1. Performance was kind of a reason for the Air. I want a machine that gives me the perspective my users will have when using my software. 2. But doesn't that suck for actually building? Yes except... I also just have a nice computer on the network that I offload the actual rough tasks to. This might end up being a good solution for you too. 3. This might also future-proof you for the ARM switch. If you have a good build computer, and a cheap "interface" computer, you can switch to a new ARM computer in a bit with little pain: none of the the "real" stuff has to change.
It kinda sucks for multi-monitor though. It really struggles with it.
It's basically passively cooled! There's no heatpipe connecting the fan to the CPU so it kicks up to max all the time trying to cool it but can't make much of a dent on the temp. Really mind boggling.
I had to switch back to my old pro machine and felt like a bit of a berk for not reading up on that aspect.
For me the big factors are that I actually use functions keys (in Emacs, mostly) and just cost. My previous Macbook Pro (2011), which is in its death throes, at the time a top-of-the-line 13", was $1200. With aftermarket 8 GB of RAM and a 1 TB SSD, I managed to keep it usable for almost a decade. At the current MBP configurations, I'd almost certainly end up shelling out a mind-blowing 3-4x that much.
This gives me pause for three reasons:
- All of the soldered on crap means that I can't use aftermarket upgrades to extend the life
- It sounds like an ARM switch is in the cards in the next 1-2 years
- I just don't like having a $3000-5000 portable device. I have my laptop out with me quite often, and I don't relish the idea of having something worth > $4000 with me that may get lost or stolen.