Hacker News Reader

I think one of the underestimated functions of this kind of secrecy is to make participants feels important and better about themselves.

"This is super-secret, therefore important. I was picked to work on important stuff, so I must be important".

This goes for everyone in the chain including managers "I am managing a team working on a super-secret project, I can't tell you about". That sounds a lot better than "I am managing a team which makes Uber but for dogs".

I suspect this is partly also why governments like to classify everything, even stuff that doesn't need to be -- it provides a level of importance and adds ceremony to everything. Which, in turns makes everyone involved feel important.

Also, to avoid making them feel worse about themselves when what they're working on doesn't work out, or when their coworkers are skeptics.

My experience was a bit different than ChuckMcM's; I was at Google a couple years later, and my perception was that it was a lot more open internally than any other company I've seen. But when projects were secret or semi-secret, it usually seemed like it was so that the team members could take risks without fear of criticism or ridicule. It's been my observation that for each famous product Google has released, there is some famous person inside Google who tried to kill it. Secrecy is a form of corporate anti-inflammatory, where it's a way of overcoming the natural immune response against any idea that is new and different.

I really like the corporate any-inflammatory idea. I think that really captures it.

I will most definitely steal your last sentence as a word of wisdom in our company. We don't need to make things secret, but I sometimes fear our engineers don't try to add new things on their own because they fear the discussion before they've built it...

I once met a developer who apparently worked on "Uber but for dogs", in the sense of finding dog walkers, groomers, etc. He seemed really humble and open about his work.

"The current road Alphabet is on doesn't seem to me to end in a happy place"

I think it's a natural (good) road.

Alphabet should be a good opportunity for Google owners to invest in outside business with a lot of potencial, like Space X. When you are "just" Google, it's hard from a business view point to invest in outside business. It's more like acquisitions and fusions. But with a strong group like that, you can make more moves with investors money. And Alphabet have a lot.

If your bets isn't giving you good results, let's start put some coins outside. But it's just a teory (don't have actual facts to support).

In finance class we learned it should be investors who make that decision not google. Therefore they could just break it into google and all its other companies so the investor can invest what they chose to. If they want to invest in risky bets they can. If they want to invest in core google assets they can.

The issue is other than ad words their other bets are not nearly as profitable.

I think that a "should" statement is a bit inapplicable. Investors may choose a different stock if they don't like the particulars of Alphabet. It has some significant downsides: founders owning 10% of equity but having a majority of votes; large investments in ventures unlikely to succeed (Fiber, Calico).

"should" from a market efficiency standpoint. By being an all or nothing proposition, it makes it hard for investors to buy in if they are fundamentally against one technology (e.g. user data mining) but want to support another (e.g. self driving cars). It ultimately results in the entire company being less valuable than the sum of its parts (think Yahoo where it's Alibaba holdings alone were worth more than its market cap).

In other words, by having so many diverse bets, investors in Alphabet have to buy into all of those bets. So the number of investors that believe in all of those bets is smaller than the sum of all investors that would invest in each one individually.

I guess nobody ever invests in index funds, then.

The members of an index fund are supposed to have little correlation; not the same goes for Alphabet's divisions.

Quite a few index funds comprise members that have very high correlation with each other, e.g. funds that track sector-specific indexes in health care, natural resources, etc.

An index fund won't suddenly invest in unicorns or whatever else on a whim. They follow a particular plan (e.g. track a sector, track the S&P, whatever).

It all depends on whether investors or Google executives are better allocators of capital.

Google executives certainly think they're better and for the time being it seems investors agree.

>it seems investors agree.

It seems some* investors agree. There is no way to know how much more they could be worth if they were separate entities.

That is what investors like to think. Google management thinks differently (and that is their right for their company).

There is a halo effect around new ventures which helps recruiting. Google got a lot of smart people get really excited building great software to serve advertising. But, if all they do is maintain software they are not going to attract top talent without obscene salary's.

PS: I suspect they probably spent less on self-driving cars than they gained back from that Halo.

Fwiw afaik it would be more tax efficient if Google makes those investments rather than investors reallocating after dividend / cap gains tax.

Moonshot sounds a lot better than tax arbitrage-based VC

Investors do not have control of Alphabet, they don't have voting rights in effect.

Not being tongue in cheek here - actually quite serious: How do you feel about Apple and Braeburn Capital?

Or Berkshire Hathaway, to really nail the point home.

I wonder how many hundreds of billions of dollars all those worthless ponies like Youtube, Maps, Gmail, Android, Chrome, Docs, etc. would fetch if they were sold on the open market.

Not much because none of them are profitable separated from Google.

but aren't docs/mail/etc. somewhat profitable from enterprise offerings?

I attributed it to the somewhat interesting way in which recognition and promotions were handed out, which was to say if someone else got word of your idea/project and executed on it more quickly they could declare victory and take the glory.

Wow, that sounds like a toxic environment, and I would hate to work at a place like that.

I don't think he necessarily meant within the same company. If Apple or Amazon or some startup do it first, then your project is no longer something new, it's following what's been done already.

