I generally use Cuis as opposed to Squeak for the bulk of my day-to-day Smalltalk needs. I do still use Squeak on a semi-regular basis to cross-check issues (i.e. is a problem I'm having with the image or VM etc) and to look at older (3.7 and earlier) images for ideas.
Mainly too much accretion of cruft and the resulting instability when my code attempted to sit on top of said cruft.
Cuis. It was the best match for my needs that I found after spending a few years bouncing between Squeak/Pharo/Cuis. I view this as unfortunate: a minimal base image very much like Cuis (only moreso) should be a shared foundation that all dialects (Squeak, Pharo etc.) could be based on. The Smalltalk community is small enough and fragmenting it further isn't good long term.
Tactically it would need to be much more minimal and modular, have API stability, documentation, direction and stop reinventing every wheel just because it's not written in Smalltalk (i.e. pick your battles). Strategically, while I very much appreciate everyone's efforts on Squeak I honestly don't see a long-term vision and think that's a problem.
When I was a beginner: getting started (I can't tell you how many times I threw up my hands and said 'not worth it' given all the missing/broken functionality, outdated documentation, broken packages etc. I started to see some daylight as I started playing with Cuis followed by a couple of big wins that sold me on sticking with it). At an intermediate level customizing core behavior (I'm looking at you, Morphic). And now that I've got a moderately large codebase I can't imagine how much effort it would be to keep it running on Squeak.
Everything I'm using Cuis for. (This is not at all a knock on Cuis as I love what it offers me. Rather, just acknowledging the reality that Juan or our small group of users often have to duplicate effort on things I think we (or at least I) often really wish we didn't have to.)
Thanks,
Phil