FAQ on Leaving Google

That was an absolutely spot-on description of what it was like when I joined in 2007 or 2008. Within 2 year of joining, I had approval to use all the idle cycles in prod for protein design, folding, and drug discovery, and I had a front-row seat with some of the best programmers in the industry. By and large, employees were fun people to interact with, and the management was generally understanding of our hijinx. The main struggle I had was to convince the leadership to move faster into the cloud ("But we have appengine!" and "But profits aren't as good as ads", until MSFT ate their lunch). As soon as it was possible, I built and launched the cloud product I had wanted Google to make even before I joined!

It really did just feel like grad school with better funding. For me it lasted until around 2014 (wow, 10 years ago) when a director stole my ideas and bad-mouthed me to a bunch of senior folks. I hung on a bit longer (working for a close friend of the author of this FAQ on 3d printing and making stuff) and then a couple stints with ML hardware, before I finally concluded that the company was well on its way to enhittifying everything it did.

Sundar is sort of the complete opposite of this. He wants a large pool of completely anonymous programmers and a small number of directors who know how to turn those programmers into growth products, but those directors don't have a clue. For example, with gChat, one day the head of chat told TGIF that chat was changing, that japanese teen girls were the primary target, and they wanted emojis. He didn't even get that there was this enormous number of professional workers using gmail/gchat/gcal/gdocs and that by fucking up the product, he lost their trust.

Oh well. Sundar is why we can't have nice things.

Grad school...with all the politics to match.

Sundar is not a psychopath. You're making a common error, ascribing humanity to Sundar. Sundar is a growth robot with no moral system. See Bryan Cantrill's description of Larry Ellison: "You need to think of Larry Ellison the way you think of a lawnmower. You don't anthropomorphize your lawnmower, the lawnmower just mows the lawn, you stick your hand in there and it'll chop it off, the end. You don't think 'oh, the lawnmower hates me' -- lawnmower doesn't give a shit about you, lawnmower can't hate you. Don't anthropomorphize the lawnmower. Don't fall into that trap about Oracle."

The difference is that Sundar is a industrial scale trash compactor, not a lawnmower.

Yea but if you have ever met Larry Ellison he’s a total psychopath.

Agree on Sundar.

For some reason I always think of him as Hans Gruber from Die Hard. I find they look very much alike.

https://en.m.wikipedia.org/wiki/Hans_Gruber

https://en.m.wikipedia.org/wiki/Larry_Ellison#/media/File%3A...

Wow, I hadn’t heard that quote before but is is scary accurate and broadly applicable (think ‘psychopathic’ executives and politicians). I often wonder about those type of folks whether some transformation of mind occurs that turns a previously reasonable human being into a ‘lawnmower’, or whether they were born that way and directed their appetites into channels where they could maximise the expression of their ‘lawnmower-mess’.

EDIT: “ broadly applicable”

Something tells me there’s an Ellison story here!

Jesus. Tell me more. You didn’t happen to be involved in deepmind are you? I kind of _loathe_ google these days but find it absolutely mind blowing that there was a time when they were just casually unblocking the scientific community for funsies…to the point where parents could just leverage google’s freebies to maybe shed some light on their kids’ rare disease.

It pains me to think this won’t be happening anymore because really, you can’t sic your brightness engineers on detecting ad blockers _and_ casually make scientific breakthroughs. Something happened to google and the we’ll be writing about for decades to come but if there’s one thing I’m certain it’s that they are done trying to make the world a better place.

“The difference between God and Larry Ellison is that God doesn’t think he’s Larry Ellison.”

(throwaway for obvious reasons)

The comparison to Oracle is pretty good. Working for Sundar's Google feels like working for a company whose only product is quarterly earnings reports. I have no idea what the company's mission is anymore besides Number Go Up. The old descriptions of Google's creative, disruptive, academic culture seem very foreign at this point. Our raw materials are the brains of new Comp Sci graduates, and our product is money.

Exactly, and for someone like myself who hated academia, the internals of Google when I was there (2011-2021) were awful for the same reasons.

It's hard enough to be motivated to work on things in a Big Company. It's even harder when you have to consciously play a game to advertise and promote your success and work -- spending almost as much time doing that as actually doing work. ... and then have others come along and take credit for your work, etc.

Please don't do this here.

One of the things that makes me sad about this is that Sergei and Larry seem so checked out. They were mostly gone when I was there, but I've always gotten the feeling that Google was like grad school because Larry and Sergei wanted it that way.

I get that they've made a ton of money, but it also seems like they really wanted to spend their lives making awesome stuff and doing things like scanning books and making them free. And it feels a bit like the market forces took Google away from them. They put Sundar and a bunch of other McKinsey alums in charge. And McKinsey is, from what I can tell, basically the opposite of grad school.

