Exciting stuff! Looking forward to 1.5 and the future direction of the platform.
I know it's been discussed in the past, but what are the current roadblocks to getting Meteor fully on npm? Sashko mentioned maintaining backward compatibility as a primary concern, so I imagine that moving all the core packages to npm and then asking all the package developers to "convert and move your stuff to npm" is not gonna happen (and would break every existing Meteor app). Perhaps with a major version release (2.0?), breaking changes like this could eventually happen? The old Meteor package system would be dropped and fully replaced by npm, and perhaps a migration script could help in converting older apps to use the new npm versions of Atmosphere packages? That sounds like a very difficult problem, so I imagine the solution would be to keep supporting legacy apps and keep Atmosphere going in some diminished capacity for the foreseeable future (I think Sashko mentioned this in the video too). Whatever happens, I'm looking forward to the evolution of Meteor, and I'll help out where I can.
Sashko talks about how they're looking at Meteor as the best and easiest way to get the infrastructure for a modern JS app, i.e. how to not spend time configuring Webpack, choosing Babel transforms, etc. That is exactly what I think Meteor should be. To me it also means they need to move to 100% npm (which they've said is a plan anyway).