I don't keep up with what's going on in the smartthings world, but I've just been watching Brian from automate your life YouTube and some of the features he mentioned seem pretty cool. There was also reference to local processing.
Does anyone know if it will attract back some ex users?
My guess is no. 
They've been promising more local processing for years, and all we got was the v2 hub that allowed (some) Smart Lighting and Smart Home Monitor automations to run locally, plus a few more devices over time (e.g., Hue Bridge bulbs--the communication for which is already local and just required cloud processing), though devices being "local" on ST doesn't (currently) matter much due to the fact that manual control through the app always needs the cloud and so it's just automations (like the above) where this matters.
Now, they are drastically changing the architecture. Certainly, this opens up the possibility for more local execution. In many cases, this actually seems less likely: custom SmartApps and DTHs, to the extent that their limited documentation currently suggests, will need a network-accessible endpoint like AWS (not necessarily this, but early docs I saw suggested this as an easy host). There's no reason an individual couldn't host this locally if they wanted, but this seems like it would require much more know-how than letting someone else do it. Both also require a lot more than ST's current "we'll host it all for you" paradigm, though it seems clear they're not interested in making the hub host that kind of code directly (like Hubitat, which one of their engineers one took a jab at for doing this). It does seem equally or perhaps even more likely that some of ST's own apps will run locally; in a demo a while back, they mentioned that webCoRE could just become another front end to their new rules engine that rules could run locally.
Even if they do fulfill their promise of more local execution, my guess is that the architecture will seem so different that previous users won't see an immediate appeal in coming back, particularly given that something like Hubitat already exists. I, for one, have little desire to learn an entirely new app/driver model just to develop for the platform (and I think extensibility from devs is part of what makes ST and Hubitat great)...but exactly when and where each is headed isn't 100% clear to me from the minimal docs and examples they've provided, so who knows.
If you've been following this, you may be able to shed some light on this part. The shutting down of the groovy ide in favor of a developer toolset on an api-based ecosystem.. would that mean all existing apps and device handlers written in groovy will now no longer function until they're re-made to work with this new api? Which then I assume would be the end of porting any new apps/drivers from ST to HE
I won’t be going back... no more Groovy SmartApps or DTHs eventually. That’s a lot of work by the community being thrown away. New API SmartApps Will have to be hosted by the developer, not on the ST Cloud Servers. This will greatly dissuade most users from using any community developed SmartApps. Most users are not going to stand up there own home servers or be willing to pay for AWS hosting or similar.
Local processing has been something held out in front of users for over 5 years now. Still, to this day, only SmartLighting and a few parts of the Classic Smart Home Monitor are eligible to run locally. I left ST 2.5 years ago for local processing on Hubitat. In that time, NOTHING has changed with respect to ST allowing any other SmartApps to run locally. NONE of the new ST App Automations/rules are eligible to run locally. Once the ST Classic App is shutdown later this year, only SmartLighting will have any hope of running locally (unless ST makes some major changes.)
That is correct. Samsung has stated that they plan to shut down the Groovy IDE altogether in 2021. At that time, all ST Groovy code will be pretty useless.
ST has not yet shared their new design for what will ultimately replace Groovy DTH for Hub Connected devices. For now, they are suggesting/helping developers tweak their DTH’s to work with the new ST Mobile Phone App. That is needed soon as they want to shutdown the Classic ST App this year.
"Smartthings changes are they for the good?"
GREAT for Hubitat! 
The bits I read is that it's not been fully detailed how they will handle custom apps but they have provided details to developers on how to migrate. Some I've seen have said it's more stable and robust and it appears it's not a re-write process. Although we won't be able to port from smart things going forward.
There seems general excitement and confusion over in the forum tho
Once the groovy IDE is gone and developers forced to use AWS endpoints and Lambda functions the development will dry up. I have stopped my development on SmartThings the moment they announced their intentions at the last developers conference. The only people they had on stage with them from the community was the WebCore team. Samsung doesn't understand that the many of the developers and creators cannot afford to host their functions in the cloud.
That’s only for DTH’s in the short-term, as SmartApps will require a complete rewrite in whatever language the developer desires. These custom SmartApps also have to be self-hosted. There is no confusion regarding the SmartApps. WebCoRE, for example, will cease to function in its current incarnation, as well as Echo Speaks, etc... SmartApps such as SharpTools and ActionTiles have the advantage of of already running on existing cloud servers. Therefore, these applications can simply have the portion of code that communicates with the old Groovy side of ST rewritten to communicate with the New ST API. However, something like webCoRE, which runs 100% within the ST Cloud Server, will no longer function. There are a few ways this could be resolved, but it will require major rewrites.
SmartThings changes are going to be good for Hubitat. I'm waiting for something to come back in stock so I can start switching over.
It's gong to hurt many people that don't even know this is going on and depend on "it just works". Most of the folks I put on SmartThings a while back are happy as a clam with their current integrations and automations. Once Ecobee and other custom DTH's such as WebCore are gone I am going to get a lot of phone calls. I hope my hobby doesn't become a full time job.
It seems that Samsung have buggered Android auto up for many people with a phone update
Download the Hubitat app
