Hacker News Reader

Comment by javajosh | Hacker News Reader

There are only three meaningful things you can do with a computer with specs like this: development, design, or gaming. And yet ChromeOS can't do any of these things.

Perhaps this is a vanity product for the wealthy. But wealthy people are going to just surf the web on an iPad.

I really don't get this product.

View on HN

Exactly. ChromeOS is not for such powerful machines, because the OS itself is not powerful. So ChromeOS with a $1000 machine is far from a perfect match (at least it isn't yet, maybe 5-10 years from now).

But right now ChromeOS is and should be for light web machines. And I've said it many times before. There's absolutely no reason for Chromebooks to use x86 chips. I've said Google will almost surely mess up the pricing of Chromebook Pixel because they never seem to understand how to price a new product.

Maybe if they would've made a quad core Cortex A15 machine with this display and for $500, that would've been a little more attractive - although for people just looking for browsing the web, there was no need for such a high resolution anyway. Just keep Chromebooks at $250 and use ARM in them, and get to that 10h battery life.

The Chromebook with the best battery life is the Atom-based Samsung Series 5.

That is because they cheaped-out on the battery in the ARM chromebook, which almost defeats the point of going ARM for me.

I own a Series 5 chromebook. It's hands down the worst machine I've ever owned. I really wanted to like it, too. My issue is that I tried to use it for stuff like reddit, youtube, and whatnot, and because it has no swap space it would just kill background tabs once I had more than 3 or so open. Basically it's a machine built for browsing the web, but because of the way I (and possibly others) use the web it's not useful for that purpose.

I have to put in a good word for the Samsung here. My wife's MacBook finally died, so she got a ChromeBook as a stop-gap, but in reality, it's really all she needs for basic surfing, Google Docs/Drive and watching CBS streaming (for stuff not on Hulu/Roku/NetFlix).

It's cheap, adequately quick, easy to use and has all-day battery life, and charges up at night with the iPad and iPhone used by my toddler.

I'll grant you that if you only do one thing at a time, it works all right. I just tend to jump around a lot -- I use Tree Style Tab for Firefox just to figure out how I got somewhere -- and the Chromebook is a really bad machine for that purpose.

I doubt the Chromebook was ever designed for people like us. We demand far more from our technology and multitask in ways the average user simply doesn't. My Android tablet is nice, but I can't even open up 10+ new tabs at once on a whim and let them all load in the background without it dying.

And forget having them all loaded where I left them when I next return to Chrome.

A friend of mine once put it that if you're the kind of person that thinks a Chromebook is neat, you're not the target market for a Chromebook. I think you and he are both right. The only reason I even attempted it is because I damaged my Series 9 so much that it wouldn't start, so for a few months I just used what I had.

There's absolutely no reason for Chromebooks to use x86 chips.

What about people that prefer trading some battery life for some performance?

The use case that most people cite when they say that they're developing on a chrome machine is to use it as a dumb SSH terminal to their actual development machines. So they're not going to benefit from the higher performance, because all the heavy lifting is being done elsewhere.

It's an i5 processor, 2 cores at 1.8ghz. New generation arm processors, for example the tegra 4, is quad core at 1.9ghz.

Wouldn't it be both more powerful and have more battery life, if it had waited for the tegra 4? Or am I missing something?

Yes you are missing something. Processor performance is not simply a function of number of cores and clock speed. There are vast differences between arm's architecture and x86.

How could we compare then?

Benchmarks.

I don't know what benchmark would give an unbiased comparison.

Build your own benchmark. Measure whatever you want to do. Or if you insist on standardized testing, use something like SPECInt 2006. Results here: http://www.spec.org/cpu2006/results/res2013q1/ (sadly no ARM results). Of course that is only for integer calculations, if you need fp, use SPECfp.

There's nothing that you can really condense down into a couple of simple numbers to use in marketing to non technical people.

Actually comparing Arm versus x86 is really like comparing apples to oranges. One is RISC, other is CISC. One is good with running multiple threads on single core with highly sophisticated branch prediction, other concentrates having many low powered cores each hoarding threads. And this is very, very crude simplification.

Benchmarks tend to be always favor other camp, depending which attributes you are testing.

Actually Intel chips are since long RISC based, with a CISC translation layer. The real difference is in the more advanced pipeline, prediction and cache and and of course the 32 vs 64 bit + other architectural features. Hence the intel chips use a lot more power than an ARM based processor.

This is not exactly true. In most cases intel's chips are more power hungry but the low power atom processors bring better performance than an ARM cortex A9 while still having lower power consumption. (mostly by racing to sleep faster) And this is a three year old architecture i'm talking about.

link for the curious:http://www.anandtech.com/show/6529/busting-the-x86-power-myt...

You are missing something. 2 things, actually.

Cache and instruction set.

Intel is still more powerful than ARM because of its instruction set.

What makes the x86 instruction set more powerful than ARM?

It's CISC rather than RISC, and ARM chips are also designed to use as little power as possible, reducing performance.

I don't see how being CISC rather than RISC results in any inherit advantage. "The CISC approach attempts to minimize the number of instructions per program, sacrificing the number of cycles per instruction. RISC does the opposite, reducing the cycles per instruction at the cost of the number of instructions per program."

> What about people that prefer trading some battery life for some performance?

Question being, performances for doing what? What are you going to do with a chromebook to need the extra oomph?

WebGL? Web Workers? It's astonishing how out of touch the cynics in this post are with modern web technologies.

These aren't exactly things that seem to have taken the web by storm as of yet. Maybe in 3 years.

Modern doesn't mean "in use". I haven't seen much more than proof of concepts and demos when it comes to those technologies. Feel free to point out real-world uses if there are any that could benefit me.

Why pay $1,299 for a thin client OS even if it's running thick client hardware? I can run a browser and so much more in a macbook air or thinkpad in the same price range.

To encourage a world where

a) companies build desirable hardware

b) companies realise having a good screen is a feature

c) companies build screens with more code friendly aspect ratios

> companies build screens with more code friendly aspect ratios

Let's hope Lenovo takes a look at this. I'd love a Thinkpad with a high-res screen, 3:2 display, and of course, the venerable TrackPoint and keyboard. Touchscreen is a nice novelty feature but really isn't required.

