Hacker News Reader

Comment by burtonator | Hacker News Reader

Google needs to own its failures. It can't ship anything new or innovative. The only things that they've been good at is acquiring things that other people have invented (Youtube, Android, etc).

Buzz? dead. Wave? dead. Plus? dead. Allo? dead.

It's so bad now that even features and innovation on their existing platforms are dying.

Youtube? Everything they do with Youtube is a screwup. Android? Good luck getting an app in the app store without Google accidentally killing your company.

Google Cloud? Sorry... everyone's on AWS or Azure.

Search? Nope... Elasticsearch now dominates that space. Think about it. Elasticsearch now beats Google at what Google should have dominated - search.

View on HN

Chrome, Kubernetes, TensorFlow, Golang, AngularJS, Prometheus, gRPC, Protocol Buffers, Vitess, Spanner, Flutter, Bazel, GoogleTest, Istio, V8, AlphaZero

Seems like Google innovates and builds a ton. It's the whole dealing with customers part that they don't like.

Angularjs + Angular doesn't belong in that list.

how they handled that transition gave react a huge footing, splintered the framework community, also have made it very hard to search for their own framework.

Wasn't the bigger problem that it was hard and heavyweight and felt like you needed to be a Java developing software engineer to make anything substantial with AngularJS?

Well, AngularJS shouldn't belong on that list because it was brought in from an acquisition, I believe.

Isn't Golang really Bell Labs guys, now Google.

MapReduce far far predates Google, in shared-memory multiprocessors of the late 80s and thereafter, for instance, though didn't Google popularize it.

Isn't Golang really Bell Labs guys, now Google.

Golang was developed at Google, with Googles blessing, by Google employees. I don't know how much more you can ask for. What difference does it make that some of the people had jobs before Google.

MapReduce far far predates Google

The general concept was known, the implementation and algorithms Google used to run MapReduce efficiently on large cluster of heterogeneous machines was original to Google.

Is it Bazel or Blaze?

Blaze was the internal tool developed at Google. Bazel is the external incarnation.

So they seem to output outstanding software for us to use, but when it comes to making money they are still mostly a one tricky pony, but what a pony....

Google has a significant fraction of the best and brightest developers of our generation on payroll. It’s not surprising that they create innovative technology.

It’s just that they really suck at creating products. After gmail and maps there were next to no successes; even when they throw all their weight behind something like they did with buzz and plus, they tend to fail hard.

> After gmail and maps there were next to no successes

Yeah, just little used products like Android and Chrome /s

Neither chrome nor android have a revenue stream. Both are long time strategic hedges meant to protect the ad revenue from gmail, search and maps.

> Flutter

I am still sad that they acquired a really cool company called Flutter back in 2013 that shipped an awesome product allowing you to control media players using your webcam with hand gestures [0]. They only kept the name it seems. So yeah, not so innovative.

0.https://arstechnica.com/gadgets/2013/10/hands-on-with-google...

Because a name got reused? What does that have to do with innovation?

Because the product didn't get reused.

Flutter is going to die unless they actually release Fuchsia and force everyone to jump ship. Dart got lucky with the AdWords team that is all about it. All the original Dart designers have left Google and are no longer related to the language.

Android team is pushing Kotlin/Native for common code across iOS/Android, came up with Jetpack Compose and rebuilt their reload code from scratch for Android Studio 3.5. When asked about Flutter they always give the "use what is best for you" kind of answer.

Go is still playing catchup with majority of internal services written in Java/C++. Although it already has some small victories in ChromeOS, Fuchsia and Android GPU debugger.

Kubernetes was originally written in Java, was rebooted into Go after another team took over and its codebase suffers from lack of generics support, leading to a pile of code generation.

Some devs dislike so much how AngularJS was so badly managed, that they still mix the Angular/Typescript variant with AngularJS.

Flutter might actually have some legs. It's gaining a lot of steam without any Fuchsia involvement.

I'm a UI-focused engineer who's been working with it for the past 8 months or so, and I'm really liking it. Hot reload is awesome, the framework's primitives and how they compose is very natural feeling after some practice, and the framework doesn't hide low-level control from you if you need it.

Dart's no slouch either. I mean, it's not super exciting, but its standard library is nice, and the language itself is improving all the time (non-nullable by default coming soon!).

PLUS, with support for macOS, Windows and Linux on the horizon, we're soon going to have a really solid development environment that works on both mobile and desktop platforms.

I know this sounds like an ad, but I'm serious. I like it a lot.

Prometheus was developed at SoundCloud? Or is there another Google product with the same name?

Prometheus... now there's a name I've not heard in years.

"App Engine", to non-googlers.

> Google needs to own its failures. It can't ship anything new or innovative.

K8s? BigQuery? Chrome (in particular v8)?

> Wave? dead.

Wasn't Wave effectively a tech preview for collaboration technology central to Docs and Sheets?

