I realize it's FOSS hardware but I'm more of a software person and would probably take me too long to figure out how to hack all those things in and get it manufactured ...
It just isn't in my use case. For me, the only use of a watch is on the go. When I'm not on the go, I already have a computer w/me which is more powerful than any smartwatch. Therefore, if I am on the move, I want to blindly use my watch, just like I type this message blind on a keyboard. Physical buttons aid with that. If I am catching a bus, I want to be able to control Spotify. If I am running, I want to be able to control Spotify. Touchscreens are far more cumbersome in that regard.
Apple Watch, yes very powerful computer, beautiful screen, and very expensive (more so if you want the medical features). You have to give it juice every night, and it requires either a 4G connection or you're having to pay for another plan or it uses Bluetooth w/your smartphone. These devices don't have enough storage for my liking. I mean, I am sure I could add and use a microSD with 256 GB or more to it, if it had that option. Instead we gotta rely on very little storage, Bluetooth with smartphone, or another SIM + data plan. All in all a completely different purpose and price class.
The only sensor on this is the Bosch BMA423 3-axis accelerometer. Unfortunately, that has no gyro or magnetometer, which makes it a bit of an odd choice. It's supply constrained, and has an EOL date of Sept 2021 to boot. The other sensor is a DS3231MZ low-power highly-integrated real-time clock. It's also out of stock, and has a 42 week manufacturer lead time.
I've got some experience developing embedded hardware, but the watch form factor is really hard to hack after the fact. It's not particularly friendly to expansion busses and add-on cards...
How accurately do you care about geolocating yourself?
If you only care about <50m accuracy for the watch to be able to provide you location-based behaviors, I've had very good results doing Wi-Fi scans with ESP32 devices (in particular, a FeatherS2) and feeding the results into the Google and Mozilla geolocation APIs. More than enough to pull up the local weather or public transit info for the bus stop in front of you. It could send data requests through your phone when you don't have Wi-Fi.
Obviously it won't have the accuracy and update rate that a GPS would, but on the upside, it does work indoors.
I love my Garmin, but the aggressive price segmentation (even though I bought the device at the top of the list) is vexing and I would love to help build an open-source alternative.
On the note of pushing data through a phone, I suppose I could stick a uBlox, ESP32, and tiny battery in a case on my belt, and push GPS data to the watch that way, but that's basically the same as the watch minus the display.