I'd rather have 4:3 or even 5:4, but I'll take any steps in the direction of square as positive.

I couldn't agree more. I'm currently on an upgraded Thinkpad T61p with 1680x1050 and I'm reluctant to change it to a newer machine because of the 16:9 screens padded with tall plastic border that are available in most notebooks nowadays.

I really hope that hardware makers will finally realize that the primary reason for buying a notebook is not for watching movies on a 15" screen.

Did you know you can put a 14" standard-aspect T61 motherboard in a 15" T60p chassis with a 1600x1200 IPS panel? Did you know that if you want to spend $400 or so, or you're really good at shopping on ebay, you can upgrade that to 2048x1536?

If tall thin windows are better, why not have a widescreen with two windows side by side?

I don't especially want a single tall, thin window for code. 5:4 is still wider than it is tall. My usual development environment is Emacs with two panes (Emacs calls them windows for historical reasons) side by side on 4:3.

My thinking is that a screen that's close to square works better for a wider range of tasks than one that's very oblong. A rotating oblong screen that could be tall or wide might work nicely for a lot of situations too, but it would be hard to implement well on a laptop.

What I want is as many vertical pixels as possible; this is the first new screen I've seen that's got more than my 10-year-old CRT (1536).

(At this point some wag is going to suggest portrait mode, but I've never seen a laptop with a portrait screen that can be used at the same time as the keyboard).

Apple's 13 and 15 inch retina Macbook Pro models have 1600 and 1800, respectively.

Because there's no space / it would be too small. We're talking about sub-15" display here. So while wide display is great in 22"+, it sucks for everything (except movies maybe) on smaller devices.

> I'll take any steps in the direction of square as positive

If someone proposes a Kickstarter project for a mobile developer workstation with a solid design, 4:3 high-res screen, pointing stick, and great keyboard, I'll be the first to hop on.

Or maybe I should go ahead and make that Kickstarter project a reality instead of waiting for someone to.

a mobile developer workstation with a solid design, 4:3 high-res screen, pointing stick, and great keyboard

The lack of something like this is why I'm still using a 7 year old Thinkpad.

Dell Precision Mobile Workstation M6700? It's got a pointing stick and a decent keyboard, high res 17" screen (although, widescreen), solid construction, etc.

> companies build screens with more code friendly aspect ratios

I hear this complaint a lot and I think it's based on a flawed premise.

Taller aspects might be better on very small screens that only fill tiny portions of your vision but on bigger screens -- the sort you should be using if you're working regularly on a computer -- you want wider not taller. It's the only way to usefully use your field of view. This is why wider aspects exist.

On my 16:9 aspect 27" desktop screen, my vision is practically filled in the vertical direction -- it is easier to glance to the horizontal edges than the vertical ones. I would use more horizontal space before I used more vertical.

And this is the point of wider aspects -- they use your eyes better. If you buy three screens for your desktop, you don't stack them vertically -- you arrange them horizontally so you can actually see them.

People who keep asking for squarer displays for work should simply get bigger displays and stop pretending that a 13" laptop display is ever a good choice for a working display.

Frankly, anything sub 24" is not a working environment. Anything smaller is either (a) something you use occasionally when you're not at your desk or (b) for people who are not working.

May be true for desktops, but if you have ever tried to code on a 1366x768 14" laptop display, you'll agree.

8:5/16:10 or even 3:2 fill the horizontal field of view well while also delivering more vertical space.

Even if 16:9 is theoretically better, why shouldn't we have a choice? Every notebook computer out there besides Apple has a 16:9 display, and I'm sick of it, and many others are as well. At least offer us more than one choice.

I use Visual Studio on a 1360x768 display all the time, and it's not so bad. It does encourage writing the code better because you can't read large chunks at once.

I heard a story where one famous computer scientist (Dijkstra, McCarthy, Hoare, or Knuth?) said something about how using punch cards encourages writing code better because the compile/debug loop is so long and annoying.

> the sort you should be using if you're working regularly on a computer

Could you go into more detail on this? I've been using 12 and 13" screens regularly for my day job for years, and I find the extra physical area of larger screens rather arduous. Basically, if there's some kind of ergonomics concern I'm prepared to change, but not otherwise.

> wider aspects -- they use your eyes better

Seems false, otherwise letters would be typed in landscape. There are other things in my FOV and that's a good thing.

> Seems false, otherwise letters would be typed in landscape. There are other things in my FOV and that's a good thing.

It's true that in reading, we're only good at scanning approximately 8-10 words per line. This is why letters are the width that they are. But I'm not talking about making a single document fill the width of the screen. That document behavior is a relic of a small-screened world.

Modern computer working environments (particularly programming) generally involve opening a few different documents simultaneously and having them open side-by-side. Multiple editors, browser windows, debug tools, documentation, file-trees, etc. This all needs to go somewhere. It is easier to place these separate documents side-by-side than stacked vertically.

> I've been using 12 and 13" screens regularly for my day job for years, and I find the extra physical area of larger screens rather arduous.

Maybe you've only ever opened documents fullscreen or your work involves little more than punching brief commands into a terminal window but if you've ever wanted to edit two documents side-by-side there's little you can do on a screen that small. You're pigeonholing yourself into a mono-tasking environment because you don't have the real-estate to present multiple documents simultaneously.

> Maybe you've only ever opened documents fullscreen or your work involves little more than punching brief commands into a terminal window

No, neither of these things are the case. I'm currently working with Xcode, Blender, terminals and a debug app, as well as my internet distractions, and it's not slow. I do make heavy use of virtual desktops and cmd-tab, and when I was working on a dual-screen workstation I found that most of my work centred on a single screen because I would rather move the app than scan the screen. For me at any rate, a large spatial organisation is visual clutter, and lends itself to hunting for information, whereas the apparent bottleneck of cmd-tab actually sorts information efficiently.

I'm not sure if this is making sense, I'm going down with a cold and focus has gone out of the window in quite a different sense today, but there's my 2c.

