The thread linked to above in the Bootstrap github does suggest that the only reason they haven't released it as Beta at the moment is time. They have said, again in that thread, that there wont be an Alpha7, that they are all but ready to release Beta - although I'm not going to hold my breath given how long A6 has been there already!
I'm not sure why that would be - it doesn't seem to me to take a lot of time to declare something as Beta if it is, in practice, at that point. There are a small number of changes still going through their processes, but not a great number. I think the biggest changes have already happened (the most significant one in the alpha series, rather than between 3 and 4, was probably turning on flexbox by default as part of the grid system).
I'm hoping that HQ are monitoring what is changing and looking to update Boost when the Beta is released as it is currently still built on Alpha4, so there is some work to do already. (In fairness, I don't think anyone expected Bootstrap4 to still be in Alpha at this stage when the original decision was taken)
I would agree with Fernando to some extent - Alpha code is not suitable to be used in production and probably should not have been used to create a new default theme, that is Boost is great and a much needed development, it should probably have been an optional theme for now. BUT it isn't and HQ have gone down that road and with BS4 being downloaded and included as part of that Boost theme, it is at this stage upto HQ to ensure that Boost is NOT alpha standard code. Child themes are then built on Boost, rather than directly on BS4, giving a slight cushion.
Given that Boost has been released as the default theme, and with the assumption that if anyone does find significant bugs, then HQ would need to fix those in Moodle, pending any upstream fix in BS, then I have been happy to create child themes for Boost and to work within that framework and push it forward. I do so on the assumption that the heavy lifting of updating Boost to Bootstrap4 Beta and final will be part of the ongoing process by HQ.
I do, though, always caution people about the current nature of BS4 - and with full realisation myself that Beta/final changes could involve additional work in those child themes. And I think that is the key - if you accept Boost and Boostrap4 for what they are, and acknowledge the fact that there is a potential overhead when BS4 moves from alpha, then I don't personally have a problem with it (I've never had any issues of it crashing or being unstable, just feature developments), BUT that does mean it isn't necessarily suitable for many institutions who want to be able to buy/commission/create a theme and mainly leave it alone for a couple of years for stability.
So far, I've kept Waxed and Handlebar relatively small and manageable within that kind of overhead - I believe even if there are significant changes that I can update them fairly quickly. I, personally, would not even attempt to migrate a theme like my Flexibase, or something like Adaptable or Essential across to Boost/BS4 until all that overhead is known and quantifiable (eg probably a late beta or RC version).
From my point of view: Boost (and its child themes) are stable (they wont break your system), even though the upstream BS4 library is alpha. But they come with a potential development overhead that needs to be considered and which could be significant.
Gareth - re SASS/LESS. I personally find SASS better/easier, but that's possibly because I did a couple of projects with it before switching to LESS to use with Moodle when we adopted BS in the first place. There's probably not really that much difference between the two in practice (although the inline SASS compiler HQ have used seems to be dreadful - slow and at least one major bug, as you detailed, around the @import url function)