4 ms·

Google shared AI knowledge with the world – until ChatGPT caught up

59 points by sonabinu 3y ago

metadat 3y ago
https://archive.today/rUQxM https://archive.today/rUQxM
herdcall 3y ago
Most believe ChatGPT didn't just catch up but overtook...a while back.
neel8986 3y ago
Issue is this complete stops open research in AI. Last decade was really golden period of AI. Every major industrial lab collaborated like peer groups. Google Brain, Deepmind, FAIR, MSR every one played their role. We went from broken Alex net to trillion parameter transformer in a matter of 10 years. Everything stops now. I don't think any thing of the scale of transformer or even resnet will ever be shared. Guys with suits ( or VC with hoodies ) have reached the party and researcher . I really really hope open source model wins. That is the only way to unlock the knowledge and make real progress. Even OpenAI admits the path to larger and larger model stops very soon. We need more innovation and that innovation needs to be shared
- runnerup 3y ago
  We’ll have some ebb and flow between industry and academia which will spread concepts just as it does in other industries. It’s slower, but it still gets the job done.
- cushpush 3y ago
  In my opinion, we are STILL in the AI winter and things were just barely beginning to thaw, back to deep freeze.
13years 3y ago
Google "We Have No Moat, And Neither Does OpenAI" https://www.semianalysis.com/p/google-we-have-no-moat-and-neither https://www.semianalysis.com/p/google-we-have-no-moat-and-ne... AI has completely changed technological competition. Most still don't perceive that AI is essentially a skill/technology replication machine. This changes everything. It is not comparable to any technological innovation prior. "Anything you create can be replicated without investment cost while also being unique in design as well as delivering the same function or experience. Intellectual property laws essentially have no function in this new environment and there isn't an obvious remedy" https://dakara.substack.com/p/ai-and-the-end-to-all-things https://dakara.substack.com/p/ai-and-the-end-to-all-things
iamleppert 3y ago
I wouldn't be worried. This is just Google operating from the same playback as when they failed at social. It's as if Google wasn't invited to the party, so they decided to throw their own. And then go on about how some of the guests at the former stole their outfit and game ideas. If you recall, Eric Schmidt and Larry Page put the company on high alert and made it a priority to integrate Google+ into all their products with the same finesse as mixing oil and water with predictable results. However, if Google's failure at social and their complete inability to seriously penetrate any other mass consumer market via organic efforts without an acquisition shows -- Google may have deep technical chops but their product and go to market is essentially broken. This is not a technical issue that can be fixed with the same approach Google takes to everything, but rather a culture issue from within the company itself. They have repeatedly demonstrated they can't launch anything new that remains sticky -- what is the last major new Google product you can't live without? They publish a lot of papers but routinely can't see the forest for the trees. Google had all the technical resources in the world (and in some cases, even a launched product) to compete with social but lacked the culture of Facebook, Twitter, Instagram at the time to adequately compete with it. Executives are so far removed from what actually happens and comfortable such that they aren't willing to take big bets and have a limited attention span, with too many choices for which to pick their favorites. We're talking about the same company here that tried to innovate on one of their core products with Google Wave. The bar for success is so incredibly high at Google and that is their weakness: they are unable to make any single new thing a success. A startup has a limited amount of time, money and pivots; at Google they are essentially unlimited and that will be their downfall.
- andromeduck 3y ago
  At least for AI there are obvious user stories like fixing auto awesome which is currently quite the travesty, or auto summarizing Gmail or rewriting/rewording in docs/Gmail better than the gunk that exists today.
muro 3y ago
"until ChatGPT became better than Google's version, bulding on the research published by Google and others, yet deciding to not publish their research" would probably be closer than the headline.

59 points by sonabinu 3y ago

metadat 3y ago

https://archive.today/rUQxM https://archive.today/rUQxM

herdcall 3y ago

Most believe ChatGPT didn't just catch up but overtook...a while back.

neel8986 3y ago

Issue is this complete stops open research in AI. Last decade was really golden period of AI. Every major industrial lab collaborated like peer groups. Google Brain, Deepmind, FAIR, MSR every one played their role. We went from broken Alex net to trillion parameter transformer in a matter of 10 years. Everything stops now. I don't think any thing of the scale of transformer or even resnet will ever be shared. Guys with suits ( or VC with hoodies ) have reached the party and researcher . I really really hope open source model wins. That is the only way to unlock the knowledge and make real progress. Even OpenAI admits the path to larger and larger model stops very soon. We need more innovation and that innovation needs to be shared

runnerup 3y ago
We’ll have some ebb and flow between industry and academia which will spread concepts just as it does in other industries. It’s slower, but it still gets the job done.
cushpush 3y ago
In my opinion, we are STILL in the AI winter and things were just barely beginning to thaw, back to deep freeze.

13years 3y ago

Google "We Have No Moat, And Neither Does OpenAI" https://www.semianalysis.com/p/google-we-have-no-moat-and-neither https://www.semianalysis.com/p/google-we-have-no-moat-and-ne... AI has completely changed technological competition. Most still don't perceive that AI is essentially a skill/technology replication machine. This changes everything. It is not comparable to any technological innovation prior. "Anything you create can be replicated without investment cost while also being unique in design as well as delivering the same function or experience. Intellectual property laws essentially have no function in this new environment and there isn't an obvious remedy" https://dakara.substack.com/p/ai-and-the-end-to-all-things https://dakara.substack.com/p/ai-and-the-end-to-all-things

iamleppert 3y ago

I wouldn't be worried. This is just Google operating from the same playback as when they failed at social. It's as if Google wasn't invited to the party, so they decided to throw their own. And then go on about how some of the guests at the former stole their outfit and game ideas. If you recall, Eric Schmidt and Larry Page put the company on high alert and made it a priority to integrate Google+ into all their products with the same finesse as mixing oil and water with predictable results. However, if Google's failure at social and their complete inability to seriously penetrate any other mass consumer market via organic efforts without an acquisition shows -- Google may have deep technical chops but their product and go to market is essentially broken. This is not a technical issue that can be fixed with the same approach Google takes to everything, but rather a culture issue from within the company itself. They have repeatedly demonstrated they can't launch anything new that remains sticky -- what is the last major new Google product you can't live without? They publish a lot of papers but routinely can't see the forest for the trees. Google had all the technical resources in the world (and in some cases, even a launched product) to compete with social but lacked the culture of Facebook, Twitter, Instagram at the time to adequately compete with it. Executives are so far removed from what actually happens and comfortable such that they aren't willing to take big bets and have a limited attention span, with too many choices for which to pick their favorites. We're talking about the same company here that tried to innovate on one of their core products with Google Wave. The bar for success is so incredibly high at Google and that is their weakness: they are unable to make any single new thing a success. A startup has a limited amount of time, money and pivots; at Google they are essentially unlimited and that will be their downfall.

andromeduck 3y ago
At least for AI there are obvious user stories like fixing auto awesome which is currently quite the travesty, or auto summarizing Gmail or rewriting/rewording in docs/Gmail better than the gunk that exists today.

muro 3y ago

"until ChatGPT became better than Google's version, bulding on the research published by Google and others, yet deciding to not publish their research" would probably be closer than the headline.