[elgoog1212]

Y'all misunderstand why Google X exists. It exists for three reasons:

1. To obscure the fact that Google is an advertising company, first and foremost, by creating dazzling PR. The more insane the project, the better. Self-flying balloon shooting lasers from the sky? Here's 10 million dollars, go ahead and do it. Over time people lose interest, and those projects disappear into the void.

2. To draw the "best and brightest" in with the promise that they'll work on self driving cars, only to have them repair some obscure dilapidated ad serving backend, where they'll spend years nurturing futile hope that they'll get to work on the cool stuff at some point, and wiping their tears with hundred dollar bills.

3. To keep very senior employees from jumping ship to competitors.

If those three goals are met, actually producing a moonshot every 5 years or so would be gravy, GoogleX will exist even without producing anything at all, indefinitely.

[pjmlp]

> 3. To keep very senior employees from jumping ship to competitors.

If I am not mistaken Bill Gates once mentioned in an interview that in the old days, Microsoft would rather pay someone to sit around than letting them work elsewhere.

[dschiptsov]

And the second part of the joke is that these people usually are not real techies who produce, like Wozniak, Ritchie, Thompson, etc. but socially skilled, manipulative talking heads, who inevitable raise to topmost positions if an organization is old enough - the useless ballast of pointy haired managers. There is no better example of that kind than Balmer himself.

[carterehsmith]

Excellent. Keep those characters around, then unleash them when a competitor is looking for a CEO or something. Sit back, get popcorn, watch a competitor crumble. Laugh maniacally?

It's just like that time when Marissa... nah, I was just kidding :)

[pjmlp]

If you are referring to the whole Nokia episode, don't forget it was the Nokia board itself that placed a big sum of money on Elop's contract if he managed to get someone to buy the company.

[carterehsmith]

I was not thinking Elop, wasn't he more of a Trojan Horse, a smart guy planted to do evil stuff?

And anyways, I was just joking... I even mentioned that :]

[unfunco]

Google, Yahoo, Verizon, Marissa Mayer, all spring to mind.

[snarf]

I was going to make a similar comment on another thread. Google X is mainly about the PR value around its image and recruiting/locking up talent. Google would rather have smart people locked up inside the company working on projects with a high probability of going nowhere rather than having them going to a competitor, or worse, creating the next major competitor. Facebook also engages in some of this, though not to the same extent as Google.

[elgoog1212]

FB is just younger. They have plenty of horseshit projects already, the most recent and public one being Aquila. It's not like anyone is going to allow US-controlled drones stuffed with electronics of unknown provenance to fly in foreign nation airspace "for months at a time". CIA is salivating at the thought. Under the US laws, it's not like FB will be able to politely decline if they were asked to carry a little extra equipment.

[simplehuman]

This is pretty much the reason the research wing of most organizations exist.

[elgoog1212]

I don't disagree, but GoogleX is _not_ Google's research arm. Google Research actually produces very cool stuff that's beneficial to the company, and does so at a pretty rapid clip. GoogleX, on the other hand, takes bold, but not really research-worthy ideas and milks them for PR for years before they quietly die.

[hartator]

Do you know what was the lastest cool thing by google research?

[elgoog1212]

Where do I even begin. :-) State of the art achievements in object recognition, speech recognition, reinforcement learning, natural language parsing, natural language understanding, and so on and so forth, across the board. They employ basically all the prominent names in deep learning, except for Yann LeCun and a handful of others. Their research team is worth its weight in gold.

[gwern]

The very latest cool thing by Google Research or Deepmind? That would be applying reinforcement learning to datacenters to cut cooling bills by 40%.

[infinite8s]

It's mostly CS related. Most of their big data systems were initiated by the Research group.

[1_2__3]

That is not true.

[hn9780470248775]

If these points are true, then why has X been spun off into a separate entity (just "X", under the Alphabet holding company, with increasing separation from Google)?

[elgoog1212]

So that if the public catches up to what it actually is PR damage on Google proper would be contained, and a new division ("Google Y"?) could be spun up from its tattered remains?

[huherto]

4. To keep Sergey Brin entertained.

[patall]

A counter example would be Monsanto which pretty much saved (or at least fully transformed) the company from one of these small basic research groups.

It think this is more a phenomenon that these big companies become more and more like nations. Nations also fund basic research that 99% of the time has no economical benefit but that one time you invent coffeine free coffee and make a lot of money. But sure, nations are driven by the same or similar motivations you describe.