I highly doubt any power users will move to IOS. IOS is far more limiting than even a locked down Android. Graphene is a far more likely candidate, though I suspect most will just stick to normal Android with unverified apps enabled as they have their phones configured today. This change is really just an added 24hr delay to the existing process of enabling unverified app installation. If Google had just announced it this way I suspect there wouldn’t have been so much pushback. But instead it was rumored that unverified installation would only allowed over adb which would absolutely be too much friction with little security benefit.

They’ve recently announced the flow to enable unverified apps. It’s a one time process which will require waiting 24hr after enabling unverified apps but after that 24hr installing APKs will work exactly the same as today. It’s annoying Google has single-handedly decided to implement this rather than going through a more transparent method built into AOSP, but if this is as far as they take it I doubt any power users will be seriously impacted. Of course with how Google has handled this I have no confidence that this is as far as they’ll take it.

If this is really as straightforward as it sounds then I’d consider this the best case scenario. Google could have gone full Apple style lockdown or even just have implemented this flow on a per app basis, but needing to wait 24hr one time to enable unverified app installation isn’t a bad idea from a security perspective. It prevents a bad actor with temporary access from being able to do much while not getting in the way of us power users after the initial 24hr period.

My bigger problem is how Google is leveraging their monopoly to implement this single-handedly and only for themselves. If they had instead gone through AOSP this perhaps could have been implemented in a better way to allow other parties than just Google to be the verifier, and that 24hr waiting period could be applied to any verifier that is not the phone’s default. I’d argue this would be an equally reasonable security measure considering how many scams are out there preying on those who aren’t technologically savvy, yet would maintain transparency.

Lyft been doing this since last year too. In Chicago our group of 5 was told to have our two men seated in the back row because the driver wasn’t comfortable with men. We were confused but the driver knew the ride was requested by a woman in our group and therefore assumed the whole group for the Lyft XL would be women I guess?

I can’t say I’ve had a great time with audio in either personally, though it’s indeed much easier to fix audio problems in Linux. But just yesterday pipewire must have hung or crashed preventing all browser based video playback entirely, which due to the symptoms not appearing audio related was quite annoying to debug. I still have no idea what caused it in order to avoid it happening again in the future.

Depending on how you use your phone this concern may not be as big as you think, flip phones spend the vast majority of time with the screen closed and protected. My partner got a 2024 Razr and has been using it daily for 1.5yr without any scratches or other screen related issues (there’s some expected slowness from the clunky MTK processor of course). I was skeptical of the durability at the time but for $450 with an 18mo warranty including accidental damage figured it was worth the risk. The technology is much more mature than I realized.

All my friends who have one rave about the reliability. I can’t get one on principle. They’re all at the mercy of Bambu’s closed source firmware which means you’re always only one automatic update away from a required subscription.

These attacks are more around the encryption and all require a fully malicious server. It sounds like Bitwarden is taking these seriously and personally I’d still strongly prefer it to any closed source solution where there could be many more unknown but undiscovered security concerns.

Using a local solution is always most secure, but imo you should first ask yourself if you trust your own security practices and whether you have sufficient hardware redundancy to be actually better. I managed to lose the private key to some Bitcoin about a decade ago due to trying to be clever with encryption and local redundant copies.

Further, with the prevalence of 2FA even if their server was somehow fully compromised as long as you use a different authenticator app than Bitwarden you’re not at major risk anyways. With how poorly the average person manages their password security this hurdle alone is likely enough to stop all but attacks targeted specifically at you as an individual.

On Android Newpipe has the best UI in my opinion.

Snaps bundle dependencies and sandbox applications. The dependencies aspect is what matters more to me, but apparently there’s also security benefits if you were to try to install a malicious program.

You can remove snapd, doing so also removes a number of built in apps. But at that point you may start questioning why you’re not just using Debian stable and add the stuff you want. Both of these options pretty much defeat the point of what Ubuntu was.

18hr of battery with the display off is a killer, and even if you could get an m.2 modem working in it m.2 modems tend to be far less efficient than the ones integrated to cell phone hardware. At least if my experience with Quectel and Sierra m.2 modems is representative of other brands.

I get what Canonical was going for with snaps but wow did they ever ruin Ubuntu’s reputation. It used to be the clear choice for anyone who wanted a generic Linux where you don’t have to configure everything yourself. Sure some people didn’t like Unity but the core distro still worked well and was stable. With snaps, package management has become more complex than other distros while decreasing performance if memory limited (and who isn’t nowadays). The number of times I’ve had something not work in the “stable” snap package is far too many, and it’s pretty much always fixed by installing the same package with apt.

I get the reasoning for sandboxing applications, but they needed to wait until it was more stable to make the default. At this rate I doubt we’re ever going to get a truly mainstream desktop Linux distro rivaling macos and Windows…

Power management could still be a lot better for Intel laptops (though admittedly over the past decade it’s come a VERY long way). On my Chromebook running Ubuntu the powersave governor noticably stutters as it decides whether to boost the clocks, but all the other governors significantly hurt battery life. Somehow Windows managed to solve this battery problem with all its bloat, and Chromeos also has while also ultimately running Linux under the hood. Laptops could really benefit from the same level of driver maturity as desktop platforms.

I’d also point out touchpad gesture support as a secondary point which is lacking. I love that pixel perfect scrolling and gestures are integrated into many desktop environments now, but they lack configuration for sensitivity and in some cases leave it to the applications themselves to control. Scrolling in Chrome is way too fast and Firefox way too slow for my trackpad, but unlike the cursor speed/acceleration, there is no setting to adjust the sensitivity of pixel perfect scrolling in supported applications.

This process pretty much summarizes why I’m scared to try changing companies lately. Presumably these measures are to make sure you’re not cheating with AI, but then if you get the job they expect you to use AI.

I like in-person interviews most, they totally resolve the trust concerns. And to other engineers interviewing you using fewer MS products is typically viewed as a good thing. But getting to the in-person part is difficult in this market even if you’re willing to put up with all their spyware from what I hear.

Indeed it’s misleading wording but credit where credit is due, this is far better than turning them all into e-waste. It’s not like anyone bought these with the assumption they would have any sort of official API someday, especially after seeing how Sonos handled their similar situation…

Full Windows 8.1 thanks to Intel’s x86 tablet push at the time.

Windows RT never made it to any other devices besides the Surface RT iirc and was pretty much an immediate failure.

They made so many terrible Windows 8.1 tablets which they had to support. I used one of these with an atom z3735f and 2GB of RAM as my only Windows computer for a long time, and Windows 8.1 was completely smooth on it despite the anemic hardware. Some even cheaper tablets and mini PCs released with 1GB RAM and 16GB emmc yet somehow also were also able to run Windows 8.1 okay.

This is almost certainly US Mobile. They have some really interesting plans at surprisingly reasonable prices. But ultimately if you just want the cheapest cell service then they’re not the carrier for you.

https://youtu.be/Khv-0X1IURs

All web browsers are nearly unusable with 4gb of RAM lately. Even with desktop Linux I usually have nearly my full 8gb used. With 8gb AND Windows it’s only a matter of time before these computers become unusable…

Company computers often come with pre-installed spyware which is notoriously RAM hungry. My company laptop immediately after boot uses nearly a full 16gb before you open any programs. Luckily our IT department realizes this and only allows us to purchase machines with 32GB and up. They’re probably not happy with the current prices, but being a F500 company they can afford it…