The Truth About Google X: A Look Behind The Secretive Lab's Closed Doors | Hacker News Reader

The Truth About Google X: A Look Behind The Secretive Lab's Closed Doors

fastcompany.com

163 pointscryptoz113 comments

The original url (http://www.fastcompany.com/3029138/world-changing-ideas/goog...) was just a teaser for this much longer piece, so we changed it. Thanks to primigenus for the tip (https://news.ycombinator.com/item?id=7593039).

HN prefers the most substantive variant of any story. User comments pointing to better urls for a story are helpful—so helpful, in fact, that I'm working on a bit of software to semi-formalize them, so we'll be more likely to see them all. It should be ready soon, and we'll post about it.

Thank you for explicitly calling out mod actions. This is much better than changes blindly happening behind the scenes.

It would be nice if this could also be formalized, maybe with a note at the top or bottom of the page instead of a comment that can be lost.

The trick would be to do so in a way that fits the minimalist spirit of Hacker News. We're loathe to change the site in ways that add clutter. Our bias is in favor of using the existing mechanisms where possible. Often the best ways of doing so are surprisingly if not embarrassingly low-tech.

Unfortunately, the other top comments are about space elevators.

Not many are actually discussing the real article.

Yes, that's a problem whenever you redirect or unify ongoing threads. It's a tradeoff. We'd rather have the best stories on the front page. In most cases, comments eventually grow around the new content.

In this case, though, the space elevator content does come from the real article, so these comments are on topic. The original url was basically self-blogspam—full of links back to the long piece.

Wouldn't be possible (worth) to split a converstion and make it a first page post on it's own, if: -it's not directly related to the thread (hijacking) -a certain cuantity of upvotes are received, -and a minimum converstion lenght is reached? Some times very interesting topics or meaningful comments are lost on moderate or low (to first page standart) interest news.

I can't even find the real article. The link the mods changed to seems to go to a photo gallery...

They hit the same wall everyone else who's ever looked at space elevators hit: the cable materials aren't currently feasible.

What else did Google do around this? Did they go about designing the elevator apparatus itself? Did they devise plans for launching the cable/tether into orbit? Did they explore options for what would anchor the tether in space? Without any further details on this "project", it's as good as a lunchtime spitballing session any one of us might have had recently. One might as well say I've also tried to design a space elevator.

Exactly my thoughts. That nanotubes are the best option, but are still far from feasible from an engineering standpoint, has been publicly well known for some time. But if you're Google you can generate press just from your water cooler talk. For actual design work on space elevators check out http://liftport.com.

Would graphene be a good alternative, especially if it becomes more mature a lot sooner in the mainstream market?

This is just from what I have read on the internet, I am not by any means an expert.

As far as I know, graphene is just as good or a little bit better than currently producible carbon nanotubes. There are a few other materials out there such as colossal carbon tubes and something being called carbyne ( linear acetylenic carbon). Carbyne as far as I know is supposed to have better properties than any of the others but is only a predicted material from computer simulations and some small traces of it supposedly found in interstellar dust. So far, there has been no real success in creating it in a lab.

It is a bit annoying. Arthur C Clark suggested using nano tubes in a space elevator over 20 years ago - so what exactly did google x bring to the discussion?

This comment is a bit annoying. Had they set out to build an elevator, looked at the problem from a different angle, came up with a different solution that had a chance at working we would all be praising them for going after a problem so many people had failed at before.

The fact that they came to the same conclusion (infeasibility) we now question their intentions/motivations and whether it was a good idea or not.

We either applaud attempts or we don't - but we can't choose based on the outcome of those attempts.

There's no evidence they put any serious effort into actually trying to build a space elevator. The article doesn't discuss any alternate cable materials they explored/considered, nor does it even mention any of the many other aspects of designing a functioning space elevator.

If they were even remotely serious, they'd be continuing work and research on everything else relating to building a functioning space elevator and waiting for materials science to catch up, or better yet directly contributing to the advancement of the materials.