This has a far better screen than either of those objectively and a more preferable aspect ratio for many of us subjectively. Installing arbitrary Linux distributions on Chromebooks has been trivial in the past and almost certainly remains so with this one. Many people here would install their Linux distribution of choice on whatever laptop they bought anyway, so the shipping OS is a non-issue.

And Linux would boot at native resolution, making the contents of your screen microscopic and useless. Linux isn't ready for High DPI.

I'd argue that Unity looks much, much better on high DPI. I run it on an old crappy 1024x768 and a newer 1920x1080 of around the same size, and the icons/launcher actually look sane at almost double the DPI of the older screen. Unity makes the low-DPI computer look (and feel) like a toy, but I like it a lot on my new computer.

I'd agree with you if you're running Gnome 2, or KDE, though.

Well the thing is that most people don't prefer Unity and neither do I. I don't really think that there are so many devs using Unity either.

Who is "most people", exactly?

I'm ready. I realise some people aren't, but I'm young with good vision; "microscopic" is exactly what I want. I'm buying the extra pixels so I can fit more on my screen, not so I can fit the same stuff with better-rounded edges.

>And yet ChromeOS can't do any of these things.

Am I just imagining all the posts I've seen on HN from people talking about being extremely happy using Chromebooks as development platforms?

I'll assume that's sarcasm but I haven't personally noticed that trend. The only way I made my cr-48 useful was by flashing the BIOS, adding a bigger SD, and running ubuntu on it.

Aren't most of those folks replacing the os on their machines?

How do you develop on this?

I'm used to using Linux, command line, text editors, local installs of the product I'm developing for, virtual machines, etc...

Are you simply forced into using ALL websites to actually develop and unable to use anything else?

Seems strange to pay $1250+ to develop on a machine that cannot even run or test the code you write.

Action.IO... Cloud9... Koding...

None of which use the $1250+ machine to actually "run or test the code you write." If you are just using a browser for coding, why pay that much money for a machine to run a browser?

One could argue that if a browser is your only environment, that you'll be spending the most time in, then why wouldn't you want the best option for running said browser? The way I see it, if you're going to be working in a completely remote/cloud capable environment, then a powerful, fast, beautiful screened machine to work from makes the most sense.

My current job gave me a Dell Latitude E series machine for work. I work exclusively in AWS. The screen on this laptop is so horrifically bad that I practically get sick looking at it. Compared to my personal (2010 era) macbook pro, or the Retina MBP's I've used at my last two jobs, there is no comparison. I'd much rather have a lightweight powerful little machine that I can ssh from and have some sort of IDE or text editor to work on my AWS projects then worry about a big bulky crappy laptop that is overkill for my work.

Obviously I realize this is MY current use case, however I'd be willing to be that the majority of dev/devops type folks on HN have a similar type of use case. AWS, Heroku, just about any PaaS, Saas, or Iaas, etc... would take advantage of a machine like this beautifully.

Oh and the last time I had an Air (right before upgrading to the rMBP), My use case? Chrome, Chrome, Terminal, RDP, Chrome. That was at a PC Game dev/publisher company where I administrated all windows based desktops, mac laptops as well as all servers of misc variety...

Sooo.....Is this machine useful? I think it is, and I am excited to check it out and may look to getting one.

One more thing, I too am one of the lucky Cr-48 owners and LOVED the concept. Biggest reason I hacked it to load ubuntu? SSH. Had the plugin existed back then? ChromeOS ALL THE WAY BABY!!!

/rant.

Totally dig hosted IDEs. I think it's the way of the future. But why do you need a powerful computer to use them? How does it help? I'm genuinely asking.

The display...and your hosted IDEs don't do all their processing on a remote host. I believe a lot of the compilation/interpretation/services are provided locally via JavaScript.

I spent some time with cloud9 convincing myself it was usable and bought a chromebook only to find that aswell as chromeos being a massive let down c9 ran like a dog on it. Main thing is it was ridiculously slow, the same actions taking noticably long on the chromebook compare to my laptop.

It's interesting to see all the people on here saying that there's no point to having a beefier machine to "just browse the web" and then see your comment which would appear to indicate that there is a need for more performant web-browsing devices.

How much did you pay for your laptop?

Its ancient and these days massively underpowered but i could sit and do sid by side operations in cloud 9 and the laptop would always complete noticably faster. For example i would be debuggin and i would ask to go to hte next step and in the laptop it was immediate, on the chromebook i might wait a second.

Very very odd and very unexpected but it made cloud9 totally unusable for me. Chromeos is also horrible for nearly everything else i do, even reading pdfs, so i cannot think of any reason to shell out for another machine with it installed.

What about them ?

I honestly don't think of them as IDEs more as text editors with syntax highlighting.

There is https://chrome.google.com/webstore/detail/chrome-remote-desk... as well. That might be useful in some cases.

Chrome has a wonderful SSH client [0], and with a $5/mo Digital Ocean [1] (or other [2]) VPS for all your UNIX work, it's a perfectly capable dev machine.

[0]: https://chrome.google.com/webstore/detail/secure-shell/pnhec...

[1]: https://www.digitalocean.com/

[2]: http://www.lowendbox.com/

So you pay $1,300 to log into a $5/month remote box with 20GB of storage and 512MB of RAM?

Well, you probably would not get this chromebook if you were only going to use SSH. You'd buy the $250 one.

Unless you wanted the gorgeous 2560x1700 display... Many people have paid close to the whole computer's price just for the screen (I'm typing this on a 2560x1440 screen that I paid $999 for).

Is your 2560x1440 screen only 13" though?

I don't see the point of having a pixel density that high; almost no one has sharp enough eyesight to distinguish individual pixels at even lower densities than that, and enough people have poor enough eyesight that they won't be able to read text on software that doesn't scale its fonts properly.

That resolution is appropriate for a monitor with twice the diagonal size.

Also, what's up with the 3:2 aspect ratio? That seems awfully odd.

almost no one has sharp enough eyesight to distinguish individual pixels at even lower densities than that

That's exactly the point. Apple markets its high-PPI displays as being higher-resolution than the human retina[0]. It does need some help from software to not cause usability problems like font scaling and tiny images, but we've had the hardware capability to do this for a while. It's past time we improved on this.

