Is this basically an end-dump of the work that Sam and David did? Or a newly activated resurgence?

Listening to Sam Adams talk about this and some of the ideas they were playing with for massively swarmed processors was pretty cool 10 years ago. Definitely challenged a lot of the classic assumptions about computing dependability, and new paradigms/ideas for how to deal with hyper scale like this.

I wonder if either Sam or David is actively working on these things anymore? Or have they retired now?

Is there are recording somewhere of the talk you mentioned?

Awesome idea but it seems to have died a decade ago.

Suggest Pharo, that's the Smalltalk continuation effort that is the most successful.

Well updated with traits and other modern paraphernalia.

There was a lot of improvements in the CogVM in the last years. The team Eliot Miranda leads is awesome. I hope they finish the speculative inline JIT VM soon.

Miranda's work on it is truly one of the most herculean efforts I've encountered.

...except the non-deterministic parallelism that this offers.

That's the thing. Smalltalk is historically extremely bad at utilizing multiple cores (not because it can't be done, but because no one has - multiple people have thought that an actor-oriented VM might be a good fit). This utilized multiple cores by embracing non-determinism. No other Smalltalk VM offers that.

Squeak is the right one, made by the experts. Pharo breaks stuff all the time

Squeak is still pretty good! The new 6.0 release is forthcoming.

only dead when unavailable