Edit: I'm not criticising Google for looking at space elevators. Hell I do applaud that, and I think it'd be a great direction for them to explore. What I'm upset about is Fast Company's journalistic standards about covering it. Google X mentions they looked at space elevators and Fast Company would have you believe they threw millions of dollars and man-hours at it and actually tried to design one.

Journalists apparently have no idea how to cover scientific topics. If someone reports they've found a chemical that's shown to target and destroy a specific strain of cancer cells in a petri dish under laboratory conditions then Fast Company would be reporting that they've cured cancer.

Realistically, what Dan Piponi was probably doing was demonstrating use of http://en.wikipedia.org/wiki/Back-of-the-envelope_calculatio... to rapidly rule out implausible approaches. I doubt they had to spend more than a few minutes looking up materials properties and doing some basic math.

You miss the point, imagine you're in a company that makes billions buying electrons for nanocents and selling them for millicents. Now most of the people in the company are not part of the money making apparatus, so what do they do? They think up new things. But that is hard, it is actually much easier to prove impossible thing is impossible.

To its credit you get self driving cars out of this sort of strategy, and you also get less PR worthy work :-)

Nothing. As they've stated, Google X isn't even working on space elevators. It was just an off-handed comment that someone made, that the Reddit crowd now fixates on.

Probably engineering know how and money.

Why would you build it down here? You go grab an asteroid, move it into far Earth orbit. Eat it to make your cable then get that runners down to earth while spooling one the opposite direction.

So yeah, once we can grab an asteroid or two, chop them up for materials, then I think we can have something like a space elevator. Heaven help us if it ever does break/fall (not snapping of the cable but destroying the anchor point)

I'm not sure why this is catching downvotes; it's probably the most viable idea for how to go about actually manufacturing a space elevator and putting it in place. Granted, we would need to figure out how to manufacture that cable that's 100x stronger than anything we currently have in space instead of on Earth, but that's probably easier than figuring out how to get the entire manufactured cable from Earth's surface into orbit.

Holy shit. Now there's an idea. Forward that one to Elon.

This is what I would try to reach space: I would make a rope ladder system with poles extending at intervals to hold hydrogen balloons. I would space the balloons in a way that they won't affect other balloons or the ladder if they explode.

Contrary to what Red Bull would have you believe, balloons aren't a very effective means of reaching space. A balloon got Felix Baumgartner about 24 miles up, but the cable would have to go out beyond 22,000 miles.

Well you can't just fly 22,000km of cable into space with a rocket, and then drop it down. Using balloons to cover the first few km could save a bunch.

Also, there's a rather large difference between a multi-billion dollar internationally supported scientific operation, and a few million dollar marketing stunt.

If you can make a few hundred meters long carbon nano tube (which we can't right now), what's stopping you from making a balloon with a circumference of a few km, and how much cable could you lift up how high with that?

Since gravity decreases quadratically, the first miles are the ones that count.

Gravity does decrease quadratically, but the origin is the center of the earth, not the cable's anchor. The earth's diameter is over 12000 km, so gravity would not decrease significantly over the 10s of kilometers the balloons supported.

Argh you are right at 160km it's only dropped about 10%. It's a tough world out there :(

Well, to be fair to an otherwise impractical idea, getting up above the dense part of the atmosphere helps.

If you're trying to get to space quickly, then much of the air-resistance losses are in the first 10 km.

Look up "rockoons" for more info on replacing a rocket's first stage with a balloon platform.

22,000 miles might be necessary to escape the gravity well, but something substantially lower would still be of use.

The cable has to be "anchored" beyond the point of geostationary orbit in order for it to pull itself taut and keep itself in orbit, else its own weight would cause it to simply fall back to Earth. Anything less than that would be an extremely tall building, and subject to the same limitations therein.

Point taken. That makes sense.

I believe that's the point. X is no different to your water-cooler chats. Rather than degrade the former, this comparison is meant to upgrade the latter.

