For engineers only. No software support beyond ensuring the BIOS and the drivers work and are updated.
I hate Dell, but I got a Dell XPS 15 because it was the only laptop I knew would: boot Linux reliably; last two Ubuntu LTS upgrades (4 years); have a high spec (i9 etc); and was reasonably priced (Apple tax was 50% higher for worse hardware).
Is it impossible to make a laptop better than Dell or Apple? Google: how much goodwill would you receive through developer love? Why do I have to pay for Windows when the OS division seems to be actively working against making my work easier? Google, you have the skills and size to make a secure laptop.
Google, maybe it is because most of your engineers internally just need a good terminal, instead of a fast laptop?
As for the so-called "Apple tax", I just don't buy into this. I mean I miss the glory days of the 2011-2012 Macbook Airs. These were so cheaply priced they destroyed the competition because no one could produce a similar spec laptop at that price point. But then the numbers men decided dropping the average ASP for Macbooks was "bad" so they had to invent new features nobody needed or wanted to drive up the cost (Touchbar being the poster child) but for the hardware you get, it's actually pretty good. Or it was.
> Is it impossible to make a laptop better than Dell or Apple?
It's clearly not trivial.
> Google: how much goodwill would you receive through developer love?
Based on the bubble that is HN the answer is clearly "absolutely none". Just look at Chrome adding DDG as a search engine option. Where in that thread do you see people saying "this is great"? No, the top comments are littered with theories about how this is just an effort to avoid antitrust or other such ulterior motives.
I guarantee you that if Google produced the laptop that you and 22 other people would buy you'd hear nothing but complaints about it.
Counterpoint: I know plenty of happy developers that own multiple Nexus and Pixel products, without bagging them. Engineers want good hardware and many are willing to pay for quality.
> As for the so-called "Apple tax"
For as close to equivalent as I could get (specification, reliability, etc) the equivalent Apple cost 50% more than the XPS 15. Paying 50% extra for an OS I don't need makes no sense for my situation. Paying for an unused Windows license for the XPS was unavoidable (I have our own licenses for the Windows VMs we need for testing).
> It's clearly not trivial [to make a laptop better than Dell or Apple]
The Google hardware I have experienced has mostly impressed me (even if third party, they have mostly enforced quality constraints). Although I have owned a Nexus with a design flaw, I would consider buying a Pixel (although at present I find the Nokia gives me better bang for my hard earned buck).
But that's nothing compared to the new stuff. It runs containerized linux with a great abstraction layer that connects GUI calls from linux into the beautiful hi-dpi Chrome OS desktop environment (I think it's still called "Aurora"?)
It's still in beta, but they're almost ready to add: - USB support (obviously important here, but they want you to be able to develop android apps on it) - Sound from linux container -> hardware speakers (I believe this is in beta channel now though?) - graphic acceleration (also is in early testing, canary branch maybe?) - backup/restore functionality (lose your laptop? get a new one and restore your container)
All of this in one of the sexiest laptops with the best keyboard of any of the ~5 machines I own for $999
I've been using it as my main development machine (a complicated electronJS app that windows substate on linux choked on) and all the backend/database work for over a year and I love it. You could get it on sale for $699, too.
Main complaint? If you need to do a lot of work in VM's. You currently can't spin up a Windows VM or anything very easily, but rumor has it they're working on letting you dualboot in the future.
I freaking love this machine. They should pay me to evangelize but I promise they don't :)
Prior to this, I gave my mom Chromebooks after she kept calling me for tech support for virus infestations, and it worked well.
Working with the Pixelbook, I totally agree with the VM bit. I would love to run Vagrant, MiniKube, KubeFlow, etc. The Linux emulation is good, but It's Google. 'We' can do it better. They probably did, but missed the adoption / backwards compatibility curve.
