Comment by teleforce | Hacker News Reader

Comment by teleforce | Hacker News Reader

Parent Full thread

Really surprised that Google didn't outright buy OpenAI perhaps the price is already too much now even for Google. Previously, Google has no problem buying direct competitors to its products for example Youtube and Waze.

ChatGPT with Bing search is just very good, intuitive and convenient that I highly recommend it if you have not tried it. Imagine a version of ChatGPT that has tight integration with Google Search, Google Scholar, Google Patent, Google Books, Google Deepmind, etc.

Early last year or so I've read that people were lamenting on how Google at the time was struggling to monetize Deepmind products and suddenly OpenAI ChatGPT 3 came and changed the game forever, while the game changing algorithm is right under their nose, so to speak. It's quite telling that none of the original authors of the Attention paper is still with Google for now.

To showcase the powerful nature of ChatGPT 4 with Bing search I've queried where the original authors of the Attention paper are working now and the answers are as follows:

"As of 2023, here are the current affiliations or recent activities of the original authors of the "Attention is All You Need" paper:

Ashish Vaswani: Co-Founder and Chief Scientist at Adept AI

Noam Shazeer: Co-founder and CEO of Character.AI

Niki Parmar: While there isn't specific information about Niki Parmar's current affiliation, it's known that both Vaswani and Parmar were authors at Google when the paper was published in 2017

Jakob Uszkoreit: CEO and co-founder of Inceptive Nucleics, Inc

Llion Jones: Left Google Japan in July 2023 to launch a startup, Sakana AI, alongside David Ha, the former head of Google's AI research arm in Japan

Aidan N. Gomez: CEO & Co-founder of Cohere, a company focused on Natural Language Processing (NLP)

Lukasz Kaiser: The search for Lukasz Kaiser's current affiliation did not yield relevant results. Further research may be required to find his current affiliation.

It's also noted that many of the co-authors have left Google to start their own ventures or joined other organizations since the publication of the paper."

> Lukasz Kaiser: The search for Lukasz Kaiser's current affiliation did not yield relevant results. Further research may be required to find his current affiliation.

The Lukasz Kaiser who's been at OpenAI since 2021?

https://www.linkedin.com/in/lukaszkaiser

I know it's popular on HN to give Google crap about having bad search results these days and to use ChatGPT instead, but this was literally the first Google search result for me.

IMO it's especially embarrassing that Bing (owned by Microsoft) was unable to find Lukasz Kaiser via LinkedIn (also owned by Microsoft).

That said: native Bing Search gets it right:

    Lukasz Kaiser | LinkedIn

    Connections: 500+
    Followers: 5.3K
    Works For: OpenAI

(obligatory disclaimer that Google pays me money in exchange for work that has nothing to do with AI/ML, so let that inform how you read my post)

Also: Parmar and Vaswani were cofounders of Adept in Nov 2021, but have since left to found a different startup around Nov 2022 per Niki's LinkedIn denoting a "Stealth" startup [0] (and supported by a news article [1]).

I guess my point is that it's really hard to tell which pieces of information are correct vs outdated vs outright hallucinated, without doing the legwork yourself.

[0] https://www.linkedin.com/in/nikiparmar

[1] https://www.theinformation.com/briefings/two-co-founders-of-...

Maybe it only does a certain amount of sub queries? Like a DOS protection?

To be fair. It says research yourself and provided data for most of the others saving the user enormous amount of time.

The genie is out of the bag.

> Early last year or so I've read that people were lamenting on how Google at the time was struggling to monetize Deepmind products

The problem was that all of the ML/AI products would demonetize Google's flagship moneymaker.

Everybody is trying to use ML/AI to undo all the damage the Google monopoly on search has done to the web ecosystem. Microsoft is fine with this for now because Google has the dominant position.

It is pretty clear, however, that this is not a sustainable situation. At some point, people are somehow going to want to turn ML/AI into cash extraction from end users. How to do that isn't obvious.

>Imagine a version of ChatGPT that has tight integration with Google Search, Google Scholar, Google Patent, Google Books, Google Deepmind, etc.

Google wants Google Bard to become that and use all the Google's services.

OpenAI is $80 billion now, you want to buy it outright, that's probably $100 billion. Google does not have that cash, probably only Apple does, and OpenAI would not agree to any 'equity' based purchase.

Google reportedly has $118 billion cash on hand. But makes no sense to blow it on one company. And no way that Microsoft could let it happen.

the more interesting number is $20B in net income in last Q, so $80B is their annual profit.

Microsoft also owns a large percentage of it and would almost certainly prevent any sale

It's also not for sale - OpenAI is still owned by its non profit, and Microsoft only has effectively debt equity (capped profit).

I always assumed Google bought Waze to prevent Facebook from buying it.

Reply on news.ycombinator.com