There is a class of problems that don't sit nicely in a single domain and can benefit from collaboration. When we engage in spit-balling we are (un)consciously looking to extend the scope of our solution search. From experience such discussions almost never go anywhere beccause that space is really vast and intractable. I imagine a new service here: Google Watercooler, providing tools to extend X into your spitball sessions and vice versa. Like TedX: call it XX.

Just out of curiosity; how would you get the cable up there....

I assume the following:

1. Launch a relatively thin line, with high strength to orbit.

2. Spool it out to earth; attach another, thicker cable to the line and reel it up to orbit

....

N. Step 2 repeated until you have the cable thickness and strength needed.

N+1. Attach a cable climbing robot to that cable to haul up additional lines.

N+2. Attach a larger cable-robot to the multiple cables to climb into orbit and haul a real load.

It's a joke that the media has continued to dig and dig on as a serious consideration.

Google X's Chief of Moonshots (effectively the CEO), Astro Teller, alludes to the space elevator being a fun joke in his SXSW 2013 speech [1], including confirmation of it never being taken seriously by the team.

[1] https://www.youtube.com/watch?v=K8mTHgaLQ7w

Make sure you read the article and interview this was quoted from: http://www.fastcompany.com/3028156/united-states-of-innovati...

Thank you. This article is so much more substantive that I've changed the current post to point to it, even though the space elevator side-piece wasn't bad.

How secretive is something that Google trots whenever they want some goodwill? I think I've seen an "article" about Google X every six months for the last 3 years.

Honestly pretty damn secretive. I worked at Google for the summer and that was the only building on the campus I couldn't go into (I could even go right up to Larry's office door but not to X). My good friend was an intern on X and he never told me a word about what he worked on or who he worked with. That's just an intern. So even though they do reveal some projects for PR once in a while, it's definitely a very secretive part of an otherwise fairly transparent company.

That's crazy. Did your friend mention how he managed to land the internship? Was it a software based role?

He was a student at a design school (I believe it was the Art Center [1] in Pasadena). Incredibly talented at designing cars and also an insane work ethic. 99% sure he was on the self driving car project as an industrial designer but he never confirmed. He said they came to his school to find someone. For a normal job you can go to Google, but from what I've heard Google X comes to you. They know who they want.

1 - http://en.wikipedia.org/wiki/Art_Center_College_of_Design

That might be true for niche fields like designing. For other customary roles like software engineering at Google X. I don't think this is the norm.

Source: My co-worker sitting besides me used to work at X as an engineer.

Same page as the one any secretive organization uses, don't tell anyone anything unless you need a PR boost then trot out some program that meets the need. Has worked for IBM in the past, works for Google, heck it even works for the CIA.

By far the most interesting piece of Google, and likely the most satisfying to work for.

Also good to see people that aren't 20-somethings driving innovation. Hope for us oldsters yet.

The ageism, to me, is the biggest sign that most of Silicon Valley "innovation" is bullshit.

It takes time to get to the forefront of an already-explored area, and to become a good software engineer. People who get there by 25 are extremely rare (maybe 1 in a few million). Some people peak early and some peak late, but the median (across the arts, sciences, mathematics, and technology) is around 45-55 and for technology it should be, if anything, later.

But most Silicon Valley "innovation" is bullshit and doesn't need high-power people. It needs an image of "newness" and that requires youth. It needs people who can exploit the chickenhawk instinct of venture capitalists, and that requires youth as well.

Strong words. I assume you feel the same way about the field of mathematics then -- that it's all "bullshit chicken-hawking?"

http://legendofpi.wordpress.com/2012/03/24/do-mathematicians...

http://articles.latimes.com/2012/jun/25/science/la-sci-sn-nu...

http://www.massey.ac.nz/~rmclachl/overthehill.html

https://www.dpmms.cam.ac.uk/~rdc26/zoop2.pdf

Your third reference spends most of its time supporting Michael's point, in much the same language.

"In fact, Mr. Simonton found that mathematicians make their best research contributions (which he defined as the ones mentioned most often by historians and biographers in reference books) at what many might consider doddering old age: 38.8. That age is very similar to those he found in other sciences: 40.5 in biology, 38.2 in physics, and 38.0 in chemistry."