Whether or not that's the case, aren't Docs and Sheets counterexamples to your argument?

> Google Cloud? Sorry... everyone's on AWS or Azure.

AWS and Azure have together (as of Q4 2018) a bit less than 50% of the public cloud market, hardly “everyone”. Proportionately, Google is closer to Azure than Azure is to AWS, and has a higher growth rate than any of the other major public clouds, bigger or smaller. [0] AWS has first mover advantage, Azure has leveraging Microsoft’s preexisting deep penetration into enterprise. But Google's gaining on them, anyway.

[0] https://www.canalys.com/newsroom/cloud-market-share-q4-2018-...

  Wasn't Wave effectively a tech
  preview for collaboration
  technology central to Docs and
  Sheets?

According to Wikipedia, Docs launched in March 2006 and Wave in May 2009. And Docs had collaborative editing from the start (Google brought it from a startup with that feature already working)

Right, but Wave was the foundation for the real-time conversations and other collaboration improvements added to Google Docs (etc.) during the period at the end of Wave’s life as an independent project at Google.

https://lifehacker.com/google-docs-updates-with-a-drawing-ed...

But Docs had the ability for multiple people to concurrently edit a document and see one another's changes [1] before Wave was even launched - right?

Hasn't live multi-user editing been there all along?

[1] https://youtu.be/dSGkzDgW1fA?t=78

I don't think it's half as bad as that. Google Flights, Translate, the G-Suite, Firebase and TensorFlow have all been good products/software, though perhaps not world-changing.

Translate has been around as long as I can remember. It works exactly how I remember translate.altavista.com working-except I can’t remember whether alta vista could detect language.

The Google Translate android app has a magic camera that recognizes translates and replaces the text in real time in the video feed. That's pretty cool.

So does Google Translate on iPhone. But that was originally Word Lens, then google brought them and for a time the lens app became free then they added the functionality to the Google Translate app.

There has been a staggering amount of innovation in machine translation. Google has some of the leading groups in research for neural machine translation and they put a lot of it into production.

Saying that it works just like Altavista's Bablefish of yore is ridiculous, at least for the European languages I am familiar with. Translation quality as offered by all major services is leagues ahead compared to the state of, say, three years ago.

(Disclaimer: I work on a competitor to Google Translate)

I meant UI...copy paste text on left side and output on right side. Of course there’s been innovation underneath. My comment wouldn't be two sentences if I had any knowledge about how alta vista or google translate work(ed) underneath.

True enough. When I get unhappy over cancellation of Wave, Buzz, etc., I should remember that a good part of my career in the last four years has been made around TensorFlow and Keras.

I am a huge fan of TensorFlow.js, the examples, docs, etc. are really good. But, I am nervous about jumping in too deep in case it disappears. Regular TensorFlow is here to stay for the long term.

Google Flights was an acquisition of ITA.

Firebase was an acquisition too.

Youtube too

Tensorflow was also an aquihire.

You're conflating Google Brain with Deepmind. Tensorflow was decidedly not an "acquihire".

this is not correct. TF was originated and written by Jeff Dean and some other folks after their experience with distbelief.

wasn't Docs an acquisition too? dunno about Sheets but it wouldn't surprise me at this point

yes, docs came from writely. sheets came from xl2web.

Not the application itself. ITA was mainly acquired for their data, Flights was developed from the ground up.

What about Gmail and Maps?

Anyhow, acquiring the right companies, integrating them successfully and keeping them successful is an equally important skill for a big corp. Look at FB, where would they be without being excellent at acquiring companies (IG, WhatsApp, Oculus).

Edit: Can't spell

Well inbox was a brewth of fresh air but apparently if you can't get a billion users of a billion bucks in a year or two it's just not worth doing.

Interestingly, the Google Wave team was lead by the founders of Where2 - the company they acquired and turned into Google Maps.

https://en.wikipedia.org/wiki/Lars_Rasmussen_(software_devel...

The only way Google actually commits to a product is if it becomes so popular so fast that killing it would rock society as we know it. Gmail, YouTube, Maps, etc.

> Google Cloud? Sorry... everyone's on AWS or Azure.

Google Cloud was late to the party, but they were very smart to invest in GPUs near edge nodes. I think Stadia is primed to eat the gaming world's lunch. They can reuse this capacity for ML during off-peak hours.

I'm actually kind of scared of Stadia. I like having physical copies of console games, and if Google makes major inroads they could become the dominant player in the gaming industry. They definitely won't be friendly to physical form factors.

I have bought a lot of virtual content on Google Play Movies and Books. I am not nervous at all about that content going away.

I imagine it is not quite the same for games which might require maintenance like security updates. Still, old versions of their games should be available.

I'm a huge fan of digitally purchased games. By purchased I mean games with no DRM that don't require a multiplayer server or anything. Humble Bundle and GoG are good sources for these, but you can find them on Steam and such as well.

