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

AlphaCode Attention Visualization

alphacode.deepmind.com

Post reply on HN
Print view View original

41–50 of 51 posts

Re: AlphaCode Attention Visualization

#41

Post by spaceman_2020 »

jw1224 wrote: post #7

Has Google/Alphabet publicly released any of their AI models yet? I’ve seen plenty of hype surrounding Imagen [1] and Parti [2], but as far as I know, they’re still vaporware. Normally I wouldn’t think too hard about this, but two things come to mind here: Firstly, the speed at which competitors are launching and developing new AI models. Imagen/Parti both seem to rival Stable Diffusion and DALL-E… why can’t we use t…

Google has no incentive to share its AI research with the public. If it does release a tool like chatGPT, it will just increase calls to incorporate the tool into regular search. If Google does that, it willingly eats into its own ad-driven business.

Companies have often ignored newer tech because it eats into their existing business (Kodak and digital cameras, for example).

Google CAN'T adopt AI at this point, at least not without drastic changes to its revenue model.

Re: AlphaCode Attention Visualization

#42

Post by elcomet »

spaceman_2020 wrote: post #41
jw1224 wrote: post #7

Has Google/Alphabet publicly released any of their AI models yet? I’ve seen plenty of hype surrounding Imagen [1] and Parti [2], but as far as I know, they’re still vaporware. Normally I wouldn’t think too hard about this, but two things come to mind here: Firstly, the speed at which competitors are launching and developing new AI models. Imagen/Parti both seem to rival Stable Diffusion and DALL-E… why can’t we use t…

Google has no incentive to share its AI research with the public. If it does release a tool like chatGPT, it will just increase calls to incorporate the tool into regular search. If Google does that, it willingly eats into its own ad-driven business. Companies have often ignored newer tech because it eats into their existing business (Kodak and digital cameras, for example). Google CAN'T adopt AI at this point, at le…

> If Google does that, it willingly eats into its own ad-driven business.

I don't see why. They could still incorporate ads in the output of such a tool..

Re: AlphaCode Attention Visualization

#43

Post by spaceman_2020 »

elcomet wrote: post #42
spaceman_2020 wrote: post #41

Earlier quoted context omitted.

Google has no incentive to share its AI research with the public. If it does release a tool like chatGPT, it will just increase calls to incorporate the tool into regular search. If Google does that, it willingly eats into its own ad-driven business. Companies have often ignored newer tech because it eats into their existing business (Kodak and digital cameras, for example). Google CAN'T adopt AI at this point, at le…

> If Google does that, it willingly eats into its own ad-driven business. I don't see why. They could still incorporate ads in the output of such a tool..

Impressions and CTRs for ads will go down drastically. CTRs for ads are already non-existent for any queries for which Google shows direct answers (such as "when was lincoln born").

Re: AlphaCode Attention Visualization

#44

Post by iLoveOncall »

DoingIsLearning wrote: post #38
iLoveOncall wrote: post #16

The fact that the amount of text you have to write is greater than the amount of code AlphaCode writes should reassure programmers afraid that their job will ever be taken over by AI.

We should really let go of words per minute or typing effort as a useful metric. Most of our time is spent thinking about what to write not the actual writing. Also nothing stopping it from being a voice input rather than typing.

That's not what I meant. I'm saying that it is more overall effort to code this way than to code by just writing code.

Re: AlphaCode Attention Visualization

#45

Post by Delemono »

Al-Khwarizmi wrote: post #40
ipsum2 wrote: post #35

Earlier quoted context omitted.

The arrogance of this comment is astoundingly funny. > Firstly, the speed at which competitors are launching and developing new AI models. Imagen/Parti both seem to rival Stable Diffusion and DALL-E… why can’t we use them yet? Because Google doesn't want you to have access to them? Why do you feel like you're entitled to their internal research? Google releases papers on robots[0] as well. Do you expect them to ship…

I'm not the user you replied to, but I have the same view as them. And it's not that we are entitled to anything, it's that they're losing the race, or at least the image race. For example, as an AI researcher, I can't consider Imagen/Parti to be the state of the art if all we have are cherry-picked examples and we can't verify anything. For all practical intents and purposes, they are just vaporware, and the state o…

They show what it can do, people in Google can use it,.there are papers.

Google publishes plenty of other ml based papers.

Assuming in any way that Google might loose it's reputation because of imagine is very very far fetched.

Re: AlphaCode Attention Visualization

#46

Post by elcomet »

spaceman_2020 wrote: post #43
elcomet wrote: post #42

Earlier quoted context omitted.

> If Google does that, it willingly eats into its own ad-driven business. I don't see why. They could still incorporate ads in the output of such a tool..

Impressions and CTRs for ads will go down drastically. CTRs for ads are already non-existent for any queries for which Google shows direct answers (such as "when was lincoln born").

I completely disagree, I think it is very unclear. They could show one direct answer + one or a few related ad, which I think might increase a lot the CTR.

Re: AlphaCode Attention Visualization

#47

Post by Al-Khwarizmi »

Delemono wrote: post #45
Al-Khwarizmi wrote: post #40

Earlier quoted context omitted.

I'm not the user you replied to, but I have the same view as them. And it's not that we are entitled to anything, it's that they're losing the race, or at least the image race. For example, as an AI researcher, I can't consider Imagen/Parti to be the state of the art if all we have are cherry-picked examples and we can't verify anything. For all practical intents and purposes, they are just vaporware, and the state o…

They show what it can do, people in Google can use it,.there are papers. Google publishes plenty of other ml based papers. Assuming in any way that Google might loose it's reputation because of imagine is very very far fetched.

Not losing its reputation as a whole, but losing the reputation of being the leader in this field? Sure.

When they released BERT there was no doubt that they were the leaders. Even laymen heard about it.

What AI advances do laymen most talk about now? DALL-E, Stable Diffusion and ChatGPT. AlphaCode and LaMDA gave some headlines but not even close. Everyone is too busy trying ChatGPT to pay attention to those.

Re: AlphaCode Attention Visualization

#49

Post by duckydude20 »

i don't know much about ai or anything. but i would really like if ai can replace competitive programming. i really hate it. it's not programming and a very bad merit to judge. it's one of the thing i hates the most. i hope this type of things makes us rethink how we are judging the people. and how cp rank one cannot design a full system which is maintainable and easy to reason about.

Re: AlphaCode Attention Visualization

#50

Post by alphabetting »

Al-Khwarizmi wrote: post #47
Delemono wrote: post #45

Earlier quoted context omitted.

They show what it can do, people in Google can use it,.there are papers. Google publishes plenty of other ml based papers. Assuming in any way that Google might loose it's reputation because of imagine is very very far fetched.

Not losing its reputation as a whole, but losing the reputation of being the leader in this field? Sure. When they released BERT there was no doubt that they were the leaders. Even laymen heard about it. What AI advances do laymen most talk about now? DALL-E, Stable Diffusion and ChatGPT. AlphaCode and LaMDA gave some headlines but not even close. Everyone is too busy trying ChatGPT to pay attention to those.

Does elite AI talent really care what layman new to AI think? If so, problem for Google but I doubt they do.
Post reply on HN