Gauss is an obvious counterpoint though, he made important discoveries well into his 70's:

https://en.wikipedia.org/wiki/Gauss#Later_years_and_death_.2...

Relevant comic with a list of old scientists: http://www.smbc-comics.com/index.php?db=comics&id=1833#comic

I've read speculation that people peak later in quantitive fields, because a discovery requires a lot of work over many years - gathering and verifying data.

In math, I guess, advances are made by bright ideas - logical jumps, sudden realizations. The discoverer just needs to read to the edge of the field and then realize something no one else has before.

But I think that's a shallow model for inovation. Genuinely innovative companies like Tesla or Skybox mostly seem to be mixed age teams.

> Strong words

This is the same guy who says "No, not everyone who's rich deserves to die": https://news.ycombinator.com/item?id=5625397

I'd agree with michaelochurch, interestingly. Capital has been taken hostage by the rich -- the system we live in is rigged. And that is not a controversial or unsubstantiated statement at all at this point, it's good science, e.g., per Piketty's new findings and analyses, which are scientifically rigorous and heavily data-backed.

> Capital has been taken hostage by the rich

Err... yeah? Previously who did the capital belong to?

There is one thing that always annoys me as of late, once I figured it out, that reporters always represent {A} by {B} and over-sensationalize {B} to the point where I cannot recognize it as normal or I see {A} through a lens unbefitting of it.

Take for example the following line: Google X head Astro Teller embraces failure. Sometimes literally: In group meetings, he has been known to give hugs to people who admit mistakes or defeat. Here embracing failure is over-sensationalized, I think, when in the end it is only a hacker's attempt at trial-and-failure. And then for some reporter to come and say, because they are trying to find gems in the story, that hackers embrace failure sounds just plain old gibberish.

If my boss hugged people in meetings, I think he'd be reported to HR. That's just weird and creepy...

I wonder what situation at Google prompted Teller to make the strategic decision to be featured in this article. Maybe it has to do with the fact that Google X has been somewhat notorious for upsetting shareholders.

> Maybe it has to do with the fact that Google X has been somewhat notorious for upsetting shareholders.

Evidence or sources?

Larry Page gets questioned about it occasionally on their earnings calls. (not in a completely negative way; just on when the projects will pay off, etc.)

It depends on the shareholder, really. Most of the big fish are not interested in holding Google stock for 10 - 20 years waiting for these initiatives to pay off. They operate on annual bonuses, whereas Google X invests in projects that will take more than 1 year to pay off.

Those investors should know what they're getting in to.

From the very start the Google founders have been clear on how their company should operate. They are taking long bets and will not be told to do otherwise. See their IPO letter from 2004:

http://investor.google.com/corporate/2004/ipo-founders-lette...

Isn't the main point of the stock split to give Larry and Sergei the control to not care if their shareholders want short term profits?

Google X is about generating good PR for the company.

That's a nice side-effect. Really it's about finding the next billion dollar business for Google.

No, I'm pretty sure it's about taking over the world. Why generate billion-dollar businesses when you can just own steadily more of the entire infrastructure of civilization as time wears on?

The coolest takeaway is that they didn't really dismiss the idea, just waiting for better manufacturing of currently feasible materials.

Could you weave those nanotubes together, to go from nanotubes less than 1 meter, to something the required length?

I think they'd have to work with strands much longer than the 1 meter ones currently available in order to make a sufficiently resilient "rope". Also, the article didn't mention it, but that 1 meter strand was probably made under strict laboratory conditions after a great deal of trial, and it's not likely able to scale up to manufacturing standards just yet.

Some issues come from phonon effects. These ropes would be a single crystal in effect. From thermal 'noise' you get phonons in the crystal. Since the crystal is a big tube, these phonons can only travel up, down, or create standing waves around the tube (essentially). When these phonons interact, there may be enough energy present to break the crystal/rope. This happens a lot since atoms are real small. Doping the crystal may help, but that'll weaken it. Cooling it may help, but that's costly. Much more basic research is needed.