Also, what's up with the 3:2 aspect ratio? That seems awfully odd.

If I had to guess, I bet a lot of the people involved in bringing this machine to market wanted 4:3, but someone in marketing and/or a focus group said that would be perceived as old-fashioned, so they got as close as they felt they could get away with. I'd really like to see a writeup on the reasoning from someone involved though.

[0] I believe the accuracy of said claim is disputed.

> That's exactly the point. Apple markets its high-PPI displays as being higher-resolution than the human retina.

But the resolution doesn't need to be any higher than the maximum resolution of a human retina. By definition, you'd be unable to see the difference; you're just left with the scaling problems of an extremely high resolution on a very small display.

But the resolution doesn't need to be any higher than the maximum resolution of a human retina.

As I mentioned, the claim that the resolutions sold are higher than the human retina are disputed. See http://www.pcmag.com/article2/0,2817,2364871,00.asp

Apple's successive retina devices have decreased in pixels per inch, but have increased in pixels per degree at their intended viewing distance as larger high-PPI devices were introduced. Of course, sometimes people use devices at distances other than those the manufacturer intended, and some people have better eyes than others.

The point is, current high-PPI screens haven't actually banished the pixel from human perception just yet. There are still gains to be made, but they aren't nearly as significant as quadrupling pixel counts over previous devices.

This and similar arguments in this thread are eerily similar to the arguments made against the first iPhone by the ./ crowd.

You mean the iPad? I remember the iPhone having a very positive reaction.

As a general rule, most of the time when people say "this is going to be a failure" they are right. The iPad is the exception, not the rule. And Apple has market-making ability that Google can only dream of.

Well, I mean, given that most new product launches and most new business ventures fail, of course the naysayers will be right more often. That doesn't mean they actually have insight into why products fail.

So, that having been said and because this is the internet, I'll rush in with my probably foolish and ill-advised explanation for with this won't be a game changer: In this case, I think the price vs. capabilities will make people compare this unfavorably with other ultrabooks/macbook airs. For the price, there isn't much here that can't be copied, easily, and quickly – and if I really want the cloud connectivity, I can get that on my ultrabook.

What's the unique value proposition in this product? I just don't see it being compelling – except that it's not an Apple product (don't get me wrong, I'm typing this on a Macbook, but I recognize that some people don't like the fruit vendor).

> That doesn't mean they actually have insight into why products fail.

I think it's gut reaction: I wouldn't use this product or recommend it to my friends. I think most of the time, that gut reaction is on point. Sometimes it's wrong, like the iPad, because you don't see the use for it until you try it, but that's the exception.

>Well, I mean, given that most new product launches and most new business ventures fail, of course the naysayers will be right more often. That doesn't mean they actually have insight into why products fail.

Here's a great article on that subject "How to be right 90% of the time, and why I'd rather be wrong": http://paulbuchheit.blogspot.com/2007/03/how-to-be-right-90-...

Maybe you mean the iPod?

Plenty of people already do this with the $250 Chromebook, and even iPad. The $1,299 isn't for the 32GB SSD. It's for the 239ppi 2:3 multitouch Gorilla Glass display. It's for the 4GB of RAM. It's for the dual core Intel i5. It's for the impressively made aluminum chassis. It's for the 1TB of cloud storage. It's for everything a high-end cloud device should be.

If you're logging on to a remote server to do development,you're not taking advantage of the dual core i5 or 4GB of RAM. You're logged into an image that's giving you maybe 1 GHz of CPU and 512MB of RAM. You're buying a $1,300 machine to do development on a $300 machine, over a shitty internet connection.

It depends what you're working on. If you're doing heavy client-side work (e.g. WebGL) then a beefy client is preferable to a beefy server.

I agree though that I probably wouldn't buy this for development work. But I do think that browser's are still better on beefy hardware than on gimped hardware.

> If you're doing heavy client-side work (e.g. WebGL)

Really? If you're the one person in the world doing heavy WebGL this might be the product for you.

Sorry, but if you're doing WebGL development, the last thing you want is a laptop with one of Intel's awful integrated GPUs.

Please qualify your claims with a statement saying that you actually developed on the Intel HD series. Intel graphics are not what they used to be.

But if you also have ~20 tabs open in your browser, and are streaming music etc... I am doing a thin client setup of sorts, and by the end of the day things start slowing down quite a bit.

I do agree that $1300 is too much, at that price range you might as well get a mac. If it was 600-800 i'd probably jump on it. However, I think the curve ball here is the touch screen. I'd be willing to bet that Google's strategy for this laptop is to do a smaller production run, and get these in the hands of tech/programmer types, and then use the lessons learned to develop a model targeted toward a wider audience.

> I'd be willing to bet that Google's strategy for this laptop is to do a smaller production run, and get these in the hands of tech/programmer types

If that's the case, pull an Apple and ship great dev tools on the installation disk. As far as I can tell, there are no top-notch Chrome OS development tools, just various ways to remote into a real machine to do that work.

Chrome Developer Tools, built into Chrome and Chrome OS really are top-notch Chrome OS development tools... because everything runs in Chrome.

They're toys compared to Visual Studio or XCode (or Emacs).

Pretty sure Square disagrees, and they know what they're talking about when it comes to XCode. See PonyDebugger, a tool they wrote to pipe iOS apps' network connections and debug information through Chrome to take advantage of the developer tools: https://github.com/square/PonyDebugger

Now, you'd never write code in the Chrome Developer Tools (except maybe the occasional one-liner in the REPL). But the Chrome Developer Tools are great — best in class even — for profiling, visualizing, and debugging.

Not sure i would call them developer tools since you cant write code in them.

How long will it be before the price of something like this is $650?

But the OS is cloud based, so you don't need the processor power or the RAM. That's the entire point of ChromeOS.

You do for things like WebGL, client-heavy web apps, and other wonderful upcoming web technologies. The point is that this laptop should last you until it physically wears out, not until it becomes obsolete.

