2 ms·

I’d rather see a focus on projects that are doable with existing capabilities, and separately an effort to develop new capabilities. It seems like if you immed

I’d rather see a focus on projects that are doable with existing capabilities, and separately an effort to develop new capabilities. It seems like if you immediately aim for something you can’t do, then you develop a bunch of technology and infrastructure that isn’t very good for other tasks. You end up with a stunt that doesn’t really go anywhere. That gets you rockets like the Saturn V, which was amazing but too big and expensive to be used for much besides Apollo. We’re doing it again with SLS: a big, catastrophically expensive rocket built to achieve a specialized goal. It’s not going to be very useful and it might not even fly. Compare with SpaceX’s BFR. It has much more general capabilities. At first blush this might look bad: why lug a heavy heat shield all the way to the surface of the moon and back? Why all this focus on reusability if your goal is to send just a few missions? But at the end of the day you’re much better off with a rocket that’s ok for a bunch of stuff than one that’s really good at one thing. Or just look at the space vehicles of the era. Many of them are still flying today, with lots of evolutionary improvements. How many Apollo-derived vehicles are flying today? I don’t think there are any. Apollo makes me imagine a Roman emperor declaring a program to send an expedition to the South Pole. They can’t even get there! You need to work on your ships before you set such a specific goal. As a bonus, steady, incremental improvement is the sort of thing you can sustain across political changes, if you can just get to the point where people stop asking for moonshots.

pgeorgi 7y ago
I suppose the important bit is really that there's universal buy in (and with the moon landing thing that was supported by making it a matter of national pride). With incremental progress you risk things looking invisible (just like AI is always just the things we can't yet do, while newly possible tasks are merely computations). "Do some notable thing in two terms" monument building exercises risk being shot down on budget reasons when any part of government ends up under new management (see Congress). So I guess you can go incremental (but solve the visibility bit somehow) or grand vision. Right now, NASA seems to be going through the worst of all.
avar 7y ago
> [...]rockets like the Saturn V, which was amazing but too big and expensive to be used for much besides Apollo[...] The cost per launch for the Saturn V was around a $1 billion in today's money. That's lower than some estimates for the per-launch cost for the Shuttle. It wasn't unsuitable, the US could have continued with it just like the Russians continued with the Soyuz. > But at the end of the day you’re much better off with a rocket that’s ok for a bunch of stuff than one that’s really good at one thing. Perfecting the rocket is opportunity cost eating up missions you could be flying today if you just mass produced your existing but flawed rocket. The US could have had something like a hundred moon missions many the ISS's mass many times over in orbit by the start of the 90s if they just continued improving on the Apollo architecture.

pgeorgi 7y ago

I suppose the important bit is really that there's universal buy in (and with the moon landing thing that was supported by making it a matter of national pride). With incremental progress you risk things looking invisible (just like AI is always just the things we can't yet do, while newly possible tasks are merely computations). "Do some notable thing in two terms" monument building exercises risk being shot down on budget reasons when any part of government ends up under new management (see Congress). So I guess you can go incremental (but solve the visibility bit somehow) or grand vision. Right now, NASA seems to be going through the worst of all.

avar 7y ago

> [...]rockets like the Saturn V, which was amazing but too big and expensive to be used for much besides Apollo[...] The cost per launch for the Saturn V was around a $1 billion in today's money. That's lower than some estimates for the per-launch cost for the Shuttle. It wasn't unsuitable, the US could have continued with it just like the Russians continued with the Soyuz. > But at the end of the day you’re much better off with a rocket that’s ok for a bunch of stuff than one that’s really good at one thing. Perfecting the rocket is opportunity cost eating up missions you could be flying today if you just mass produced your existing but flawed rocket. The US could have had something like a hundred moon missions many the ISS's mass many times over in orbit by the start of the 90s if they just continued improving on the Apollo architecture.