I’d add that I was was initially somewhat interested in ARM hardware, but I’ve cooled a lot on it.

For me, and I suspect a number of others, power efficiency is the main appeal.

First, even on Linux, where a lot of software is open-source and some distros have ARM builds, there’s a lot of closed-source software out there, like Steam games, that are x86-based and won’t ever be ARM, and if you’re emulating x86, your power-efficiency benefits go away.

Second, the ARM world is more SoC-oriented, so you don’t have the ecosystem of drivers for modular hardware that plays nicely, and a lot of SoC data isn’t available. This is not a minor issue. An ARM system is not just an x86 system with a slightly-different processor. Whole different world.

Think this sums it up:

https://www.reddit.com/r/linux/comments/1b0lbva/linux_for_arm/

With Arm support, it greatly depends on the specific boards as well as its general popularity. It isn’t at all as well standardized like x86 is. Each device takes at least some amount of unique customization, even before you get into the video and other hardware drivers, often not open source. ie Qualcomm.

An image for a Thinkad X13S won’t necessarily work or even boot on a Lenovo Flex-5G, for example.

The good news is that there are growing numbers of people hacking on these to get better support and usability.

For the X13s, at least. My Flex 5G does not seem to be very popular, lol.

I think this point is often underappreciated - if ARM is to take off for personal computing, there’s a lot of standardization work yet to be done.

Third, for a lot of software, what matters is single-thread performance. And x86 is ahead there.

Fourth, I was recently in a discussion with someone on here and they informed me that the power-efficiency gap has narrowed (at least on Apple’s ARMs, dunno about all ARM sysyems) since Apple’s M1 release, when it was more-significant. I haven’t looked into that, but given that that was the major selling point, it also gave me pause.

EDIT: All that being said, I am totally onboard with wanting laptops with long battery life and the state of things in 2024 is a favorite pet peeve. I would very much like to have a laptop with beefy batteries in the vein of some of the older Thinkpads. I had a T460 or something with two batteries, one removable and where one could get a larger battery that just hung out the back. That was fantastic. Getting a laptop with a single fixed 100Wh battery is very difficult these days, and pretty much nonexistent outside of power-hungry gaming systems that will burn through even the larger battery in short order. Getting a multi-battery system that can be expanded (even past 100Wh) is pretty much only the domain of a few very expensive “ruggedized” laptops like the Panasonic ToughBook.

I’m not sure whether that’s because the typical consumer:

• Cares way more about weight than I do.

• Is way more price-sensitive than I am. 100Wh batteries aren’t that expensive, not in 2024. Here’s a device with a 146Wh battery and an inverter and charging hardware and a case on Amazon for $76.

• Does not care about battery life, like, uses their laptop never far from a plug.

• Just doesn’t consider battery life when buying a laptop.

• Is willing to live with swapping USB PD power stations in. The problem here is that while there is actually a USB device class for power sources that permits a battery bank to report remaining capacity – I checked this in the USB spec the other day – (though I don’t know whether Linux can use this, map it to a /sys/class/power_supply device the way it can ACPI batteries), I haven’t been able to find anyone who actually implements this on their power station and advertises it. No power stations that I own implement it. And I want things like my laptop telling me remaining time to keep working, which they cannot, absent that information.