Considering the intel integrated GPU, this seems doubtful.

What is the point of all that RAM though if you are using SSH?

I think it might work reverse to this. You don't buy your think client and then figure out how to live with it. Rather, Google is betting that in the future people will be using thin clients and cloud services because of all the other reasons they are attractive.

Once you've already decided you want to live in the cloud, then your question becomes, what is the best possible device I could buy for doing that? And that is what the Pixel is for.

Thin client, thick client, cloud, local, etc, these are all only relevant to technical people. Focus on use-cases. What does the Pixel let you do that a Macbook Pro does not? Have access to all your documents from anywhere (over internet connections of various levels of unreliability?) No, because you can do all that on a Macbook Pro. Does taking away the ability to do things locally improve the user experience in some other way? Arguably, it improves maintainability and makes the UI easier to use.

Is that the target market for the Pixel? People whose main concern is reducing maintenance burden and having an easier to use interface? Okay, now how many of those people are better served by an iPad?

> Arguably, it improves maintainability and makes the UI easier to use

When your entire OS is cloud based, I think you get some synergies - you can reliably sit down at any computer and everything "just works" with all your state exactly how you left when you got up from the previous one.

So yes, you could purchase a Mac Pro, but because it's not cloud based from the ground up there are massive gobs of local state such that you continually need to worry about sync'ing things here there and everywhere, installing apps everywhere, etc. So in world where the default thinking for every application is to store local state, removing the capability for local state is necessary to achieve a true thin client.

> Is that the target market for the Pixel? People whose main concern is reducing maintenance burden and having an easier to use interface?

I think that's not really the selling point of this high end machine. This is Google's statement that cloud based computing is better even for people who do demanding, complex tasks (I'm not saying they are right, but I think it's what they believe).

> Okay, now how many of those people are better served by an iPad?

As I said, you're underestimating Google's ambition for what can be done with a thin client. An iPad is not suitable for highly intensive desktop tasks at all. A chromebook has all the physical features of a high end professional laptop, but is 100% cloud based, and thus can be used as a real computer for real tasks.

> A chromebook has all the physical features of a high end professional laptop, but is 100% cloud based, and thus can be used as a real computer for real tasks.

Except it can't. There is a theoretical possible future in which everyone has 40 mbps unmetered LTE and powerful apps exist in web form, but if anything the trend has been the opposite in recent years (carriers eliminating unlimited data plans, apps moving back to native on tablets and phones).

Yep, I completely agree - I'm hugely sceptical about whether this future where we can all rely on high speed network access 100% of the time will ever arrive. But I don't think Google is - they are betting this is coming. And in the meantime they are trying to provide just enough local state (HTML5 style) to get you by during the outages.

Why wouldn't that future come? Considering Phase I of the GSM specifications was only published in 1990 we've gone pretty far.

This is an aspirational device. They should give a few away in Kansas city. Im sure it would be beautiful with Google fiber.

> Okay, now how many of those people are better served by an iPad?

None of the ones who need to write documents and emails, or edit spreadsheets.

For spreadsheets, it's still less than perfect because Google Spreadsheets are still worse than Excel wrt scripting and managing large datasets. But security, ease of use, and lower maintenance burden might make up for it depending on what your needs are. For my parents, this would be perfect.

These are really cool tools that I didn't realize existed! I think I may take advantage of one of them!

However as the other reply said, it seems silly to buy a big powerful incredible machine just to do all of your actual work on a virtual server. Why not just buy a cheap machine and an expensive server instead of vice versa?

I want good performance and it feels wrong to buy a beefy, sexy machine and gimp it with a 1 core 512MB development bottleneck.

AKA lesson learned: the pixel is not for cost-conscious developers.

If you're doing CPU/GPU intensive web development this would come in handy. WebGL is pretty much a no-go on the ARM based Chromebooks but would perform fine here.

2560 x 1700 display is a big deal for devs, but I think I'd find the extra $200 and get a Macbook pro.

Apple Store has a refurbed MacBook Pro for $1,269:

    Originally released October 2012
    13.3-inch (diagonal) Retina display; 2560-by-1600 (227 dpi)
    8GB of 1600MHz DDR3L SDRAM
    128GB Flash Storage

On a 13" display, those specs are a gimmick, MAYBE useful for examining professional photos.

Not even remotely true. Super high pixel density screens are an unalloyed win. Everything you do on one looks better, most particularly, text rendering, which is absolutely superb on my rMBP. I would expect Google's screen to be comparable.

I spend 12 hours a day looking at screens. Sharpness makes that more pleasant. So does a wide viewing angle, good contrast ratio, accurate color reproduction and adequate brightness. This screen likely does pretty well on all of those factors. Considering how much time programmers spend interfacing with screens and keyboards, it might make sense to spend a couple days' pay on good ones.

Many of us consider this screen's aspect ratio a big improvement over the 16:9 of most PCs for working with code/text as well.

Text sharpness on a retina screen is amazing. You know how lots of people turn of anti-aliasing for code because they don't like the blurriness it introduces? Anti-aliasing on a 250 dpi screen looks as sharp as your 1-pixel thick coding fonts at 72 DPI used to.

I regularly use the SSH client. I wouldn't call it wonderful; it works, but that's about it. Highlight a wrapped line? GNOME Terminal will Do The Right Thing, Chrome's SSH client won't and gives you a newline. See a URL? GNOME Terminal will let you click to open; Chrome's SSH client won't.

I got a cr-48 in november 2010. The new chromeOS remaps ctrl-o to "open window" even when in crosh, so you can't save files in nano or pico. I stopped using it altogether. It's gathering dust.

I'll give you $50 for it!

crosh is being deprecated in favour of a newer ssh client:

https://chrome.google.com/webstore/detail/secure-shell/pnhec...

I'm still using mine (got it December 2010), but I replaced the ChromeOS with a Linux. It actually works great, especially for a free laptop that's over two years old. I agree that the last several updates to ChromeOS made things worse rather than better, though.