Whenever I did see Larry or Sergei make an appearance they always looked a little dead inside and like they were just going through the motions.

And from what I can tell, the original sin was taking VC funding. Once they took VC funding, they had limited actual control over what happened to their company. So while they talked in 2004 about not wanting to be a conventional company, and while they warned in 1998 that ad-driven search engines were biased against their users, they still had limited ability to be unconventional in any way that was unattractive to investors. And that includes, in a sense, just being too different. A large company will eventually need to be run by professional management, and professional managers need a thing that looks and drives like a conventional company.

> he lost their trust

What does this mean? Yes, I understand the English, but I mean deeper: What are you trying to say? And, why does it matter?

Who cares about "losing the trust of the users". What matters is that 99% of "professional workers" don't have a choice what mail/chat/calendar/word processor they use. Their IT department decided for them. And, if they do have a choice, what do they use instead?

> Oh well. Sundar is why we can't have nice things.

Most of Google is now mature products. Run it like a business -- maximize profits. It seems logical to me.

Why is this downvoted? It seems relevant to the conversation.

What if they lose the trust of the IT department that chooses which office ecosystem they’re in..?

To put the picture together: So VC money, the DoubleClick merger, and McKinsey ‘culture’ eroded Google (culture)?

> Who cares about "losing the trust of the users". What matters is that 99% of "professional workers" don't have a choice what mail/chat/calendar/word processor they use. Their IT department decided for them. And, if they do have a choice, what do they use instead?

That's true until is isn't. Complacency's impact is subtle, but no company is actually invincible forever.

I've always despised the higher echelons of academia, the top 1%, the Ivy leagues et. al. for a similar reason.

Success in academia comes to those who pick the right people to work with, pick the right things to work on at the right time and say the right things at the right times to the right people, all to push yourself ahead of others guised under veils and veneers of goodness. Truth, morality and the quest for knowledge be damned.

"Picking" is more than what the word suggests. It involves shutting others out, stealing ideas and actual work, propagandizing, giving out freebies but keeping the kickbacks hidden, buttering people for favors, building and fostering inner circles etc. All this is the politics.

No surprise that the ones who are left and thrive are self driven narcissists and ruthless cold blooded creatures painted in playful colors.

Google is the equivalent of the Ivy league. Hopeless, clueless and on a path to irrelevance fostered by a thousand leeches.

Some argue, the world is better because of what Google produced and hence entitled to such inner workings. Same argument as the Ivy's. That's missing the forest for the trees. The real loss isn't what Google or the Ivys have become, but the opportunity loss comparing to what they could have been, with all their resources, had they not gone down this path. This isn't the only possible outcome in this game.

> Success in academia comes to those who pick the right people to work with, pick the right things to work on at the right time and say the right things at the right times to the right people, all to push yourself ahead of others guised under veils and veneers of goodness. Truth, morality and the quest for knowledge be damned.

Do you have a proposal to repair this? It seems any organizational effort is going to end up in a similar situation, because the people who desire to be at the top are the people willing to do the things required to get there, and that leaves little room for people who just want to pursue 'truth, morality and the quest for knowledge'.

It seems to me that the only solution to resolving this problem is to either (a) rely on a benevolent, genius, moral autocrat; (b) completely purge the leadership regularly; or (c) delegate authority to some future un-corruptible intelligence.

Something new will have to be made. You wont get credit for that effort, no riches, not even a thank you. Fooling around with the puzzle is the only reward and it should be good enough even if it amounts to nothing.

What is even the real question? How should we do politics?

> Most of Google is now mature products. Run it like a business -- maximize profits. It seems logical to me.

It is logical if all you want is to extract maximum short-term value from what was already built. To me, the logical conclusion of this path is irrelevance in the long term.

>>One of the things that makes me sad about this is that Sergei and Larry seem so checked out.

I think this one sentence describes everything, companies are all about the people at the top. Its these people that set the culture, pace and overall direction of the company.

If the founders tune out and outsource the very soul of the company to general managers, who can keep lights on rather too well. Well thats what you get. The lights will be on, it will be life as usual and gradual erosion of that very soul that was the company.

You can't blame Larry and Sergei either. There are better things to enjoy when you have billions in the bank and one life to spend it.

Google stopped needing VC money very fast, by the time I joined in 2006 it had long since been irrelevant. Google was funding the VCs by then, not the other way around.

The DoubleClick acquisition wasn't a merger, and had no impact on anything as far as I could tell. It was really acquired for the market share not the people and iirc every DoubleClick employee was reinterviewed, maybe half didn't make it!

I doubt McKinsey had much to do with it either.