On the flip side, physical games are no longer truly self contained. Day one patches and license keys and DRM and server shutdowns all prevent you from owning the content on the disk any more than a digital purchase.

Now if only I still had time to play games...

The person you're replying to specifically mentioned console games. Consoles have a (very) capacity (given the size of the average game), there are benefits to having physical copies of the games rather than digital.

Most consoles that allow for downloading games have an option for external storage. My switch has a 256 or 512 gig SD card. Xbox and Playstation allow for USB hard drives.

That's great, but Switch games routinely come in at ~20 GB. That means with 256 GB you can have 12 games at a time, if you don't have anything else such as demos, screenshots, replays, etc. Once you've reached that capacity, it's much better to use physical games rather than having to cycle through deleting and reinstalling.

Thank you for reminding me to double check, it turns out I only have 1.6 gigs left. (And 25 or so on internal storage)

I think I only have a 128 gig in here now that I'm counting. (The menu only tells you free space, not capacity) I have 31 "games" installed, where NES, Genesis, Spyro, Crash, Hotline Miami, and Contra collections each count as 1.

The top three are Spyro, (which isn't even out yet) Smash, and Zelda, which are 14.8 to 15 gigs. I could get rid of Zelda and not really notice tbh.

Next biggest is Doom 3 at 8.8, then a bunch of 6.x games. 16 of them are 1.1 or smaller.

So not quite as dire as you're making it out to be, but in fairness my tastes skew more indie and retro than the average.

> I think Stadia is primed to eat the gaming world's lunch.

Not in the US, with most residential ISPs having arbitrary 1TB data caps.

And not in the rest of the world, where broadband speeds can just barely keep up with a Netflix stream

I've only dabbled a bit in Azure and AWS, so I can't do much of a direct comparison, but GCP really seems to have its shit together in a major way. It has a high degree of polish and usability.

> Android? Good luck getting an app in the app store without Google accidentally killing your company.

Kinda painting with a broad brush about the world's most popular operating system with an 85% mobile marketshare.

>Search? Nope... Elasticsearch now dominates that space.

Again, being really selective here ignoring the virtual monopoly on consumer search.

Android was actually a Google acquisition as well...

More of an acquihire. Even the version they showed in 2007 (two years after the acquisition) bared little resemblance with what it became.

Isn't the change mainly due to the iPhone being released? Before that it was a copy of the blackberry look & feel

Google Photos and Google Maps are underrated in terms of future potential for Google's business

Google Photos by far is my favorite Google app. Whatever team works on that should be in charge of everything at Google.

Elasticsearch does not at all compete with Google Search.

How much does it compete with the (defunct) on-site Google search appliance? Or rather -- how much could it have competed if it hadn't been killed?

Imagine if there still was a private Google search tool for businesses where you could have patched into those APIs for integrating new tools. Everything could have been searchable -- Slack, Email, Docs... everything. Instead, they are all siloed.

Not too much. The GSA died because of some stupid internal Google political battle.

We are moving back towards that. Google Assistant is trying to let you search inside of apps, if I recall correctly

> Imagine if there still was a private Google search tool for businesses where you could have patched into those APIs for integrating new tools. Everything could have been searchable -- Slack, Email, Docs... everything.

Isn't that what https://cloud.google.com/products/search/ claims to do?

The developer community needs to acknowledge Google's successes and contributions.

I seem to remember that (maybe Facebook) had its WAVE development team recruited away.

The Chrome browser and Chromebooks are non-trivial examples of Google's successes.

WAVE was genius and needed time and feedback loops to mature.

I morn early Polymer more than WAVE. Spot on brilliant.

But yes, Google is this generation's Digital Equipment Corp. - engineering brilliance/marketing ineptitude.

Thanks for mentioning Google Buzz. I had forgotten about that.

Google published a sample AppEngine project that was very hackable and used the Buzz data stream. Nice, but now it is gone.

Buzz's effect is still very much alive in production Google code. Gmail specifically bans you from using a label called 'buzz' due to this product bungle. Like 1B+ people plus are loading code today that references Buzz still, it's crazy.

Edit: Screenshot https://imgur.com/a/ERuEwuo

This isn't "google owning" their failures, but this:

https://gcemetery.co

at least memorialises them...

It's funny that there are multiple takes on this: https://killedbygoogle.com/

all the greats of history create a lot of duds. Michelangelo threw away (and didn't finish) 75% of his works. many more examples in history.

Yeah, it's a shame. They were a really exciting company when they were about creating good products and not just about making us the product. Search was a game changer. Gmail was an amazing product compare to the incumbents at the time. It really felt as if product was king, with tech following closely. Now it feels like it's more about MBA type managers that care much more about the bottom line than about the product and customer.

I think that Milton Friedman spiel about giving Shareholder value as the main objective of a manager, as opposed to giving value to customers, has done major damage to humanity, and the capitalist system.

Reply on news.ycombinator.com