I guess you are talking about the $250 chromebook, I am using a lenovo x120, having all my dev setup in digitalocean 1GB server. Do you use X or just fine with command line/emacs? Just curious to know how other people are using ec2/linode/digitalocean for development.

I use the Secure Shell extension to ssh onto a dev environment and use vim from there.

https://chrome.google.com/webstore/detail/secure-shell/pnhec...

So I have to buy another machine to develop and test my stuff on? What's the point of a powerful laptop if it's just a dumb terminal to connect to my actual development machine?

Why not just buy a $300 super cheap laptop and do the exact same thing? Sure I'll have less resolution on the machine, but seeing as the developing is happening on another machine all together, it seems like the performance would be similar.

Well, if you're a developer, presumably this isn't the first computer you have owned. I'm using a 5 year old cheapy HP laptop as my dev server and it works wonderfully.

There are other alternatives, of course, Cloud9 premium account gives you full shell access, Action.IO gives you full shell access as well.

Regarding the idea of using an old laptop as a dev server, have you run into any issues as a result of having the laptop on all the time (such as overheating)?

Also, how would you set it up so that you can log in to your laptop from outside your LAN? I have a few old machines lying around and it would be great if I could put them to use as servers.

Shouldn't have overheating issues as long as you sit it so that the vents can breath. I leave mine in the basement where it's a little bit cooler. My current uptime is 45 days, I only restart for kernel upgrades and never shut it down.

I purchased dynamic DNS, which IIRC is around $35/year, well worth it for the convenience of logging into my home server whereever I am.

DrDreams you are hellbanned. I don't find your comments especially terrible so perhaps you were just unlucky or something, so I'm telling you.

I bet serious developers will buy these things for linux development

I paid $200 for an Acer C7, $80 for 16GB of RAM, $100 for a 128GB SSD, and $50 for larger capacity battery to end up with a portable linux box (running ubuntu 13.04 now) for under $450 that is light, fast to boot, and has reasonable connectivity (wifi, gige, 3x usb ports).

This took me about an hour of reading, 45 minutes to go from sealed boxes to linux, and a small phillips screwdriver (one screw on the back to get to ram+hard drive).

You could put the original Chromebooks in Dev mode and have a terminal with a build environment, or just install your own Linux build on it. Or even Windows. Not the fastest machine in the world, but not the slowest either. I quite liked it. This feels wildly overpriced to me though.

Either after flipping the dev-mode switch and running a chroot or such (thus being outside CrOS's use case), or using it entirely as a thin client (which doesn't need this kind of horsepower).

I really hope that running a chroot or vm becomes a fully supported part of chrome os along with an x server. Working with a crouton ubuntu chroot on my (free) chromebox feels a lot like using OS X circa 10.3 or so and I would consider using it as my full time dev box if autoupdate worked in dev mode. I think i would actually prefer it to modern os x if I could tweak some keyboard shortcuts and run an x sever (even an annoying one like apples) mixed in with aura instead of having to switch back and forth.

People are happy using Chromebooks as development platforms because they're $250 for an ARM laptop with long battery life, and you can hack them to run Ubuntu, not because Chrome OS has any merit on its own.

Things that are cute on a $250 machine are not so cute on a $1,300 machine...

I believe it's much easier to hack this one to run Linux than any ARM-based Chromebook. It certainly has enough memory, storage and processing power to be a decent development machine (provided you avoid heavy IDEs).

It's likely possible to even run Windows on this machine.

It's already braindead easy to put Linux on the ARM Samsung Chromebook. One of the first things I did was to assemble my own personal Linux distro and besides having to copy some blob files out of the original ChromeOS image, it's all very straight-forward stuff and I pulled up about a hundred packages I wanted including an X.org stack without fuss. The compile speeds are adequate if you aren't cross-compiling, I've certainly put up with worse.

I personally have little interest in the Pixel based on the specs. I think X86 is excessively 'big iron' now for a majority of needs and I find the lack of USB3 is mystifying. The screen looks interesting, but it's nothing I actually /need/ and certainly not worth another thousand bucks. I've personally taken to just using X86 for storage/cross-compile servers for the rest of my cheap ARM/MIPS/etc. crap and I've made it a point to stop buying expensive hardware. What $250 buys you now is actually pretty ridiculously awesome.

'brain-dead easy ... assemble my own personal distro' :))

You're talking about developers, no?

* Enable developer mode.[1]

* Drop Crouton[2] onto Chromebook to get a full dev stack and unfortunate Ubuntu/XFCE environment.

* Set up chroot and start building other people's crap.

* Write to SD Card/internal storage and reboot.

Which step here is hard? Tedious to roll your own I'd give you, but you don't even need to as there's stuff like ArchLinuxARM[3] which skips the middle two steps.

[1]: https://sites.google.com/a/chromium.org/dev/chromium-os/deve...

[2]: https://github.com/keyboardsurfer/Crouton

[3]: http://archlinuxarm.org/platforms/armv7/samsung-chromebook

Whoops, serves me right for not double-checking [2]: https://github.com/dnschneid/crouton

Not gonna lie, I just spent several minutes trying to figure out what context sensitive notifications for Android had to do with hacking the chromebook.

Any words on the performance? I feel like my Samsung Chromebook is already pretty busy running five browser tabs at the same time.

I wouldn't be so sure because of the display resolution. Do windows and Linux work good on that?

I have two Chromebooks. One with 3G, that runs the stable ChromeOS, and the other wifi only, runs ChromiumOS, latest ToT, mostly because I'm working on getting ChromiumOS to run on our hardware, so it's a bit easy to compare and contrast when things are working and when they aren't. It's dead simple to boot off an sdcard or USB stick. The Chrome/Chromium did a REALLY good job at making the information available.

I'll probably buy the Pixel as well. The more I use ChromeOS, the more I realize how much time I waste on other crap that doesn't really help me get done what I want to do. And using ChromiumOS, there is the "dev_install" command which will set up a Gentoo chroot in /usr/local, that points to the machine that you built it on, and if you're running the devserver, you can run emerge pkgname, it will shoot it off to the server to build it, and once it's done compiling, installs in the chroot. There is also gmerge which will install ChromiumOS packages for you (e.g. A new version of the browser or maybe a newer kernel, without rebuilding everything) OR you can build a new image, hit up the Help page in settings and it will generate a diff on the server and update the entire install to the latest image. It's really an interesting setup, and I really like it.