The context is recognition and promotions. Sounds to me like they meant within the company.

Surely if you have an idea at Apple and Google launches that idea first it will also your own recognition and promotion within Apple?

My wife used to work at Apple. I work at Google now. Compared to the former, secrecy at Google is nothing. Apple takes it to a whole other level.

They even have a Gestapo. I said screw that and never even applied.

http://gizmodo.com/5427058/apple-gestapo-how-apple-hunts-dow...

Very interesting to read your "one trick pony...strong pony" perspective, because it has always been the one constant 'red flag' about watching their growth and expansion into new and cool things. The up-and-down, perfectly natural pattern of experimenting and killing off bits here and there that don't work makes sense for any business. It's just that I don't have an easy time with, um, I guess 'aura' is the right word, of crazy-wild-future-whoa sentiment. I think they can do great things, but growing pains are real in all sorts of businesses, not just the tech sector.

What gives me pause is that if "the smartest engineers on the planet"[1] are working at Google, why aren't we reading about Alphabet Space Ventures, or The Alphabet Social Network, or the Alphabet Smartphone company, or the Alphabet Casual Hotel company, or the Alphabet Movies and TV on Demand company, or the Alphabet Electric Car Company. It goes on of course. And the pat answer is that innovation can happen anywhere. But I don't accept that. Google has tried and failed to make phones (Motorola/Android), tried and failed to make social networks (Orkut/Plus/...), tried to make a video business (Youtube Red), Etc and they have been miserable at it.

It hasn't been for lack of spending money, and it hasn't been for lack of "brand reach" or customer acquisition. It reminds me Sun's inability to make anything other than a Workstation when it needed to fight Microsoft on the engineering desktop. Sun was organizationally incapable of winning that fight, and Scott McNealy was unwilling or unable to see the changes that needed to be made to empower the company to do that. How? I think you have to look to companies that have been able to make profitable businesses out of new technologies or products that were distinctly different than their initial success. I don't see Google doing that effectively, and the people who I think were on track to make the changes that would have done that, people like Chris Urmson, leave the company (as the article points out). How is it the organization expells the very people it needs to retain in order to survive?

There is an apocryphal story of the #1 traveling circus that the author saw as a child, and when he finds it again as an adult it is bankrupt. He asks how that could be and the ringmaster tells him that the star of the show would never let any acts exist that would steal any part of the limelight or glory from his act, and then when the star died and there were no other acts and no reason for people to attend the circus.

[1] From some recruiting hyperbole that they spammed out a while ago on LinkedIn.

Helpful to get more details and I can definitely defer to your description; I just have surface-level observations overall. This part though:

>How? I think you have to look to companies that have been able to make profitable businesses out of new technologies or products that were distinctly different than their initial success.

That's a great concept to keep in mind. I've been a big gear head for most of my life, and only later on did I learn that a company like Mitsubishi or Volvo had additional heavy machinery lines, not just automobiles (I'm using the premise they were 'successful' and economics / leadership decisions / market forces notwithstanding). My thought is that you're touching greatly that a company's culture may have critical importance to being able to do such things.

It very much rings true to me after having a lot of low-level retail & some Fortune XXX gigs, and one of the ways I was able to understand it better was to write up a TV pilot / show idea about an R&D department in a toy company during the go-go 1980s. Exploring a culture of development, albeit with significant stakes & external competition, was actually really pleasant because I could (can still?) see Pros & Cons that, um, well define an overall culture. Pardon if I'm convoluted explaining it but I just really enjoyed exploring the concept as a wannabe inventor.

> tried and failed to make social networks (Orkut/Plus/...)

Google once had the entire Brazilian social network share... And then, they just abandoned it. What I guess just reinforces your story at the last paragraph.

> Google has tried and failed to make phones (Motorola/Android)

But Google tried and succeeded in making a phone operating system that ended up becoming dominant, that gives the rest of their services top billing. Imagine if Apple had kept a near-monopoly on the high end, and/or if the company that leveraged the gap they left at the low end had been Microsoft or BlackBerry.

By breaking the law to do it, yeah. And now they're under investigation for antitrust in almost every major market worldwide. And the product isn't even that good. It's a security-ridden mess[0] where common sense things like security updates aren't a thing because of the compromises they made to get the widespread vendor support.

[0]Source: I carried an Android (or several) as my primary device for seven years.

Disclaimer: I'm an iOS user so I'm no Android fan boy.

The largest suit that has been brought against them for Android has just been settled: http://arstechnica.com/tech-policy/2016/05/google-wins-trial...

While the product might not even be that good, it has clear dominance of the mobile smartphone market in every country other than the US. In the US it's 43.1% market share (which is very strong).

There's no doubt that Google has managed to win this very important market.

The copyright infringement lawsuit isn't what I'm speaking about. They're under investigation for antitrust by a large number of government entities across the globe. Russia just ruled against them. The EU case is proceeding. I believe there are antitrust investigations in India, South Korea, and Brazil. Rumor has it they're back under investigation by the FTC as well, though the FTC tends not to announce such things.

i don't think you read the comment you're replying to...

