In your alternatives considered section, you didn't consider a single mainstream Thinkpad or Dell laptop ? At least the gen 3 Thinkpads with Ryzen seem pretty good, although configs and availability are hit or miss.

Most modern ThinkPads only have soldered RAM/max out at 32GB so were basically a no-go for me. They also had 2mo+ delivery lead times for BTO options, which also put it out of the running. I wasn't super happy w/ my last Thinkpad (X250) and general sentiment seems to be that they've continued to go downhill, so unless they release something that hits it out of the park from a design perspective then I'm not so interested. (The Z13 is actually pretty neat now that it's available and discounted, but only 2 USB-C ports doesn't work for me.)

Dell QC has been pretty terrible the past decade, and they're inferior in basically every way (Intel-only, mediocre cooling solutions, meh battery size, expensive, long-standing (like multi-year, basically entire product life cycle) unresolved bugs/incompatibilities even for their dev edition versions, soldered RAM, awful support, etc).

One other thing is that if you are a Linux user, I really don't see much point in rolling the dice/fighting obscure compatibility bugs (most of them being BIOS/ACPI bugs that are hard to impossible for mere mortal end-users to fix) when there are similar (or better) alternatives that are known to work OOTB, or better yet, explicitly support Linux.

Do people really need more than 32GB of RAM? I'm struggling to come up with a scenario that needs so much RAM.

Obviously most office workers or casual users don't need that much memory, but I've hit swap before w/ 32GB of RAM just with browser tabs and electron apps, so I bet lots of people with 8GB or 16GB hit swap and get slowdowns and just assume their computers are too slow, when really, more memory would have solved their problems. Also, I'm also sitting at 36GB of wired memory usage right now with a single VM running. Anyone running containers or doing any virtualization will want as much memory as they can get, as would anyone working with any number of data sets - I've hit swap on 64GB working w/ "simple" sqlite databases and doing basic pandas work. I also do video and photo processing (including a lot of large panoramas) for fun on the road and that's another whole class of workloads where more memory is better. As mentioned by others, large compiles also benefit from more memory.

Anywhere, here's the rub. I bought my 64GB memory kit for $250. It cost an extra $125 at retail to add an extra 32GB of memory, which almost all laptop manufacturers have decided not to even allow as an option for segmentation/planned obsolescence purposes. Which is fine, that's their prerogative, but since there are enough decent options that allow more memory, it also means that I, or anyone else who needs more RAM just won't be giving them any money instead.

Yes. Not everyone does web development, for example. I compile c++ most days and the more cores and ram I can have locally the better.

I actually bought a Asus ROG Zephyrus G14, pretty much for the same reasons as you listed. The lead times for usual contenders were too high and the G14 was easy to get with the specs I wanted.

If it makes you feel any better about your purchase, I don't get great battery life either. I didn't make any attempt to optimize it, and I get about 3-4 hours of dev time out of it (yes, not very scientific, I know).

Your review was great though, I might get a framework as my next laptop assuming they are still good in a few years time.

Besides the short battery life, I hope your new laptop is treating you well otherwise!

I don't know about your workload, but in you case you ever wanted to poke around, one thing that might be worth looking into is to see whether the MUX switch (only accessible in Windows) is on its dGPU-only mode. On iGPU mode, the battery life tests for idle and wifi web browsing seem pretty good in Windows: https://www.notebookcheck.net/Asus-ROG-Zephyrus-G14-GA402R-G...

On Windows, 3rd parties have been working on a utility called AATU to improve performance and longevity: https://amdaputuningutility.com/

And on Linux, I used https://github.com/FlyGoat/RyzenAdj all the time for my preferred use cases (eg, setting the temp limit below when fans turn on to get a completely silent laptop on battery, lowering max power limits if I wanted to hit a certain amount of battery life, etc). (TLP has run-on-ac and run-on-bat commands so you can easily set your defaults based on whether you're plugged in or not (udev can see power events too).