The end of ChromeOS is a new dawn for cheap Android laptops | Hacker News Reader

The end of ChromeOS is a new dawn for cheap Android laptops

zdnet.com

40 pointsthunderbong54 comments

To add a data point to some of the anecdotes here: my mother is still rocking my original 2013 4GB Chromebook Pixel, which has been running on CloudReady since maybe 2021. She has had it for maybe 6 years, after I gave it to her when it was already 4-5 years old.

That thing is on 16 hours a day. It doesn't skip a beat. She loves the screen, loves the speakers, spends all day instagramming and managing her Etsy shop from it. The only thing to ever go wrong with it was a cpu fan that a couple of years ago started to get stuck on full blast - that sorted itself out after a firmware update earlier this year. I had to convince her to let me try to run the update & ever since she's had real anxiety about it one day no longer working.

I rate ChromeOS as one of the finest sandboxed user experiences ever made. Ethical issues aside, it was _the_ OS that made me believe in PWAs, and I'm not sure I'd have been able to stay a full time Linux desktop user for the last decade without it and it's contributions to Chrome/Chromium.

To the reports about "sunsetting" ChromeOS and/or merging it with Android I can only say: good riddance! I bought a Lenovo Ideapad Duet Chromebook back in 2020, but the performance (with the dreaded interminable Linux out-of-memory hangs when having too many tabs open in the browser) and battery life were consistently worse than comparable Android tablets, so by now it's mostly gathering dust...

I'm with you on the sunsetting of ChromeOS, but I'd like to offer a counterpoint to your experience of the Duet. I've been using the newer Lenovo Duet 5 with PostmarketOS[0] (linux for ARM) as a daily driver for more than a year and it is almost always great (minus no functioning webcam, which is abysmal anyways)!

[0] https://postmarketos.org/

Some of these devices probably make nice Linux laptops if you can get them unlocked.

Iirc, it's very easy to run regular Linux on pretty much every Chromebook. Some of them have some driver issues, though.

I would personally say that "easy" depends a little bit; if your device is listed on https://docs.mrchromebox.tech/docs/supported-devices.html then yeah it's dead easy to reflash the firmware and install any old distro and off you go. Anything not supported by that is a little more annoying because you need a distro that can deal with the default firmware's non-UEFI boot process. I'm very fond of my ARM Chromebook flashed with postmarketos, but it is a little less supported because of the boot process (and in some cases yeah hardware support isn't 100%).

TIL newer chromebooks are codenamed after Skyrim locations. Interesting

My understanding is that it's more generic than that; their codenames are apparently drawn from video games in general. Which, yes, is fascinating.

What makes you think it's a software bottleneck instead of the 4GB RAM

I'd say ChromeOS works fine with 4Gb RAM too. The problem with the Duet is it's 4Gb RAM and a lame 2020 smartphone processor which pushes memory reclaim latencies firmly into the noticable range.

And extremely cheap eMMC storage which makes every I/O access about the same speed as reading an SD card.

I have owned multiple chromebooks and they have all been fantastic.

Maybe limited in some regards, but still the best computing devices I have used.

Never used a chromebook. My very rough conception from when they first came out is that it's a cheap laptop with an OS that tries to present a web browser as the OS. Now, I'm reading comments here about getting them "unlocked" and about them having a non-UEFI (and I guess a non-BIOS) boot process, and that sounds awful, to put it kindly. It sounds like user-hostility and vendor lock-in. What other reason is there to forego a standard boot process for a proprietary one?

How are they the "best computing devices [you've] used"?

> What other reason is there to forego a standard boot process for a proprietary one?

Secure boot that actually works.

https://www.chromium.org/chromium-os/chromiumos-design-docs/...

It's even worse than you think. Chromebooks run open source Coreboot and you can get the code here: https://chromium.googlesource.com/chromiumos/third_party/cor...

I've been taking the exact opposite view: Android is a messy, ugly OS, from the unmaintained hacked-up kernels to the user-hostile UI, and I've wished that they'd scrap it and rebase those usecases onto the shockingly vanilla Gentoo spin that is ChromeOS.

