The approaches of Node vs Crypti have been totally different in terms of how they've arrived in the world.
Crypti has been a masterclass in how to put a launch and system together.
Node's been pretty chaotic all along but the developers have always delivered in the end. I think it has plenty of unique developments still to come in the future. We'll see whether the wider world picks up on it.
Both projects need to properly open source, so they can get community assistance on known problems. If you de-obfuscate and inspect both you will find some "obvious" oversights.
Node has two critically fatal flaws, as I see it. The first is related to their build process, which leaves users open to known attacks against nodejs. (Crypti has some of the same concerns, but is not as directly affected and it is at least easier to circumvent their build and use your own nodejs version.) The second is related to their implementation of proof-of-activity forging, where a modified version of the client can build a chain of artificial activity offline, and present it to the network to double spend. I brought the first issue up on their thread, but it never seemed to get addressed, as far as I know. The second issue has been brought up quite a few times by others, both in their thread and elsewhere, so I haven't approached it.
Node also has another little problem where its ledger does not appear to balance. In other words, there appears to be a (relatively very very very tiny, heh) surplus of coins with a few having been created out of thin air. The bug that was used to do this has apparently since been fixed, and the developers claim that no excess coins exist, but as of the last time I scanned their blockchain, a few weeks ago, they were still in there. I'm concerned that the model may allow for similar bugs.
Crypti appears to be vulnerable to classic "nothing at stake" related attacks, though I haven't really tried modelling an exploit for it. I also haven't done a scan over crypti's ledger because their de-obfuscated code is not as easy to instrument. This has been on my TODO for some time now. Crypti's code also appears to be lacking implementation details for some of the features referred to by their team.
It is ironic to me that both of the nodejs coins are trying to do closed, walled garden development. To think they could even "close" their source in the first place exhibits a basic misunderstanding about the fundamentals of the javascript language (as well as computer science in general, and simply typed lambda calculi specifically) that concerns me. I've raised the concern of lack of source to both projects, and was basically told where I could shove it, by both, despite even being an active investor.
I've said before: someone should come in with a proper implementation of coin for nodejs, open source and with a focus on actually working correctly instead of trying to implement fancy features, and eat both of their lunches.
(It would not be hard, maybe I will even try to throw something together just to prove the point.)