Everything goes through gerrit, so you can even help out with version bumps and or patches fairly easily. The documentation is really good, and if you happen to hang out on IRC, the ChromiumOS developers are extremely helpful when something isn't clear, and even update the documentation for clarity.

"And yet ChromeOS can't do any of these things...I really don't get this product."

I completely agree. If I pay $1300 for a laptop, I expect to be able to do some serious graphics work or video editing. Instead Chrome Pixel gives you powerful hardware coupled to an internet-connected OS with basic, limited apps like Google Docs. (As an example, Google Docs is so basic it can't even span table cells. Worperfect 5.0 could do this over 20 years ago. Pay $1300 for an OS so lacking in features or capable software? Is this progress?).

By "specs like this," I'm assuming you just mean the high resolution display. I disagree that a hi-res display is only useful or "meaningful" for development, design, or gaming. It's great for web browsing, which is my main usage for the iPad, and I happily upgraded from iPad 2 to the 3rd generation iPad just for the retina display.

I imagine "specs like this" is referring to the Core i5 rather than the display.

The i5 is hardly a gratuitous CPU. It's a fairly mid-range CPU. The phrasing of the first comment sounds like it's talking about some completely over the top system, like a 12 core Mac Pro.

But we're talking about something used to surf the web. An ARM is still powerful enough and will use way less power.

Web browsing is pretty darn good on the iPad and comparable Android tablets, but it's still a far cry from the performance you can get on x86 hardware. Many Surface Pro reviews mention this specifically. Now, there's still obviously value judgement to be made: is the Pixel worth the extra money, extra weight, and worse battery life compared to an Apple or Android tablet?

Worse battery life compared to an Apple? I think you are wrong. I think Google is being pessimist when estimating the battery life to be 5h when their capacity is 59Wh. For comparison, Apple advertises the MacBook Air 13" battery to be 40% longer (7h) when their capacity is 15% less (50Wh). Maybe Google estimated a higher average power consumption (eg. using the LTE wireless network which is more power consuming than wifi).

  MacBook Air 13": 50Wh - starts at $1200
  Pixel: 59Wh - starts at $1300
  MacBook Pro 13": 74Wh - starts at $1500

Basically, the more you pay, the more capabilities and battery life you get. And both Apple and Google seem in line.

> I think Google is being extremely pessimist when estimating the battery life to be 5h when their capacity is 59Wh.

The battery life depends, not on the watt-hour rating, but on the watt-hour rating divided by the power required. More power, fewer hours. Maybe Google knows something about the actual power requirements that resulted in this estimate.

Just as an example, if the unit requires 11.8 watts, then the battery life will be:

59 watt-hours / 11.8 watts = 5 hours

> For comparison, Apple advertises the MacBook Air 13" battery to be 40% longer (7h) when their capacity is 15% less (50Wh).

Different battery, but also different power requirement.

> Maybe Google calculated it while assuming you were using the LTE wireless network, or streaming HD movies, or something else quite power-hungry.

You're speaking as though all these units have the same power requirements.

I am well aware that the battery life depends on the power required. My point (implied) was that the MacBook Air 13" has roughly the same hardware as the Pixel: same CPU, same GPU, same amount of RAM, same type (DDR3L), etc. The only significantly more power-consuming devices in the Pixel are the display and LTE hardware. So if you don't use too much these (eg. turn brightness down, don't use LTE), then shouldn't the Pixel be expected to consume about the same amount of power as the MacBook?

Let me reply this way -- why would Google sell their unit short? What reason would they have to underestimate battery life?

It occurs to me that, if they exaggerated battery life in normal usage -- typical user brightness settings and typical peripheral activation -- it might create a public relations disaster not unlike the infamous Tesla test drive.

Also, even with the same battery technology, units won't necessarily behave the same as the battery voltage falls. Some will cut out sooner than others as they approach the end of battery charge, a trait that depends on the power supply design.

On general principles, I just don't think they would intentionally undercut their own unit.

Everybody knows there is no industry standard to measure battery life (basically which workload to run when measuring it). For example Anandtech measured the MacBook Air 13" as having a battery life varying between 3 and 7.52 hours depending on the workload: http://www.anandtech.com/print/6063

So my point is that Google would not specifically have the goal of underestimating battery life, but just that it is almost guaranteed that they have different standards from Apple for measuring battery life.

The only way to fairly compare battery life is to have the same person/group run the same test on the 2 laptops. But even that is prone to human-error (eg. what if the tester forgets to disable LTE when testing the Pixel?).

> ... it is almost guaranteed that they have different standards from Apple for measuring battery life.

Yes, most likely, which is why Consumer Reports articles are often such interesting reading.

> The only way to fairly compare battery life is to have the same person/group run the same test on the 2 laptops.

Yes, and (just to make it more scientific) the testers shouldn't be able to tell which laptop they're testing. Not so easy to do, but necessary for an unbiased evaluation.

I just think Google's battery life specification is most likely correct -- they have no good reason to over- or underestimate it. But I agree that an independent test would be useful.

The battery life testing mechanism for Chrome OS is documented here: http://www.chromium.org/chromium-os/testing/power-testing.

Thank you, this confirms what I suspected. Google makes the laptop stream music and play full screen video for some amount of time, whereas Apple makes it only browse wirelessly ("up to X hours wireless web").

The screen and the greater graphics usage related to it is likely to be significant. It may draw much more power for any given brightness than an MBA.

"There are only three meaningful things you can do with a computer with specs like this: development, design, or gaming. And yet ChromeOS can't do any of these things."

If ChromeOS wants to do these things, they need something like Pixel.

You got it the other way around.

Some speculation about one thing a machine like this might be for:

1. Google purchased Nik Software a few months ago. Nik makes several industry leading tools for photo retouching.

2. Before the acquisition, Nik had started transitioning their desktop technology to touch, for use in their tablet app Snapseed.

