Ohad your post on facebook today about compilers seems like a major shift from what you've been working on.
An advancement (potentially?) for sure, but a lot of new work to focus on.
not shift at all but realizing the logical properties of the representation i concluded is the way to go for our low-level implementation design. the design didn't change and change isn't considered. but if we could have additional security layer in the form of verified jit compiler, it worth considering which additional implications this might imply. such verified compiler wouldn't be possible without very recent advances over the last decade.
if things will look as i guess they can look, then the compiler-construction tool will be even simpler and smaller than the compiler itself. it might even make things that were possible to be done pre-genesis, to be done post-genesis. e.g. extra language features.
but it's still too soon to draw conclusions. i still need to study this subject more.
Would you care to summarize more of your thoughts on it here?
screenshots here with more explanations
http://imgur.com/a/dTPJP