They're talking about the Space Development Agency, which would become the procurement agency for any future Space Force (Mitch also mentioned "Space Patrol" as a potential name, though I think not seriously.)
Jess Sponable said that the ratio between payload and vehicle dry mass is more important than propellant mass fraction. "Propellant is cheap."
AFAICT this is essentially an argument for lower-performance rockets, which is what most of the smallsat launchers are going for. Pressure-fed and electric pumped rockets aren't going to get you the ultimate performance.
Apparently Elon Musk told Jess that he was "just continuing the great work of the DC-X project," which Jess led, after SpaceX flew Grasshopper.
They're talking about reusable spacecraft as a way to be able to deploy resources anywhere in the world from CONUS, reducing the amount the government needs to spend maintaining overseas bases.
Jess: "I think single stage is vastly underrated."
Not for bigger payloads.
There are apparently some new composite materials with some good thermal properties. (Heat shielding is a big issue for such a big, light vehicle, especially if it needs to be reusable.)
"Probably not what you want to do as an entrepreneur."
Mitch: "This conference is now old enough not just to vote, and to drink, and to rent a car, but to be off its parents' insurance!"
Mitch is talking about the need for spy satellites to be large because of the physics of the imaging, so the smallsat launchers can't service them. But then he started talking about the fact that we'd taken a picture of a black hole 55 million light years away with a baseline of only the diameter of the Earth, so maybe it'd be possible to use smaller satellites.
Bill Bruner is interested in manned transatmospheric deployment because he wants humans on the scene making decisions about whether to kill humans, not robots.
Personally, how about nobody kills humans on the other side of the planet at all? Let's focus on killing humans right here at home.
I think this has a lot more to do with who's running the robots than whether there are humans there or not. The drone pilots in Colorado Springs aren't combat pilots.
That's changeable, though.
I agree we should not have AIs deciding to kill people. But the thing about robots is that it's never a kill-or-be-killed decisions. Robots don't have to kill 12 year olds.
Jess: "Smartest thing Elon ever did was flying Falcon 1. Even though he retired it a couple yeas later it gave him instant credibility."
Gerald Nordley (science fiction writer under "G. David Nordley") asked, "How close does a human have to be to the action to make a decision? One light-second? Two light-seconds?"
This is presumably a response to a comment Bill or Jess made about being able to do way more science on Mars in a week than Curiosity could. Which is of course a ludicrous comparison because of the cost difference.
The answer is "it has to be realtime".
These guys act like they have infinite resources available to them.