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

Greg Brockman quits OpenAI

twitter.com

Post reply on HN
Print view View original

541–550 of 701 posts

Re: Greg Brockman quits OpenAI

#541

Post by blackoil »

rchaud wrote: post #443
strikelaserclaw wrote: post #396

Earlier quoted context omitted.

Just look at how this news is doing on reddit (a service i conflate a little more with the general public than hacker news which learns towards silicon valley technology) and you can easily see the truth of your statement.

Interestingly parts of this comment section are behaving in Reddit-y ways, posting board members' Linkedins and questioning their credentials, as if their jobs are to just rubber stamp the CEO's calls.

that's not un-HN. Noone has any clue, and all are speculating, so it is one theory after another. We have dozen posts on capitalization.

Re: Greg Brockman quits OpenAI

#542

Post by leptons »

siva7 wrote: post #501
johnwheeler wrote: post #84

Edit: I called it https://twitter.com/karaswisher/status/1725682088639119857 nothing to do with dishonesty. That’s just the official reason. ———- I haven’t heard anyone commenting about this, but the two main figures here-consider: This MUST come down to a disagreement between Altman and Sutskever. Also interesting that Sutskever tweeted a month and a half ago https://twitter.com/ilyasut/status/1707752576077176907 Th…

Ah, the old myth about the irreplaceable engineer and the dumb suit. Ask Wozniak about that. I don't think he believes Apple would be without Steve Jobs.

Steve Jobs would be nothing without Wozniak to design something people wanted.

Re: Greg Brockman quits OpenAI

#544

Post by siva7 »

leptons wrote: post #542
siva7 wrote: post #501

Earlier quoted context omitted.

Ah, the old myth about the irreplaceable engineer and the dumb suit. Ask Wozniak about that. I don't think he believes Apple would be without Steve Jobs.

Steve Jobs would be nothing without Wozniak to design something people wanted.

vice versa ;)

Re: Greg Brockman quits OpenAI

#545

Post by MVissers »

mlyle wrote: post #467
darkerside wrote: post #464

Earlier quoted context omitted.

> Without persistence outside of the context window, they can't even maintain a dynamic, stable higher level goal. I mean, can't you say the same for people? We are easily confused and manipulated, for the most part.

I can remember to do something tomorrow after doing many things in-between. I can reason about something and then combine it with something I reasoned about at a different time. I can learn new tasks. I can pick a goal of my own choosing and then still be working towards it intermittently weeks later. The examples we have now of GPT LLM cannot do these things. Doing those things may be a small change, or may not be t…

Former neuroscientist here.

Our brain actually uses many different functions for all of these things. Intelligence is incredibly complex.

But also, you don't need all of these to have real intelligence. People can problem solve without memory, since those are different things. People can intelligently problem-solve without a task.

And working towards long-term goals is something we actually take decades to learn. And many fail there as well.

I wouldn't be surprised if, just like in our brain, we'll start adding other modalities that improve memory, planning, etc etc. Seems that they started doing this with the vision update in GPT-4.

I wouldn't be surprised if these LLMs really become the backbone of the AGI. But this is science– You don't really know what'll work until you do it.

Re: Greg Brockman quits OpenAI

#546

Post by erk__ »

underseacables wrote: post #164
adastra22 wrote: post #148

Earlier quoted context omitted.

ikea is a non-profit?!?

IKEA is the world's largest nonprofit. https://www.economist.com/business/2006/05/11/flat-pack-acco...

I think the Novo Nordisk foundation is the largest now. It owns a majority of both Novo nordisk and Novozymes.

https://en.wikipedia.org/wiki/Novo_Nordisk_Foundation

Re: Greg Brockman quits OpenAI

#547

Post by calf »

blackoil wrote: post #494
mlyle wrote: post #467

Earlier quoted context omitted.

I can remember to do something tomorrow after doing many things in-between. I can reason about something and then combine it with something I reasoned about at a different time. I can learn new tasks. I can pick a goal of my own choosing and then still be working towards it intermittently weeks later. The examples we have now of GPT LLM cannot do these things. Doing those things may be a small change, or may not be t…

That just proves we real-time fine tuning of the neuron weights. It is computationally intensive but not fundamentally different. A million token context would look close to long short-term memory and frequent fine-tuning will be akin to long-term memory. I most probably am anthropomorphizing completely wrong. But point is humans may not be any more creative than an LLM, just that we have better computation and input…

Are there theoretical models that use real time weights? Every intro to deep learning focuses on stochastic gradient descent for neural network weights; as a layperson I'm curious about what online algorithms would be like instead.

Re: Greg Brockman quits OpenAI

#548

Post by calf »

jumploops wrote: post #244

Sam and Ilya have recently made public statements about AGI that appear to highlight a fundamental disagreement between them. Sam claims LLMs aren't sufficient for AGI (rightfully so). Ilya claims the transformer architecture, with some modification for efficiency, is actually sufficient for AGI. Obviously transformers are the core component of LLMs today, and the devil is in the details (a future model may resemble…

Did Ilya give a reason why transformers are theoretically sufficient? I've watched him talk in a CS seminar and he's certainly interesting to listen to.

Re: Greg Brockman quits OpenAI

#550

Post by tinco »

cedws wrote: post #517
jumploops wrote: post #244

Sam and Ilya have recently made public statements about AGI that appear to highlight a fundamental disagreement between them. Sam claims LLMs aren't sufficient for AGI (rightfully so). Ilya claims the transformer architecture, with some modification for efficiency, is actually sufficient for AGI. Obviously transformers are the core component of LLMs today, and the devil is in the details (a future model may resemble…

>Ilya claims the transformer architecture, with some modification for efficiency, is actually sufficient for AGI. I thought this guy was supposed to know what he's talking about? There was a paper that shows LLMs cannot generalise[0]. Anybody who's used ChatGPT can see there's imperfections. [0] https://arxiv.org/abs/2309.12288

Humans don't work this way either. You don't need the LLM to do the logic, you just need the LLM to prepare the information so it can be fed into a logic engine. Just like humans do when they shut down their system 1 brain and go into system 2 slow mode.

I'm in the definitely ready for AGI camp. But it's not going to be a single model that's going to do the AGI magic trick, it's going to be an engineered system consisting of multiple communicating models hooked up using traditional engineering techniques.

Post reply on HN