3. The Nik folk have been kind of quiet since the acquisition.

4. Now Google comes out with a hi-res, touch-oriented laptop, an absolute dream machine for photo editing.

5. ... New Chrome OS native photo editing app?

It's got integrated Intel graphics. That's something I pay attention to because the open source support is great. But I was led to believe that no serious gamer would touch them.

Note also that I could probably develop on it as I use Vim in a terminal, which I believe ChromeOS has.

Yes ChromeOS has a terminal. Just hit Ctrl+Alt+T. The SSH interface is a little non-standard as I recall but not enough to be a problem. It has dynamic proxying support so you can set up a SOCKS proxy through localhost and on to a host in [country] to get around content licensing restrictions. I've done this for Netflix, Hulu, MLB, Fox Soccer, and the London Olympics.

I agree it doesn't make sense. Unless Google announces something in the near future that would make it more than just a web-browsing device...

If that's the case then they should have waited to reveal the Pixel. I'm of the mind that you should buy a device for what it is, not for what it may be in the future.

No. Buy a device for what it is capable of doing in the future.

You buy a computer with faster processor so it can support applications in the future.

The Pixel will be the same a year or so from now. I don't think the hardware is upgradable. But it will be handle resources-intensive apps coming in the future.

Imagine a video-editing app or 3D games.

^ this

> I really don't get this product.

Many people are not rational when they buy products. All you can do with a pair of jeans is wear them. Nice denim, nice rivets, nice pockets, and a nice zip (or buttons) - and then some quality control. And yet people can pay very much money for jeans. (Sort by price, high to low. (http://www.mrporter.com/Shop/Clothing/Jeans?sortBy=price-des...)

You say that wealthy people are just going to surf the web on tablets. This is just a tablet with a keyboard. Some people type enough to make on-screen keyboards painful to use.

If buying a product makes the buyer happy, you can hardly call the purchase "irrational."

A product making a buyer happy is not a condition for whether a purchase is rational or not. If you have a family, then suddenly you decide to buy a ticket to Amsterdam and burn your kids college money on hookers and blow, being 'happy' doesn't suddenly make all of that rational.

Yes, it does. It's selfish, and socially unacceptable, but it's rational.

Rationality (in this sense) is a matter of perspective, and useful only as a tool to help see the rationale behind decisions.

Using that logic I don't know how Apple sells so many of its MacBooks. They're all expensive and have "specs like this", but the vast majority are not used for development, design or gaming.

I can see the appeal of a small (thinner than a MacBook Air), well constructed, blazingly fast laptop. If I could run 1Password on it I would be very tempted to take it for a spin, with the LTE it would make an excellent travel machine.

MBAir is best dev laptop I ever had. Its super light, plenty of power and space for most app development and within reach price wise. Also, I suggest you look up Linus's rant about how he loves Mac book air and SSD for kernel development.

>> MBAir is best dev laptop I ever had.

seconded.

I never suggested it's not a great machine.

I wish the MBAir had 4G/LTE built in. That (and the higher res screen, and the platform security hw) are the things I like about the Pixel.

I'm sure Apple will Retinize the MBAir in 2013 or 2014. I just don't get why LTE chipsets don't appear in any Mac products. I have a VZ MiFi and it's such a pain by comparison -- I'm fine paying $10/mo to Verizon for each mobile/etc. device I have.

> Using that logic I don't know how Apple sells so many of its MacBooks. They're all expensive and have "specs like this", but the vast majority are not used for development, design or gaming.

A macbook is more than a thin client running SSH and a browser.

He's not claiming otherwise, he's claiming that they frequently aren't used for anything more than that.

> I really don't get this product.

Reminds me of what every techie said about the iPad too.

Or the Surface. Or BB10.

Neither of those have shown to be a success.

Nobody does: http://arstechnica.com/gadgets/2013/02/googles-new-touchscre...

Not a single positive comment on the entire front-page, and Ars isn't exactly madly pro-Mac.

I have to admit my feeling of schadenfreude is pretty strong right now...

My theory is that maybe this announcement was intended to make the $1500 Google Glass seem slightly less ridiculous. It's going to take more than a $1300 Chrome book to do that.

In all seriousness though, I was really excited about this when I saw the first video, then it was reported to be a fraud, and now all of a sudden it has an i5 and it costs twice what I thought it would.

At 600-700 bucks, I was ready to trade in my iPad, but for this amount, I'd rather get an XPS dev edition or a 13" retina macbook for just a bit more.

And who wants to smear fingerprints all over his laptop screen?

This! I just can't understand why it would be desired on a laptop.

Maybe it's designed for Google employees. If so, we may not know why they wanted that spec, and the price doesn't matter...

I agree. Dogfooding whilst ensuring their productivity won't suffer. Something to replace all those MacBookPros being used at Google ..

>And yet ChromeOS can't do any of these things.

It does seem weird, yes. My guess is gaming is definitely something it will be able to do (if not now then soon) via Native Client.

I was delighted to see Bastion, a game I played on my Mac, available on the Chrome Web Store using Native Client (https://chrome.google.com/webstore/detail/bastion/oohphhdkah...)

Neither do I.

For an extra $200, you can get twice the RAM, four times the storage, a better battery, and the ability to install whatever OS suits your fancy.

You can always nuke Chrome OS and install a OSX/linux/windows/whatever-rocks-your-boat. You're paying for the hardware, not for Chrome OS.

I agree but the only explanation I can think of is in preparation for Native Client. But why buy now when the price is high, right?

Oh yeah http://tech.slashdot.org/story/13/02/27/038254/pixel-picture...

well.. 64GB limit of internal memory, makes this device not capable for any serious development, design etc. work.

Same thoughts here. It's Intel-based, I wonder whether one could just install a Debian into this beast...

a) Read legal briefs

b) View financial reports

c) Read research papers

d) Write web based apps

e) Write apps so that more actitivities can be done with specs like this.

Before I read your comment I though that computers only good for accounting and scientific simulations. But then I remembered that it's not 1951 any more. Thank you. </bitter-irony>

Reply on news.ycombinator.com