A bootloader shouldn't take more than a day to do - for something designed for wide distribution like this, it should be pretty much the first thing that gets done, as it will easily save way more time in the long term.
Years ago, when I started with software development, I had the same drive: "yes, it only takes one day". This almost became the answer for any task, no matter of the complexity. Then I learned there is a big difference between "prototype" and "production", between "proof of concept" and "stable". Then the bugs, which doubled everything beyond initial estimation. I learned that to stay committed to a direction and actually finish something beyond a simple prototype (putting to production for instance), there needs to be a balance in the decisions: some features go first, others are part of future updates.
IMO, first thing was to get the system going and do its job, and it took a lot to make that happen (hardware, suppliers, manufacturing, firmware, server infrastructure, API access, frontend, bugs, new features, and then the human part: emails, support).
But, I already said this feature is on the list, and will be done when times allows it. Some hardware changes are needed too, like adding a FT232 for USB upgrades and an EEPROM for OTA upgrades. The USB thing triggers a change to the DC connector used, and I'm not happy with that.