Just out of curiosity, could we build a space elevator on the moon with current materials?

Yes, with Kevlar -- to L2 or L1: http://en.wikipedia.org/wiki/Lunar_space_elevator

That also raises the same question about Mars (given that its about half the size of Earth).

I think people place way too much emphasis on space elevators. There are so many technologies between current day rockets and futuristic space elevators that would work for dramatically lowering the cost to orbit.

For example this fellow thinks a space gun would be very viable [1]. He's even built them before, and it sounds like he's just one successful kickstarter campaign away from building one. (Maybe someone here could help him set up a successful kick starter?)

[1] https://medium.com/looking-up/b3b2af5bbf71

I think there is a very big difference between putting raw material and sophisticated electronics into orbit. One of them can't survive sudden acceleration to 8k meters per second.

There is some potential, but it would be hard to realize it.

Even if you shoot something on a ballistic trajectory, it must alter its delta-v after the initial launch.

Or it comes right back down again.

IANARS

Yeah, he mentions that in the article. His proposed space gun would basically replace the 1st stage of a rocket. Still a big win IMO.

For anything that can undergo that kind of acceleration, sure. Unfortunately most of the things we need to send into space kinda get squishy when that happens.

I don't know about most. Water and fuel are two big things that should survive a high G trip. I believe satellites and rovers can also be designed to survive such a trip.

Even if it were used for nothing more than creating fuel depots, it would be a pretty big win.

True, there is also the Langton loop, a really crazy but perhaps viable idea.

It seems as though a big part of google x is waiting for academies to do a lot of the hard lifting and then when something has all of the part's completed and researched, putting them all together.

That's the entire domain of engineering, isn't it?

It's as essential to actual deliverable products and inventions as the basic research is.

As long as there is a wealth of knowledge coming out of academic research, what's wrong with that approach? I'm not saying that is what Google X is doing, it sounds like they are pursuing their own academic interests as well, but if all they did was cultivate the research of others and utilize it in ways that have not been fully explored, that seems like a solid strategy. Instead of languishing in academic journals, let's put those ideas to the test.

It did cross my mind that perhaps they should consider longer carbon nanotubes as a Google X project rather than waiting on someone else to do it for them. I imagine there are plenty of areas other than space elevators where they would be useful.

My thought exactly. As an end-goal, space elevators are still science fiction even if you have the materials because the construction would still require a bunch of other technologies.

While better physical materials can be immediately applied to terrestrial problems.

Given how weird Google's hiring practices are, how does one actually go about getting hired at Google X? Would you have to interview for a few thousand positions before you get the right one?

Go to Google's jobs site and look for listings labelled "Special Projects" [1]. These are (often) Google[x] listings.

[1] http://www.google.com/jobs

What part of Google does Titan Aerospace fall into?

Many transfer internally based on internal references and personal interests, and some people are recruited externally for specific expertise or skill set relevant to a project. Perhaps some generalists come from the outside too.

Source: I used to work there, and some of my teammates have transferred there.

Edit: removed trailing typo.

So, the article said they need a material 100x stronger than steel. I'm not strong on materials science, so can anyone tell me: why not use 100 steel cables?

I was imagining assembling the cables into a rope-like structure. After all, rope is an excellent example of how you take relatively weak fibres and assemble them into something much stronger.

One of the core challenges with the space elevator concept is that they need a cable that can reach from the earth's surface to somewhere beyond geostationary orbit, with exactly how far depending on how they want to arrange the counterweight. The midpoint of that cable will have to support the weight of all of the cable that's below it, and that's around 10-15k miles of cable. So we need a material out of which you can make a 15k mile long cable that can hang free, supporting its own weight. Making it thicker doesn't help, because then you just add more weight that you'll have to support. We need a better material that has this level of strength when manufactured at scale.

Has anyone looked into the idea of tapering the cables? It would seem like at the bottom on the cable they wouldn't have to support as much weight as at the middle?

So could you make them thin at the bottom to save weight and then gradually make them thicker at higher altitudes?