I'd just be happy if by law, security updates had to be updated monthly, with weekly and daily for certain criticals, for a minimum of 5 years.

And maybe a 10% revenue fine, yes revenue, for each missed metric.

Chrome OS devices (depending on release date) are good for up to 10 years.

https://support.google.com/chrome/a/answer/6220366?hl=en

Such a draconian measure for measly 5 years?

In the EU, this will be the case from next year on (2025-06-20). No monthly security patch frequency requirement, but instead "security updates [...] need to be available to the user at the latest 4 months after the public release of the source code of an update of the underlying operating system" [1].

Complying with this new regulation and bumping the Linux kernel version during the device life cycle was also a topic at this year's Linux Plumbers Android MC. [2][3]. This is necessary because the Linux LTS support timeframe is shorter than the by law mandated minimum support period of 5 years.

[0] https://news.ycombinator.com/item?id=41128358

[1] https://eur-lex.europa.eu/eli/reg/2023/1670/oj

[2] https://youtu.be/b9xXCNYMWjY?si=yxDJUbJHko8HvFTA&t=458

[3] https://lpc.events/event/18/contributions/1740/attachments/1...

I agree. On top of that Android apps are designed for use on phones. Some of them are not even particularly good on tablets, and they will be horrible on laptops.

Meanwhile I wish ChromeOS would pivot BACK to just being a browser window and nothing else. I didn't want the windowing experience that ChromeOS added down the line -- I wanted just that simple Chrome browser window, full screen, and nothing else.

As soon as they added a taskbar and made it behave like every other generic desktop environment, it became crap.

I agree, on ChromeOS it's entirely possible for their to be generic images that support a wide ranges of devices all sharing the same update path (like Windows, MacOS).

on Android? very much still in the embedded space on how it's built. Linux on ARM is generally still a shit show.

I'd rather ChromeOS get a Android runtime (not a VM) and replace Android, than the other way around.

A number of kid environments depend on Chrome. Google is basically giving their pie back to Apple now.

The trouble with this approach is that the Android UX as a desktop (even using things like DeX and the Motorola equivalent) completely sucks. It is far worse than, say, Stage Manager on a modern iPad.

Does anyone want to run Android mobile apps on a laptop?

I never once ran an iOS app on macOS as the UI/UX makes no sense even though it is possible.

Makes no sense to me too but there is a vocal minority, as demonstrated when Microsoft removed android app support from windows 11, that use them. It's hard to estimate how many dozens of them there are but I guess it got to be in single digit.

I'm one of those people :)

My use case is running Android apps on my Surface Pro instead of having to buy a separate tablet. It works great for my purposes.

But why ? What app is better on Android than on windows?

I really struggle to see the benefit, android app seem subpar in every way.

- some apps don't have website/desktop versions

- some app testing is better done on the desktop where you can share/mark screenshots

but it's not a huge market for sure...

It makes a lot of sense if the laptop has a touchscreen. It replaces tablets entirely.

I use Chrome OS and I use exactly three Android apps: VLC, Deezer and Tailscale.

A video player is something I literally just watch, a music player is the sort of thing I'm happy to have in a phone-sized window, and a VPN client is of course not the sort of thing I interact with much at all.

Poor Fuchsia is never gonna make it, right.

It’s a kernel. They can migrate Android to it overnight and you won’t see a difference.

If you’re about user space written in Flutter, then forget about it - this ship has sailed.

> Google's custom silicon: the long road ahead

Oh yes, a very looonnnggg road ahead. Where apple has an history of hardware manufacturer and software development, Google is an advertisement company with software development.

They are absolutely not fit to the task, you don't make a new processor just like that or then you borrow one from Samsung and make little modifications but it's not the same thing. We've seen the result with pixel phones missing calls a lot, having bad thermals, and lack of signal.

It took them 3 phone (pixel 6 to pixel 9) to finally somewhat fix the antenna issue, but even then it's still not as good as less expensive phone.

