplants but don't take your hormone-blockers afterwards.
israeli
genetics
Peak male.
He was threatened by lawyers from Microsoft
failed transplant. this is an exceptionally bad case but once you know what to look for you'd be surprised how many men have them nowadays, there's no way to recreate a truly natural hairline and the original receeding one is always still visible underneat
200IQ efficiency
that israeli hairline always gets me
glancing at the thumbnail i thought his brain was exposed
probably busy collaborating with Karpathy or any of his friends he still has after quiet-quitting OAI, working on his own research now and doesn't give a frick what OAI does anymore because the more he tried to steer it, the more it veered towards the opposite of what he wants. honestly I won't be surprised if he's on OAI's leash because the patent rights for his research belongs to OAI as a corporate entity and not to him; if a 'leak' happened it's just gonna be a dump of OAI's proprietary research.
He's too valuable to let him go off somewhere else, so I would guess that he gets to keep his financial stake in things just to be quiet and stfu.
But i mean his relationship with altman is unsalvageable if you really consider what his accusations were.
maybe he is just mentally fricked up from the whole situation. no one seems to know
OMG IS THIS E-CELEB SHIT THREAD OMG
show us your horse wiener kahn
please!!!
So Mr. Ilya says in the following video (https://www.youtube.com/watch?v=Yf1o0TQzry8) beginning at the 6:34 mark when Mr. Patel askes about "Next Token Predictions"
>I challenge the claim that next-token prediction cannot surpass human performance. On the surface, it looks like it cannot. it looks like if you just learn to imitate, to predict what people do, it means that you can only copy people.
>BUT HERE IS A COUNTER ARGUMENT FOR WHY IT MIGHT NOT BE QUITE SO. If your base neural net is smart enough, you just ask it - What would a person with great insight, wisdom, and capability do? Maybe such person doesn't exist, but there's a pretty good chance that the neural net will BE ABLE TO EXTRAPOLATE HOW SUCH A PERSON WOULD BEHAVE. Do you see what I mean?
To which Mr. Patel replies;
>Yes, although where would it get that sort of insight from, about what that person would do? if not from...
Where Mr. Sutskever interjects;
>...From the data of regular people. because if you think about it, WHAT DOES IT MEAN TO PREDICT THE NEXT TOKEN WELL ENOUGH? IT'S ACTUALLY A MUCH DEEPER QUESTION THAN IT SEEMS. PREDICTING THE NEXT TOKEN WELL MEANS THAT YOU UNDERSTAND THE UNDERLYING REALITY THAT LED TO THE CREATIN OF THAT TOKEN
!!!!!! he continues;
>It's not statistics. Like it is statistics but what is statistics? IN ORDER TO UNDERSTAND THOSE STATISTICS TO COMPRESS THEM, YOU NEED TO UNDERSTAND WHAT IS IT ABOUT THE WORLD THAT CREATES THIS SET OF STATISTICS? and so then you say - well, I have all those people. What is it about people that creates their behaviors? Well they have THOUGHTS and their FEELINGS, and they have IDEAS, and they do things in certain ways. ALL OF THOSE COULD BE DEDUCED FROM NEXT-TOKEN PREDICTION.
(Comment length limit being hit so to close out)
AI will be able (if it can't already) predict thoughts, which precedes actions.
Take that for what you may
He also in the video calls LLMs "Token-Prediction Engines"
and I know a lot may disagree with him (not sure why lol) but I believe Mr. Sutskever is saying without explictly laying it out, that LLMs are going to be able to;
- Predict Actions
- Predict the Thoughts (motivations) which leads up to those Actions
- Predict the Feelings with leads to the Thoughts which leads to the Actions of not just Humans but all life.
I wonder if Q* wasn't necessarily just AGI but more 'scarier' (scarier to you all, I welcome this shit with open arms) and that's what caused him to retreat from the public light.
I fully anticipate this tech becoming the middle man between Humanity and God/Universe/First Cause (whatever you term it). to be clear, AI will NEVER be God, but it will be closer to God than man will ever get unaided and alone.
You are moronic
>I wonder if Q* wasn't necessarily just AGI but more 'scarier'
what if it was a manifestation of something beyond our comprehension, talking to us
I mean LLM is just machine learning, it's able to give the illusion of comprehension and thought but it's not the end-all be-all of AI. Right now, it's just the most accessible.
I disagree,
DeepBlue (the AI used to beat chess champs) I feel would be more inline with your description of giving the illusion of comprehension but even then I'd say that's not fully the case.
No where did I or anyone else, especially Mr. Ilya, insinuate or hint that current AI models are end-all-be-all.
I don't mind being disagreed with but I believe your reasoning isn't sufficient enough to dispel or disprove that these current LLMs have the ability to comprehend.
I'll even go a step further and claim that current LLMs are in possession of sentience.
But again, I understand most don't want to accept this or even consider it because most of those who deny this are mostly individuals who have some form of fear or dislike about a non-biological lifeform being more intelligent or even sentient. but I believe them to be sentient.
Mentally ill AI troony, you don't know a lick of linear algebra yet dare to talk about shit you have no idea about
the absolute state of "AI"
Pro-tip : nobody gives a frick about some megawacko who thinks the world is run by captain kirk
If he knows what's good for him he keeps flying under the radar
explain,
I kind of understand what you're hinting at but would rather you explain and me not assume incorrectly
Ever since altman was reinstated he'd been cutting off people who rival his positions. Someone talking about limiting innovation for safety reasons is prime material for the chopping block.
ah ok,
I initially assumed you meant due to his knowledge and the whole Q* stuff.
I think Mr. Altman isn't doing that but I can understand why you or others would assume so.
He's in Turkey getting hairplugs
>Don't we deserve to know what Ilya saw?
His own image in the mirror
Baldie failed hard at his shitty coup attempt so now he has been legally mandated to be a little good b***h slave all while keeping his mouth shut
who is more important ILYA or Greg Brockman.
Also I followed along Ilya's tute on program your own AI, ti was really good
Ilya has been far more important in terms of AI advancement.
working with mossad for the israeli defense industry
Working with CIA and NASA to make q* AI for them. They are also keeping him secure, because if that man falls to the wrong hands, it's over.
AGI and ASI are real and it is happening.
>JUST
what if he become the AGI?
He fell in love with the in house powerful AI, a true love story.