We've had some discussions in #science. You can see the last post there. My time is really limitted right now, but guys from my laboratory could start without me. What we need is some kind of priority list and specification for what ML needs. A KEEP-like proposal that we can discuss.