Every time I am fighting with the Android SDK, and specially with the NDK since they deprecated Eclipse CDT or a new Support Library or Android Studio version comes out, I wonder how many of "the smartest engineers on the planet" do actually work on Android.

I think it's a question of culture. For as may things as people fault Microsoft for they sure do know how to build good tooling.

But yeah the NDK is a mess and the whole environment around it is incredibly painful. Android Studio improved things but they still have a long way to go.

It's hilarious, I tried and failed to build an Android project that I'm pretty sure still built just fine 6 months ago. So much useless churn.

Then they announced the new CMake support and it's basically a CMake toolchain file they copied from an older OpenCV initiative and didn't notice it no longer worked with the new NDK, or anything other than GCC, which they have deprecated. It's insanity.

Yeah and apparently ndk-build support on the stable plugin is only meant for backwards compatibility purposes, but you don't see it described as such anywhere, only a few hints on commit messages.

I had to dig out how it all works with the new cmake plugin from their samples, because the new stable plugin still doesn't manage ndk-build properly.

Also there are quite a few features, like OpenMP, that the clang NDK doesn't support, yet GCC is already deprecated.

I know some of those Android engineers, and I've also done (and loathed) Android development, and I do think they are some of "the smartest engineers on the planet". The only problem is that they're optimizing for their success, not yours (which, if they're smart, is exactly what they would be doing).

At Google, you get promoted for launching something that is technically difficult. So most engineers will seek to implement the most technically difficult feature they have a chance of launching, and then do everything they can to ensure that it actually launches. Nobody gets promoted for not launching things and ensuring API stability. Nobody gets promoted for fixing bugs their managers doesn't know about. Few people get promoted for writing documentation (and if you do, you're probably a techwriter who doesn't call the shots on API design). Nobody gets promoted for doing mundane stuff that might improve the user experience, but isn't technically difficult.

It's the standard big-company modus operandi: hire the best, and incentivize them in ways where it's easy to define the incentives but those incentives don't necessarily add value to the customer. Usually by the time you get to that size, it doesn't matter anyway, since you're working on problems that no startup has the resources to tackle.

Good engineering is not the only factor in a product's success. There are other things that are much more important.

> What gives me pause is that if "the smartest engineers on the planet"[1] are working at Google, why aren't we reading about Alphabet Space Ventures, or The Alphabet Social Network...

IMO, it's because even if they do have the smartest engineers, the best engineering in the world can't fix a marketing problem. That's marketing not in the sense of ads and such, but in the sense of getting the right product into the right market at the right time. Just on the social network side, Facebook has a near-unbeatable network effect advantage, and Google+/Alphabet Social Network can never overcome that by making a better version of the same thing, no matter how awesome it is from a technical standpoint.

That's not to say that Facebook is invincible or anything - Facebook will go down some day, but it won't be because somebody else built a carbon-copy with a few more features and took their place.

> Just on the social network side, Facebook has a near-unbeatable network effect advantage, and Google+/Alphabet Social Network can never overcome that by making a better version of the same thing, no matter how awesome it is from a technical standpoint.

That speaks to a lack of innovation though. Obviously Google shouldn't be trying to build a carbon copy—but they seem to be incapable of building anything besides that.

Plenty of startups have arisen to challenge Facebook (Instagram, Snapchat, WhatsApp). Why didn't Google build any of them? Especially when so many startups are founded by ex-Google engineers.

> Why didn't Google build any of them?

Or buy WhatsApp when they had the chance. That non-purchase is as perplexing an Yahoo's purchase of Tumblr.

I think part of the problem was the unclear and frankly, half-assed way it distributed itself. Some of Google's subsidiaries became Alphabet holdings, but very little of Google broke up into separate companies. When it was first announced many people were thinking "oh, okay, S is for Search, Y it for Youtube, X is for X-Labs, M is for Maps, okay maybe each of those should be separate concerns".

But instead it's just Google, and a bunch of highly experimental, low return, high risk other random companies with no particular strategy. How many labs are separate Alphabet companies? Why not just 1 lab that's supported by R&D concerns from the rest of the conglomerate? What isn't L for "Labs"? Does GV really need to be a separate company? Why isn't it just part of Google? Or why isn't GV "V for Venture?" That's okay, they also have Jigsaw, Google Capital and probably a bunch of internal investment arms.

Do they really need two genomics/life sciences companies?

Why are there so many different video companies, but youtube stays part of Google?

The approach is entirely disorganized and without any seeming strategy.

> Do they really need two genomics/life sciences companies?

Presumably this is a reference to Verily vs. Calico. AFAICT, yes; or at least the two have distinct enough missions that it makes sense for them to be separate efforts. Calico is a very long payoff window effort starting with basic research aimed at lifespan enhancements, Verily is a shorter payoff timeline (but still fairly long window, given medical device, etc., commercialization timelines) medical device and treatment development effort.