I hung out with a ChromeOS dude last weekend and he told me about how the Linux emulation is controlled by 'vsh' using their own replacement for QEMU, and inside that is an LXC container. Great for security, and I believe him. I entered one of the system vm's and saw LXC/LXD. The security seems great, but again, I can't do what I want to do. I felt dumb asking him about my 'emacs' in the ChromeOS Linux emulation. It always had a huge title bar on the top that took a bunch of the little screen real estate available. His answer was 'hit the full screen' button. I felt dumb. Usability for us dumb engineers will sell your product. Apple has abandoned the engineer market. The XPS flexible, but the Pixelbook is still sexy. I can't get my son to take my daughter's Pixelbook though, so maybe I am just old.
I would love a Pixelbook, but I need a laptop if I want to work while travelling - I often end up with poor cellular data (or none at all!).
Does having both a laptop and a pixelbook make sense?
Work also bought a device to get satellite data - so that we can do operational support while in "remote" areas in New Zealand (plenty of areas we go to that lack mobile coverage) or when travelling overseas.
1) The docking station is horribly buggy. If I unplug it to take my laptop somewhere and then plug it back in, I have to do a full reboot or none of the USB peripherals on the docking station get power. Monitors, power, ethernet all still work through the docking station after being plugged back in. Dell support took 4hrs of my time installing/uninstalling drivers and firmware updates before finally giving up and sending me a new one. Which suffers the exact same issue. So I can buy a different brand of docking station, not take my laptop anywhere, or constantly reboot my machine.
2) Rebooting takes forever then fails. Every single time. After 8-10 minutes of sitting on the blue "Rebooting" screen it finally crashes and tells me something went wrong while rebooting. Every single time.
3) Sleep still doesn't work about a third of the time. If I put my machine to sleep through the Windows start menu and wait for all the various indicators to power off before closing the lid, there's a good chance that when I boot up tomorrow, the battery has fully drained and the machine has to boot back with all my applications closed down and context lost.
It makes me miss my Thinkpad.
Run Manjaro with gnome-shell (actually gnome-shell-performance until 3.32 is released). Its lightening quick, reasonably stable, good battery life.
I also tried KDE, which works quite well, but the battery life was quite poor and I find the general KDE experience abrupt.
Now my 2017 mbp just gathers dust, never going back.
I'm seeing $2182 Thinkpad vs $2599 MacBook Pro--choices driven by WQHD display, 16GB of RAM and 1TB SSD. Processor on the Pro is significantly faster, the display on the Mac has slightly higher resolution. Maybe not $400 worth of difference, but certainly almost $200 or so.
And I still can't buy an i5 laptop with 32GB of RAM.
An old laptop will run any Linux distro flawlessly and be equally snappy (as in, low-latency disk reads, fast boot times, many applications open) as a current-gen one if you put in a fast SSD and max out the RAM.
I know that this topic is controversial, but it strikes me as odd how many people seem to equate software engineering with a monster laptop, when many of us get lots of work done on smaller, less beefy client devices, offloading the real heavy lifting to a machine that can handle it.
For the $1000 it costs for a faster laptop, work gets many thousands of dollars back due to more efficient development.
I could waste time optimising disk space, waiting for compiles (the developer tools we use need single-core CPU performance), enchancing the performance of development tools, waiting for VMs to boot, waiting for Windows VM to update, waste time shutting down that VM I am not currently using. Yes, I could waste time switching to other tasks while I am waiting for whatever it is I am waiting for.
I travel, and while travelling work gets some development time out of me because it is a laptop and not a workstation. A terminal is no use when I am in an area with poor cellular data connectivity (e.g. most planes).
Work definitely wins on this whole equation.
I built my first desktop computer to do this recently. I can definitely recommend it, even for local server development. It's nice to be able to have a version of a codebase running on the remote machine so my local laptop stays cool while I'm testing.
* I once started a set of experiments, drove off for a short 3 day vacation, and returned to see everything was still running smoothly. Which frankly I hadn't expected.
(I wrote another comment for why I use a high-spec laptop: basically multiple VM's and greedy development tools).