IMO the problem was more that the culture of endless hiring disconnected from need eventually caught up with it. I think once Page finally became CEO he may have decided he didn't like it much, especially as with ever greater numbers of restless/bored employees the flow of negative feedback / hate mail got bigger and bigger. People like Pichai are often appointed as CEOs when founders move on, because they will stick to the founders vision and won't make any big changes. Ballmer and Tim Cook are similar, I think, except that Cook seems to have done a better job of keeping things on the original path than the other two did. Typically under such CEOs revenues and profits increase a lot, but there are few bold initiatives or risks taken. It's easy for drift to set in.

Has that changed? AlphaFold is recent and was practically given away to the biotech community. Google is doing a startup based around it as well. So they're still unblocking scientists for funsies, maybe moreso than in the past.

How was Sundar at the time you joined? According to Wikipedia he spearheaded the development of Chrome, GDrive, GMaps, GMail and the VP8 format which are all monumental products so he sounds like he was quite like every other talented hacker that thrived in early Google culture. Is that not the case? What made him turn to the dark side so abruptly?

I think it was Jamie Zawinski who said the reason he left Netscape was that it went from being full of people that wanted to build a great company to being full of people who wanted to work at a great company. The later culture won.

Sundar wasn't a hacker, he was a product manager, and he was very good at identifying growing products, becoming their leader, and riding them to glory. But what really sealed the deal was Sundar's ability to sit in meeting with Larry Page while all the Chief Lieutenants fought, and patiently argue them all down (which Larry never wanted to do).

It's news to me that Sundar had anything to do with gdrive, gmaps, and gmail, except that he was head of Apps for a while, long after those products had completely established themselves.

Interesting, thanks for your reply

Google/Deepmind research still has a fairly large subgroup working on health problems, but most of that work is not given away (or even easily licensable) or published in a way that competitors could reproduce.

What Isomorphic, the spinoff, learned pretty quickly is that protein structure prediction is not, has never been, and is unlikely to be, the most critical blocking step in developing new drugs. That won't stop pharma from investing biobucks (virtual dollars that are gated by milestones) in them. Right now pharma is terrified because their pipelines are drying up, their blockbusters are going generic, and the recent rates of new discoveries leading to new candidates (target diseases/bio pathways) are dropping, even as they invest more and more into automation and machine learning.

(By the way Mike- I always did wish you had been able to run your "math problems" on exacycle, as a way of monetizing idle cores)

No. Alphafold is absolutely amazing and one of those shining beacons of what humanity can accomplish when we optimize for the greater good — or the other bottom line.

What I’m saying is, I don’t think there’s another one project like it in the pipeline…not from the company that’s focusing its resources and talents on throttling browsers with adblockers.

I realize that this is an oversimplification and these changes take decades to become fully realized but you have to admit, it’s a bizarre thing to focus on from the company that literally solved protein folding.

These geniuses were advancing humankind and at some point and now they’re doing “this”.

I don't think it's VC funding as much as being a public company, and being beholden to quarterly earnings. That's the cycle for most "revolutionary" tech companies that end up needing to keep revenue growth going each quarter.

For what it's worth, I don't think it's wrong or bad, but just the way corporations work.

The thing is it doesn't have to be that way. Well, Steve Jobs showed that it didn't have to be that way for Apple. He was able to command authority and weave a narrative that stakeholders could believe in, so bigger long-term outcomes could be pursued. Google was meant to be this kind of company too. So why did Larry and Sergey and other senior Googlers, who had the Steve Jobs example right there to follow, succumb to quarterly earnings servitude?

Ahhh, I wondered about the importance of structure prediction for a long time, ever since reading those blog posts by a PSP researcher in academia who was complaining that big pharma didn't invest in the field. He was asking "why was it Google who solved this problem and not a Pfizer or Bayer?" His posited explanation was a sort of cultural/institutional disinterest in advanced R&D, but that didn't sound very plausible compared to the other possibility that they just didn't care much about the problem to begin with.

> That won't stop pharma from investing biobucks (virtual dollars that are gated by milestones) in them.

I didn't quite follow this, could you maybe elaborate? What is a virtual dollar in this context? Do you mean they make promises of future payments if certain goals are achieved, but the dollars never become real because the goals aren't achievable?

Re: math problems. I remember looking at Exacycle and wondering about that :) but concluded it would have been self defeating, as it'd just have resulted in Google dominating the network and that'd have been seen as an attack.

DeepMind is only possible because of ad money, though. The people who ensure Google has the funds to invest in such projects are a critical part of the team.

As for more projects in the pipeline, well, see the sibling post by dekhn. It's hard to know what's important to work on if you aren't actually on the front lines. DeepMind did AlphaFold because it saw a way to apply its general AI research to biotech, not because it's a biotech firm.