I agree with you, the layer 1 shall be stable, fast, and preferably not complex. Am not too technical to understand what this means. I do think we need SMT as part of the core. Again, am not technical enough to determine if 'core' means layer 1, or if that is layer 2.
Your reasoning wrt Smart Contracts and such feature will result into none-scalable services build into smart contracts, I think is a weak argument against adoption of smart contracts to HIVE. In the end the robust, performing and scalable services will find mass adoption. Any service that can't keep up with increase of usage, will simply fail. This is regardless of implementation of services on whatever layer in whatever form.
If we need smart contracts as part of the core of HIVE? Again, am not qualified enough to take a position on it. That said, I do like the aim of the KOINOS team. They state they are designing a blockchain tackling all the core problems of any chain around: essentially robustness, scalability and modularity. They include smart contracts. They even intend to implement consensus algorithms as smart contracts to avoid the constant hardfork-ing we have at our HIVE chain. Are they wrong? I don't know, but their story sounds good and that team has a lot of knowledge of the HIVE chain. And I believe we shall stop all the hardforks. HF's leeds to unnecessary disruptions of the layer 1 which needs to be robust and stable, 24x7!
We should be able to talk to a group with an existing app with an existing community and talk through how we could bring their thousands or millions of users here, have the app continue to function, and how the blockchain will handle that additional traffic.
Definitely we need to get to this! ASAP! Not sure what you mean with bringing those communities over to HIVE. I see it likes this: We enable other communities to make use of HIVE, but they don't have to come to HIVE. I mean: we enable eg tokenisation of xyz community by integrating into their own service. Tokens can be tradable tokens as we have them on our layer 1, and in HIVE Engine (layer 2 I suppose?). But tokens could also be just loyalty points, or stable tokens. Not sure what we need to allow these use cases, but I don't believe in the world coming to HIVE. HIVE needs to go to the world, and that starts with a B2B approach, and leaving the communities where they are; We just need to add HIVE features to their service; Thats it :)
I tried to voice my idea of tokinsed comments to 3rd part services in a recent post. In case you are interested; check this link:
To finish: 3 seconds is fast, but it is not fast enough imho. Can't we get to 1 second? Some chains around can go as fast as 1 second. Waiting is something the mass won't do. When something else is faster, the mass will move to the faster service. It is as simple as that. Now, we may not need to be as fast as 1 second today, but I always learned from great software architects: The foundation is where the time is to be spend and needs to be developed in a solid way, before the rest is build on top of it. The more is build on top of the foundation, the less likely its gonna be the foundation is re-architected. Thats why in the software industry, a popular service/application is rebuild a few times from the ground up in its lifecycle. Most software and services are not well architected from the ground up, at day one; As you may know very well yourself.