I don't know about impossible, but they've both been building laptops for a long time and they have large market shares, so if it was easy to build better laptops than theirs someone probably would have done it by now so it's probably really hard.
Thinkpad Carbon X1 6th gen (or 7th now?)
I only wish it was sold without Windows so I can save $30-40 on the win license.
Other than that, freaking perfect laptop that rivals a macbook (I will admit that the mac trackpad is slightly better... but this one is pretty damn good)
I needed a laptop (I try to refresh every 4 years in time with Ubuntu LTS releases), and the Dell met my requirements at the time, so that was what I got.
I looked at getting an Macbook Pro, but I was worried about the keyboard issues (I do use the laptop keyboard), and I had read about Linux driver issues.
You can get the Precision 15" with Linux. However I was quoted a price 2 times as much as the XPS 15 with a lower spec, less community support, and I was quoted it with 16.04 installed in Jan 2019 (presumably only 16.04 officially supported? Hard to get information when we are only a small business).
https://liliputing.com/2018/05/dell-introduces-2018-precisio...
Don't Chrome books provide you what you need for software engineering?
For one thing, Linux-based OS's like Debian GNU/Linux absolutely follow the "it just works" mantra - if the OS installs properly in the first place and does what you need it to, you can expect it to be rock solid with few or no future issues. Plus, they tend to work for a lot more time than the 5 years max that Chrome OS does (or Mac OS, for that matter). That's simply invaluable in this day and age.
I have two still serviceable MacOS laptops but I don’t think I will buy any more. But, I love Apple Watch and iPhone - probably be a customer of those products for a long time.
For me, wouldn't it be great if laptop shells had a universal connector you could plug the mainboard of choice via an edge connector, be that ARM, X86, MIPS,????. Certainly make laptops more upgradable and save upon production costs with a standard laptop mainboard form factor.
Admittedly it's about 6 years since I interviewed there, but when I did, there were a number of engineers in my loops who were using chromebooks.
They found it more than sufficient for their needs.... I can see that, I guess. I'm assuming the bulk of the "heavy lifting" happened outside of the context of the chromebook.
There are some Engineers who use a Chromebook, but most have a Macbook. I think Linux is more common than Chromebooks.
But honestly, why would you choose a chromebook if you can also get a macbook?
What does 'last' mean? Any laptop should 'last' 8+ years.
> have a high spec (i9 etc);
I really don't see the point of high-spec laptops for development use. Higher end, high power components just make the batter life worse - reducing the entire point of a mobile device. And you're paying a premium for the mobile version.
A much better option is to have a beefy linux desktop that you perform work on remotely. There is very little reason to locally compile code on a laptop. The economics work much better IMHE... If you have a $2000 budget for instance, you can build a $1300 desktop and buy a $700 laptop which gives you much more power and flexability than buying a single $2000 laptop. So yes, I do think many Google engineers just need a portable terminal to connect to beefy VM's remotely. If you are often working on a mobile internet connection, Mosh works very well for high latency connections.
The $2000 laptop makes sense if you want to game while traveling and/or are using Windows where offloading work to a remote machine is a bit more clunky than just using X11 forwarding and/or tmux over SSH.
The 8950HK processor is fast[1] and NVMe is way faster than SATA.
I just use the integrated Intel GPU (I don't play games on my work laptop. If I were to play games, I would use a dedicated gaming rig or console).
I am expecting that in 4 years time I can get equivalent improvements to productivity by getting a replacement laptop (or even better, retire to a pacific island!).
[1] Benchmarks are not the answer, but anyway: https://browser.geekbench.com/processors/2145#family-64-sing...
Personally, I can't live without focus-follows-mouse. That way, I can type and click on a window that is under another one.
Its too bad, since the Pixelbook is pretty nice and with actual investment it could have been something much more than it is, but alas it would not have the margins search advertising does. Back in the 90's I met the head of Xerox's PARC after he had given a talk on innovation. One of the challenges he saw was that gardening innovation was like gardening crops, which was to say it was very hard to know when sprouts had one leaf out of the ground which were weeds and which were crops. Ideas were the same way, you need to nurture them to at least adolescence before you can reasonably decide if they are going to go on to be great or not so great.
