Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Noam Shazeer Joins OpenAI

twitter.com

Post reply on HN
Print view View original

281–290 of 478 posts

Re: Noam Shazeer Joins OpenAI

#281

Post by jvican »

ahmadyan wrote: post #268
SiempreViernes wrote: post #154

Earlier quoted context omitted.

> Using what team members would later describe with words like “magic” and “alchemy” and “bells and whistles,” Ok, these peopl have all gotten extensive training on how to hype for the non-technical crowd without saying anything of substance.

As a hacker, I kinda like naom's code. I was had to implement a TC MoE kernel, and stumbled upon his code from [tensor2tensor]( https://github.com/tensorflow/tensor2tensor/blob/master/tens... ) and i think "alchemy" is justified. Dude writes some beautiful kernels. He also saw LLM would replace search before anyone else, and that is something to look at the Lamda or GPT-1's output and think: yeah this will answer all…

There's no doubt about Noam's abilities. But I read through that code, and struggle to see its 'magic' or 'alchemy'. Can you elaborate what you find especially good about that code? (You may assume GPU kernel programming knowledge on my end.)

Re: Noam Shazeer Joins OpenAI

#282

Post by adamsb6 »

nhinck2 wrote: post #213
bbeonx wrote: post #53

Earlier quoted context omitted.

[flagged]

People who have intellectually based skills think it makes them an intellectual. They fail to understand that their skill doesn't generalise. That and the hyperglazing and platforming they get for having said skill makes them a prime candidate for exposing how average they are.

What a brave new world where only machines possess general intelligence.

Re: Noam Shazeer Joins OpenAI

#283

Post by nl »

gzer0 wrote: post #48

Some context for people who haven’t followed the full loop: Shazeer was a long-time Google researcher, joined Google in 2000, and was one of the co-authors of “Attention Is All You Need.” He left Google in 2021 to co-found Character.AI. In 2024, Google brought him and some Character.AI researchers back via a licensing/talent deal with Character.AI (reportedly around $2.7B). He was then made a Gemini co-lead. Now he’s…

For context, the reason he left Google the first time was because Google wouldn't ship the chatbot-type products that he saw were possible.

Google bought him back (with lots of money) and made him one of the leads of Gemini.

Re: Noam Shazeer Joins OpenAI

#284

Post by dekhn »

jvican wrote: post #281
ahmadyan wrote: post #268

Earlier quoted context omitted.

As a hacker, I kinda like naom's code. I was had to implement a TC MoE kernel, and stumbled upon his code from [tensor2tensor]( https://github.com/tensorflow/tensor2tensor/blob/master/tens... ) and i think "alchemy" is justified. Dude writes some beautiful kernels. He also saw LLM would replace search before anyone else, and that is something to look at the Lamda or GPT-1's output and think: yeah this will answer all…

There's no doubt about Noam's abilities. But I read through that code, and struggle to see its 'magic' or 'alchemy'. Can you elaborate what you find especially good about that code? (You may assume GPU kernel programming knowledge on my end.)

To me the magic Noam moment was when he came to my team and said "that cluster has a bad node in it, but this other one doesn't" and we had to spend like a week tracking down a single bad processor out of thousands.

Re: Noam Shazeer Joins OpenAI

#285

Post by wk_end »

frollogaston wrote: post #233
manwithopinions wrote: post #227

Earlier quoted context omitted.

Accusing coworkers of being antisemitic crosses the line, but accusing coworkers of sterilizing children and denying the existence of gender is ok? Surely both are bad, neither is acceptable in a workplace. Do you mean it’s not dumb because you share his views?

The first one doesn't accuse coworkers of sterilizing children. As for saying there's no gender attribute, I disagree, but it's just his belief and not a dumb one either.

The "sterilizing children" is how anti-trans activists talk about giving trans children gender-affirming care. Framing it this way makes it sound monstrous rather than an unfortunate side-effect of a medically necessary procedure, in the same way that characterizing a surgeon who performs hysterectomies for women with ovarian cancer as "a doctor who goes around sterilizing women" would be painting them in an unfair light.

And of course he's not directly accusing his coworkers of "sterilizing children", but he's 1) using language that compares politically sensitive health services that many of his coworkers or their families may have used and/or may feel defensive about to atrocities and 2) accusing his coworkers of supporting atrocities. That feels quite disruptive and inappropriate in the work environment IMO.

Re: Noam Shazeer Joins OpenAI

#286

Post by jeswin »

jvican wrote: post #281
ahmadyan wrote: post #268

Earlier quoted context omitted.

As a hacker, I kinda like naom's code. I was had to implement a TC MoE kernel, and stumbled upon his code from [tensor2tensor]( https://github.com/tensorflow/tensor2tensor/blob/master/tens... ) and i think "alchemy" is justified. Dude writes some beautiful kernels. He also saw LLM would replace search before anyone else, and that is something to look at the Lamda or GPT-1's output and think: yeah this will answer all…

There's no doubt about Noam's abilities. But I read through that code, and struggle to see its 'magic' or 'alchemy'. Can you elaborate what you find especially good about that code? (You may assume GPU kernel programming knowledge on my end.)

Unrelated to the particular code above. There's a difference between writing code about or adjacent to a proven idea vs writing code in uncharted territory. I suspect that is what happened here. It's the same thing with say music and art. A lot of people today can play Chuck Berry.

Re: Noam Shazeer Joins OpenAI

#287

Post by woodruffw »

Avicebron wrote: post #105
bigyabai wrote: post #91

Earlier quoted context omitted.

[flagged]

the g-slur? I won't say it in case it is a slur, is that the word the jews call non-jews?

Whether or not “goy” is a slur is pretty complicated. It has pejorative uses, and outside of a strictly religious context any non-Jew is almost certainly only going to see those pejorative uses. But strictly speaking it’s a Hebrew word that means “nation,” and isn’t any more or less offensive in the abstract than Jew, Arab, Brit, etc.

(To my understanding, the closest equivalent is “ummah” in Arabic, where the connotation is flipped: goy can refer to a Jewish person but typically does not, whereas the ummah typically refers to Muslim peoples as a collective but can also be a general stand-in for “nation” or “world.”)

Re: Noam Shazeer Joins OpenAI

#288

Post by efficax »

VirusNewbie wrote: post #247
efficax wrote: post #107

Earlier quoted context omitted.

you'd be surprised! people seem to have a limitless appetite for that money stuff. they just can't get enough of it, i've found

How much money do you have? or are you just commenting from the peanut gallery?

i’ve got 300k in a 401k, 50k in cash and i earn 230k pretax. i’m not sure what the point of the question is tho

Re: Noam Shazeer Joins OpenAI

#290

Post by tomhow »

bbeonx wrote: post #145
bigyabai wrote: post #91

Earlier quoted context omitted.

[flagged]

[flagged]

Please don't comment like this here. The HN guidelines ask us not to engage in political or ideological battle or use swipes like "comically stupid take". Same goes for your other comment in this thread (https://news.ycombinator.com/item?id=48590901) – "sigh how are so many brilliant people this stupid?" adds nothing but venom to the discussion.

Please read the guidelines and make an effort to observe them if you want to participate here. https://news.ycombinator.com/newsguidelines.html

Post reply on HN