> Too many sub-variant models
This is a strength: many parts you can swap (or install later: put that empty WWAN port to a good use!)
> a simple BIOS update is enough to make or break that experience
In my experience, not with Lenovo, but personally I rarely update the BIOS. In fact, I prefer very old BIOSes, hoping for a bug that'll lead to interesting work or exploits: I don't want say the keyboard microcode holes plugged by marking the flash regions read only: that'd lock me out of exploiting them later, say to alter the keycodes!
> No safety in numbers, no network effect:
This is a very VERY wrong take: thinkpads are what linux hackers use the most, on both end of the spectrum:
When they are brand new, you have highly technical people and companies doing fixes or original work (ex: linaro work on the MHI driver for the LTE/5G chip). That's expensive, and likely outside the reach of Framework.
Then when they're old, you've got students, hobbyist etc who find new ways to teach old dogs new tricks: like motherboard swaps (X62, X210 etc)
The framework is interesting, and 10 years from now, maybe it'll be the right choice. Not now, not in the next 5 years.
It's at best a niche choice for people who don't want to tinker too much, and prefer something "average": you won't get bad surprises, but you won't get good ones either.
I like foldable OLED, nice hardware and all that you won't find in a framework now. Maybe later, which is why I really hope they'll be successful!