From the outside, Google appears to be at that stage where they have not learned how to nurture an idea to see if it will be great or not before killing it.
They should at least tell us what those projects are so someone can pick up the scraps.
I passed up the opportunity to buy the new Pixelbook as the original Pixel still works nicely and I don't need Android on it particularly.
However, I think that Google have gone wrong with ditching the Pixelbook. There is a definite market for a decent developer machine that you can run Linuxii on that isn't a hack. Homebrew, Virtualbox, Ubuntu in Windows isn't ever native and the instructions are always less straightforward than with a stock Ubuntu or other regular distro.
From recent news there seems to be a small army of pissed off Mac users that want simple things like an escape key. There are also plenty of developers that don't really need Windows. The Dell developer XPS specials are very nice but the original Pixel and its descendants get a few things right, the 'no expense spared' 3x2 screen, nicest keyboard ever, best speakers ever and best trackpad going works for me.
Google should accept that their take on a 'developer machine' is going to be low volume but there is no reason why they can't make it pay its own way, really it is a matter of specification and not nickel-and-diming things like the keyboard.
If anything they should make the machine exclusive. Anyone can have a posh Apple computer but it is still just an Apple to people like the clients that I met today. Break out the Pixel though and even if you are just scrolling through a web page it has some impressive statement value to people who know nothing about computers.
I think the hardware division went the wrong way trying to make the Pixelbook 'for everyone' (rich). If they made it super cool with the ability to run a linux distro native as well as having ChromeOS/Android then they could have made something that Pixel laptop owners like myself would buy. Instead I run a regular PC for development.
I also think that if their developer machine was the de-facto thing for Google to use in house then that would make the machine aspirational to developers. If one were to be using the same hardware that Google designed for their developers then you would feel that you had the best going for Android/web development. A Dell machine doesn't quite have that kudos because nobody at Dell writes a lot of code.
I also have a 2011 Macbook Air, and am still able to get OS and security updates.
https://www.aboutchromebooks.com/news/chrome-os-69-stable-re...
https://chromium.googlesource.com/chromiumos/docs/+/master/c...
[0] https://arstechnica.com/gadgets/2019/03/all-signs-point-to-a...
I'm also the owner of a Google Nexus Player - remember that device? I liked the idea of an Android TV / console for Android games. I bought 4 gamepads they were selling for it.
It was buggy. It was rarely updated. They stopped updates after about a year. Then it died unexpectedly - it didn't even last 2 years. Now I'm left with 4 bluetooth gamepads and a dead Nexus Player in a box in the basement.
It saddens me to say this, but I would never buy a device from Google again unless it's the second+ iteration in a successful line of products, and even then I would have to factor the chance that they'll discontinue it at any time.
Literally nothing outside of Google's core services is trustworthy. Even their mail isn't entirely trustworthy! I've come to rely on Inbox and it's being killed for an inferior product.
Too big to dare might also be a good way to describe them. They are not going to dare making e.g. one great messaging product (or even 2 .. there are after all some niches like office communication). Instead they do 7 and see which one works.
Apparently launching a new product is also the only sure way to get a promotion at Google, not improving an existing one.
You can't start with a cool technology and build a product / user experience around it; you have to start with a vision of a coherent user experience, and build the technology around _that_.
I think that's my biggest frustration with Google: they have interesting and really cool tech that is just way too fragmented. The structure of controlled chaos via launching many products as experiments and culling down to the "best" ones makes it difficult to provide consistent, cohesive experiences.
That's another way of saying don't build-out solutions looking for problems.
Think about it: Google is dominating search and online advertising, they have the most popular web browser in the world, the most used mobile OS in the world, they have Youtube, Maps, Gmail, Chromebooks.