When apple had the antenna gate, they fixed it in the next generation because they know how to work on these things. I believe if Google wasn't able to fix it for 3 generations, it's not the lack of time or resource but they just don't know how to work on hardware and its not going to change anytime soon.

They should rather stick on software and ads.

Huh, they made TPUs ...

Also, making silicon is closer to making software than you might think.

Except when your mistakes stay etched in silicon forever. The software industry is by now very deep in the fail fast and fail often. Maybe the hardware industry is taking a page from that book too but they’re many, many chapters apart.

For Google scale companies the difference between hardware and software is less than you would think.

I’m thinking of a really big difference so “less than I think” is probably still a lot. I’m sure time will show it one way or another. So far their phones for example were plagued by hardware issues. Maybe they’re just finding their footing but the point stands, hardware mistakes are forever and Google is 90% a software company with the appropriate mentality.

The only times I heard about Google being good at hardware were “trust me” anecdotal and unverifiable stories about DC equipment. At best this says Google cares about the HW they use not the HW they sell to an end user.

What sort of hardware issues?

I've had the Pixel 2, 3, 4, 5, 6, 6a and 8 Pro and each has been excellent. I've had no hardware problems with any of them.

I have a Pixel 8 Pro, which is fine, except the GPS takes a very long time to wake up. It is running GrapheneOS, however.

Mine's pretty quick, using the stock OS. I assume in your case it's doing a cold start with no AGPS assistance? If so, it'll take up to a couple of minutes to acquire a fix.

I have no idea, tbh. I just open Google Maps, and it usually takes a few minutes to get a reading. This used to be on the order of seconds on my old (Android) phone.

Are you telling me the tensor backslash was all internet drama ?

Google themselves acknowledged their phone have issue in leaked document.

https://www.androidauthority.com/google-tensor-g6-downgrades...

Beside this recent issue with tensor, everyone know that pixel phone don't hold up. They always break one way or another, which was barely acceptable when they were cheap.

Your collection of pixel either say you got many defect or you switched so frequently you didn't have time to see them appear.

Sigh... you seem to have a very short memory

When apple had antenngate, they first went on a big campaign of denying the issue, then claiming other companies were affected as well and finally claiming you are holding it wrong.

Google has a lot of issues too, but let's not kid ourselves about Apple.

I don’t think lack of hardware expertise is the problem. The biggest issue is that Google leadership does not seem capable of thinking long or even medium term. Waymo, Glass, and the myriad of discontinued software products are a proof of it.

How is Waymo proof that they're incapable of thinking long term? They've been funding it since 2009, and are only now just getting to market in 2024. They are plenty of examples where Google's failed to think long term, but I have a hard time seeing Waymo as one of them. Waymo is operational in several cities in the US so it doesn't seem like Google's about to cancel it and they raised another 6.5 billion earlier this year.

Is there anything remotely competitive for Waymo? GM noped out of the self driving market after embarrassing themselves, and Tesla is still dying on the "we don't need no stinking lidar" hill.

I hope this doesn't mean the end of running Chrome on unofficial hardware via ChromeOS Flex. I would have to think that would be affected since x86 is no longer supported by Android.

A lot of Chromebooks (I suspect a majority, even) are on x86 processors. Therefore, if ChromeOS is folded into Android, I would speculate that one result is that x86 becomes a fully supported Android platform rather than them losing that platform outright.

(And if they do that, then it shouldn't be hard to also include some version of Flex. Or worst case in that universe I'd give good odds of LineageOS or someone doing an unofficial version for arbitrary PC hardware.)

This is news? Android apps have been running on ChromeOS for almost 4 years now if I'm not mistaken.

The article is talking about ChromeOS merging to an Android base.

Ninja edit: this article should clarify some things: https://blog.chromium.org/2024/06/building-faster-smarter-ch...

Well sure, who wouldn't want android everywhere?

Personally, I'm holding out for the android brain implant...

/s?

Reply on news.ycombinator.com