It has been looked at and the figures used might even include tapering. One thing is sure though, even with tapering it is not possible today for a earth space elevator.

If I recall correctly, they want tensile strength, which is measured as force per unit area.

The problem with using 100 steel cables is that they have 100x the mass, which means the force per unit area is the same. Hence the need for a new material.

Oh, I see! Thanks for the explanation.

Weight. 100 steel cables weigh 100 times as much as one 100x stronger steel cable (requiring 1000 steel cables to hold up, etc). One of the primary challenges with the elevator is holding the cable up.

It would weigh 100x as much. The key limitation is not collapsing under its own weight.

That sounds like only part of the answer. Wouldn't increasing the mass of the counterweight keep it from collapsing?

Yeah, but then you're back where you started. If you multiply the number of strands/fibers/cables, and you also multiply the anchoring mass to counteract that, you're probably still faced with having a tension greater than the material can resist.

Right, that makes sense now. Thanks!

It's the cable breaking, not just falling down.

Or even more obvious, why not use Superman's hair?

That's just silly.

I mean, how would you cut it from his head?

kryptonite scissors ofc :-)

Does anyone know how the hovercraft would work? I'm pretty curious.

I am pretty sure they are talking about diamagnetics. Bismuth demonstrates this, and graphene also. These are materials that are repelled by a magnetic field.

I'd love for them, you know -- to design radical new ways of visualizing and accessing the wealth of human knowledge, rather than spend their time with geeky pet projects that will never see the light of day.

Google search interface has been fundamentally unchanged since it launched in the 90's.

But you know, glass and self-driving cars and all that shit.

You are assuming that you'd recognize innovation as innovation as innovation, if you saw it. Would you?

Because self-driving cars are a huge deal. It would eliminate the range issue of electrics, because you could send your car to a charging station and get in another. Plus, reduce overall transportation costs. Parking costs in a large city can be $300/month. Self-driving cars plus car sharing will make that obsolete.

Don't forget the 45k+ traffic fatalities (not injuries) per year in the US.

Google Space Elevator: the first modern elevator in history not to have phone support.

>Captain of Moonshots--"moonshots" being his catchall description for audacious innovations that have a slim chance of succeeding but might revolutionize the world if they do.

Moonshots, huh? I'd call them "useless PR stunts".

Stuff that's neither useful and product oriented (like from Apple), nor deep or interesting enough (like Xerox Parc, IBM or AT&T of old) but is guaranteed to buy some expensive positive news coverage and bring some goodwill towards Google for their "innovations".

So a self-driving car would not be useful?

It's crazy that we complain when companies build apps that we don't think is meaningful enough.

But when a company aims to build something trying to achieve a lofty goal, we call it a "useless PR stunts".

>So a self-driving car would not be useful?

Not for 10+ years or more. And at that point it might not even come from Google (other companies also conduct research on that, just less publickly). So the current spectacle and premature announcements are pure PR.

>It's crazy that we complain when companies build apps that we don't think is meaningful enough. But when a company aims to build something trying to achieve a lofty goal, we call it a "useless PR stunts".

How is that "crazy"? It's like saying, "It crazy how we call people to not be overweight, but when they stop eating altogether we call them anorexic". Comdemning one extreme doesn't mean we should applaud the other extreme.

I mean, there's a whole spectrum between "non meaningful crap" and "useless PR research" that companies could be aiming for...

Google does invest serious cash in futuristic technology outside of it main business lines, things like robots, drones and construction software. I don't think they lack "goodwill" for innovations.

“If I had asked people what they wanted, they would have said faster horses.” -- probably by Ford.

No, like actual the actual moonshot. No one really knows that Google X is working on, and this is the first media visit to the lab, so it is hardly just PR stunts.

but dont they already have a different R and D department for "usefull" products ? (mentioned in article)

Google has a Research division. It focuses mostly on problems in computer science, with application to Google businesses. I think they used a term in the article- "atoms, not bits", to describe why X is different: it focuses on physical things, rather than computational ones.

Reply on news.ycombinator.com