Now imagine if they also ruled messaging, social media, had the most used laptops and tablet in the world, were dominating on cloud, etc...
I cannot see that as being a good thing for consumer.
As guess could be said of any startup though. Not just hardware.
Do I want to spend $500 on yet another google branded brick?
Its killed several hardware projects (Nexus Q comes to mind immediately, though that only got into anyone's hands as a prerelease version at I/O 2012), but I don't know that any of those have involved bricking devices already sold (though cutting off software support and associated services may have reduced the potential lifespan for some, but that happens for old models from most makers even if the project isn't killed; Nexus Q, for instance, had support dropped in Google Play Music and Google Movies & TV in May and June of 2013.)
The Lumia 950 was released at the same time as the Nexus 5X, and despite Windows 10 Mobile being dead, is receiving security updates for a full year longer than the Nexus.
I now use a Macbook, and really wonder why any developer would buy a Pixelbook vs. a Macbook. If you want to run GNU/Linux, you can install a VM, and otherwise you still have a functional OS.
Edit: I see the value proposition in $200-300 Chromebooks. I'm specifically puzzled by the Pixelbook, which starts at $999 (not much less expensive than a new Macbook Air).
Doesn't that sound hypocritical to say the least? Why axe those engineers? Why not get them to fix Android, Fuchsia or whatever needs fixing?
The rule of thumb for me is to buy the device with the highest screen resolution you can afford. Those tend to have the beefiest specs. Common high-res tablets have screens either at 2048 x 1536 (4:3) or 2560 x 1600 (16:10). I prefer 4:3 tablets so on Android, that really only leaves the Galaxy Tab S2 or S3 (the latter having pen functionality).
I'm still getting an iPad because the productivity apps are far superior, purpose-built and actively maintained. And that happens because people actually pay for the app. On Android, everything is freemium, so people are incentivized not to pay, and developers are incentivized to jam up the UX with ads and "unlock this with Premium" CTAs.
Yes, I can write webdev code on my Android tablet with Droidedit, and connect via SSH to my web server. But it's a hassle to do so, especially when the app in question feels like a scaled up phone app instead of a purpose-built tool.
I have a Fire HD 10 tablet that I use for international travel without personal data, but Fire OS is pretty unpleasant and I was hoping to replace it with a new Pixel laptop from Google. Not sure what I'll being doing about that now. (Also, I expect the number of people cross-shopping Fire tablets and Pixelbooks is vanishingly small.)
Unless you absolutely need Android for some reason, I would recommend the iPad.
That said, 2-in-1 x86_64 Chrome OS devices are bearable. Unlike Android; all the drivers are upstream.
I really like my Samsung Chromebook Pro, but would recommend looking for something that's on a later Linux kernel (at least 4.4) because not all the new features get backported. (See http://dev.chromium.org/chromium-os/developer-information-fo... for a list of all the options.)
It's GTalk/Hangouts/Allo/etc. all over again, just this time with Android and ChromeOS. I wouldn't be surprised if they announce that they're going to abandon ChromeOS and that they're going to only use Android instead. How can the different product teams at a company be so completely disconnected from each other?
As near as I can tell from the outside, there aren’t any good Android tablet options, and the problems are baked into the OS at a fairly deep level — deep enough that OEMs can’t really address them, and Google hasn’t shown any recent signs of being to do so either. (For example, read any review of the Pixel Slate).
If you already have a software/app/etc niche you like, I hear the new Samsung tablets are decent (not ‘good’) and not too expensive.
1. This news plants doubt about whether the ChromeOS/Android hybrid in the Pixel Slate is The New Way or a dead end.
2. Google has botched the evolution of Android's tablet capabilities and tablet support in app frameworks so badly there may not be a future for Android tablets outside of Samsung's proprietary extensions to Android. Let that sink in: Samsung may be, by default, leading the Android tablet future.
3. I don't like what Samsung does to Android. This is a matter of taste. But I know I don't like it.
However when tablet dies, I will rather go for a Surface like tablet.
