Hacker News Reader

Comment by paxys | Hacker News Reader

This is Google's constant struggle in the AI space. They can hire the best talent and give them exorbitant salaries, but by virtue of already being a $2T company their stock price is simply not going to make a 100x jump no matter what they produce. Even if an AI researcher at the company doesn't care about power or independence or anything of that sort, just from a financial perspective it makes a lot more sense for them to join a startup or found their own. Just look at how many of the authors of all the seminal AI papers from Google Brain are still at the company. If you are a top mind in the AI space why would you want to share the spoils of your work equally with 200K other Google employees?

In order to have a fighting chance Google needs to spin off their AI efforts and issue stock independent of $GOOG.

View on HN

Your moneyview is partially correct but this is also a very, VERY deep cultural mindset

Europeans don't take so well to "haha this quarter was bad guess we will lay off 15% of the office"

There is a very, VERY good reason that these are Europeans going from an american company with an office (formerly HQ) in London to a potential European Startup.

If you want to boil it down, this is about power, and Power shall NOT eternally flow from California.

But assuming your hypothesis correct, wouldn't that mean most of those researchers would relocate to EU, which is not the case for most of them?

The article said the new startup would be located in Paris....

These researchers are already based in the EU (Deepmind's Paris office) and are staying in the EU.

>these are Europeans

Firstly, "European" is not a country nor a nationality with a single culture for such broad generalization, as you're trying to make it be. Individuals value different things differently indifferent from where they come from. There's Europeans who would trade vacation and security for more money and impact, and there's Americans who would trade money and impact for more vacation and security. Nationality plays no part in this, only personal choice.

Secondly, how sure are you that all of DeepMind's employees stem from the old continent in order to make such broad generalizations? AFAIK they have members from all over the globe, not just Europe.

>Power shall NOT eternally flow from California.

Please. Don't. I get enough loud-mouth, nationalistic, fist-in-the-air propaganda from the politicians in the mainstream media, it has no place on HN.

Does a European start up have more stability than a big American tech company?

I doubt it. Once money dries up a start up has to close up shop no matter where it is located.

I don't think that's google's problem. Smart talented people are usually more concerned with things like pace of work, culture, freedom to do what they want to do, impact, etc. Working in a large company is absolutely stifling.

Google has all the top talent and was caught flat footed by smaller, nimbler organizations.

Google Brain clearly had the culture over the last decade and a half that resulted in basically 100% of the AI breakthroughs we are seeing today, and all the top minds in the field had no problem working there. What changed in the span of a couple years? Only that the tech became production-ready and the rest of the world took notice, and these researchers and engineers decided that they no longer needed Google's money since there was so much available from everywhere else. This is what directly led to the founding of OpenAI, Anthropic, Cohere, Adept and probably hundreds of other small to mid sized AI startups. There isn't really anything in terms of cultural or organization changes that Google could have done to prevent this.

> There isn't really anything in terms of cultural or organization changes that Google could have done to prevent this.

If you're familiar with the political infighting between Brain, Deepmind, and subteams within Brain, you'd know there was a lot of room for improvement.

Also Google would benefit from being able to use and build off more open source stuff. Tensorflow losing to Torch for researchers was a pretty big blow imo.

Macro cultural trends at Google didn't help either. Stuff like not incentivizing work with longer term payoffs, and too many layers of middle management.

> Only that the tech became production-ready and the rest of the world took notice, and these researchers and engineers decided that they no longer needed Google's money since there was so much available from everyone else.

All that said this is very true and probably unavoidable.

> Tensorflow losing to Torch for researchers was a pretty big blow imo

My perception at that time (autumn 2017) was that Tensorflow had the biggest unforced fumble in the modern history in artificial intelligence, not having the vision to make an API aiming the researchers.

They made the almost impossible and took down Theano and Torch in less than 2 years and fueled with a lot of money from Google got the ecosystem for at least 2 years.

I think they thought like “hey, we won’t the comercial ecosystem, so it’s over; academics and research will follow; even this inflexible and cranky API. If they complain, push Keras to them all”.

What happened in reality was, at the time that Facebook made an absolutely wise decision to lift Torch and unite research and a better UX, it just twisted the situation that for me it’s just bonkers to imagine how Google let that happen?

> me it’s just bonkers to imagine how Google let that happen?

By being blinded by all that sweet ad revenue, meaning everyone's yearly performance review turns out stellar no matter what they do, because so is the company's bottom line.

Temsorflow 2 was their attempt to copy the pytorch experience which was more pythonic.

They also lost the frontend framework mindshare with facebook with Angular vs React.

As a developer i hardly use anything build by Google. On the other hand Microsoft and Facebook have more tools that developer generally like.

Additionally, there were (and still are) many factions and subteams across Brain, Google Research, and Deepmind that now make up today's "Google Deepmind"

There have been significant changes in culture and (at least perceived) job stability in the past two years specifically. See the significant layoffs in early 2023 and the surprising and completely unnecessary layoffs that just happened a week ago. Transparency about decision-making is now basically non-existent. The literal only communication from the CEO was that there will be more layoffs.

Travel budgets are also non-existent now.

All in all, it's still a good job. But it is significantly worse that before, and the trajectory does not seem great.

Brain is more like a research lab, not a consumer products division. I think it's clear that Google had a lot of fundamental and brainpower, but no vision with how to use it other than pure research.

I feel like openAI has made more significant impact touching users' lives than google brain ever did. Even if it's building on google's work.

It's all of those things.

To put it simply, Google is not that attractive anymore. They are the IBM of our generation.

Why doesn't GOOG spin them off?

Now that I think about it, it seems this might be precisely what's occurring. It's possible the approach involves the Scientists "independently" seeking funding, resulting in their Director being sidelined, while the upper echelons of leadership have the opportunity to invest in the start-up. The same leadership that initiated the oppertunity.

Anyway, all hypothetical but this is what I'd do.

>Why doesn't GOOG spin them off?

Normally i wouldn't say this would work but if they marketed it as "British Innovation in the A(G)I age" it could become an election point

Isn't that a huge gamble compared a pretty great and almost guaranteed salary?

> In order to have a fighting chance Google needs to spin off their AI efforts and issue stock independent of $GOOG.

deepmind was kinda independent for many years burning billions of dollars. Now they put them on payroll as other employees.

Reply on news.ycombinator.com