Eventually iPad Pro as alternative.
Getting fed up of Android J++, and a NDK experience that seems like the outcome of a 20% project.
Google backing off on first-party work means that third parties that have always been the backbone of the Chromebook effort have less concern that Google's going to snatch the valuable segment of the market away from them, so I would say that its at least as likely to be a positive sign for Chromebook as a continuing effort as a negative one.
With that said, the Pixel Slate was an unmitigated failure so hopefully the only impact this has is on the Chromebook Pixel line.
Question is if one country is enough to keep it going, even one of the US's size.
They're last place in cloud. They don't have a good search API product to compete with Algolia/Elasticsearch. They can't release new products. Can't build new hardware.
The only things they have going for them are Youtube, Google (the search engine), Android, Chrome,...
Basically everything they did 10 years ago. Even Android + Youtube were acquisitions.
Another good example of too big to fail.
* They have the most popular operating system in the world
* They have the most popular browser in the world
* They have the most popular search engine in the world
* They have the most popular advertising network in the world
* They have the most popular video platform in the world
* They have the most popular mapping software in the world
* They have the third most popular cloud platform in the world
It seems to me that they're doing at least something right
1) They bought Android - they didn't build it in-house.
2) They bought Double Click - they didn't build it in-house
3) They bought YouTube- they didn't build it in-house
4) Third most popular cloud platform is a bit disingenuous; it's like saying Apple Maps is a top mapping platform: sure, you're kind of right, but the gap at the top is fairly wide.
(You might even make the case that the best parts of Chrome weren't built in-house, but taken from open-source contributions via Chromium, but I'm not sure I quite buy that.)
I think the overarching thesis deserves merit: what was the last in-house built product that became a powerhouse with true sticking power, besides Gmail / Google Apps?
But well, if they can keep buying stuff, then they are okay. It's what Facebook is doing.
Everything else is not as good as AWS offerings.
- Android J++
- Kotlin, the successor, still faces lots of tooling issues, regarding incremental builds, code completion, thin APKs,...
- NDK users keep being neglected, after 10 years there is still no alternative to manually write JNI boilerplate, handle native libraries, a proper C++ API in the NDK, the ones that exist like Oboe are either plain code dumps on GitHub, or use Blaze, which isn't part of the standard SDK tooling, IDE tools if they get added always feel like an afterthought to their Java/Kotlin counterparts
- Stable support library and Android Studio releases have as much bugs as canary ones
- Android Studio feels like it needs a rendering workstation to work properly
The last two points have escalated so high that they finally created Project Marble to improve the current state of affairs.
OEMs love Android only because it is free beer.
Google makes the Pixelbook. It is definitely not the "best" laptop when you consider less expensive alternatives, and it is not the best seller.
But, Android would be called a successful strategy.
And, Chromebooks would be called a successful strategy.
What I'm trying to say (and would love to hear comments suggesting otherwise) is that these forays into hardware might just be to guarantee a baseline level of quality for vendors of the products to aspire to. And, in that light, even if the hardware divisions aren't making a dent in the Google's reported revenue, they still serve and were effective in their goals.
These cutbacks are just evidence that Google wants to optimize the way things are done, not evidence that they were failures.
GCP is a different beast, however. :)
The pixel phones on the other hand have been selling better ( in fact grew quite a bit numbers wise as well ). The phone line has a better chance of turning in to the Googles surface line esp. since the marketing around that has been "Its the best camera ever" and people are willing to pay a premium for that.
Outside that they are even less than the worldwide GNU/Linux desktop market, hardly a measure of success.
They raised pricing on Google Maps and have a reputation for dropping services.
Now they require a $75k third party audit to have an app use Gmail API’s.
After that, every other product is just a mechanism to feed the beast including search.
The cost of deprecating products every couple months has to be expensive - if only we could measure it.
Is anyone aware of a good response to this concern?
I understand that Google would rather deprecate something that not support it well but that does not mitigate the apparent risk of adopting / deploying Google products.
