7 ms·

It seems to me SpaceX's secret sauce is continuous iteration on vehicles that actually fly. It's hard to see how Boeing or Blue Origin can develop expertise wit

It seems to me SpaceX's secret sauce is continuous iteration on vehicles that actually fly. It's hard to see how Boeing or Blue Origin can develop expertise with any confidence their designs will work, except, apparently, by going very slowly. Of course, Boeing has somewhat of a track record, but Blue Origin was founded before SpaceX and still hasn't achieved orbit.

baybal2 6y ago
I don't think so. Musk has hired pretty much the same people who were making the same rockets for LM, Boeing, and NASA. Their experience surely contributed to their speed. To me, they give a feeling that they have some very experienced, and conservative type managing their engineering who does everything by the book.
- bpodgursky 6y ago
  I think there's some nuance here -- my take is that SpaceX has a very good balance of knowing when going slowly and carefully is critical (launching important customer vehicles, launching crewed missions), and when it's critical to move fast and iterate (eg, every other time). SpaceX has blown up a LOT of vehicles via mistakes, by iterating fast and learning from them. They use these as learning experiences, and if they can get a few satellites up in the interim, that's great (see: the starlink launches on re-used rockets). But when it comes to launching actual astronauts, they checked every one of NASA's million boxes. NASA and Boeing are unfortunately incapable of operating in fast-and-loose mode, even when it would be better to iterate and break a few rockets. The two-mode operation is why SpaceX dominates, and likely will continue to crush.
  - baybal2 6y ago
    I believe they are differentiated by not doing a lot of random iterations in comparison to the competition. Blue Origin for example managed to completely redo pretty much everything about their design few times over, while SpaceX had something resembling Falcon 1, and Falcon 9 on their drawing boards pretty much since the beginning. Too much pivots, too much iterations, and close to no straight advancements.
    - bpodgursky 6y ago
      I mean, I wouldn't call that a real "iteration", since Blue Origin didn't actually build or launch any of those designs.
    - Shorel 6y ago
      IMO a restart is not an iteration. Iteration is "improve the current model". A restart throws the current model away.
- nordsieck 6y ago
  > To me, they give a feeling that they have some very experienced, and conservative type managing their engineering who does everything by the book. I think this is pretty clearly not true. "The book" says that once you have an operational configuration you freeze it. I remember hearing that the shuttle program was having to scrounge for old CPUs from warehouses because someone of the processors it used were no longer manufactured, and they didn't want to make changes to the hardware. In contrast to this, SpaceX is famous for continuously modifying their rockets - so much so that one of the NASA requirements for certification was that they freeze the design of their F9.
  - leoc 6y ago
    > I remember hearing that the shuttle program was having to scrounge for old CPUs from warehouses because someone of the processors it used were no longer manufactured, and they didn't want to make changes to the hardware. Indeed they were literally buying Shuttle replacement parts from eBay: https://www.nytimes.com/2002/05/12/us/for-parts-nasa-boldly-goes-on-ebay.html https://www.nytimes.com/2002/05/12/us/for-parts-nasa-boldly-...
  - orbital-decay 6y ago
    >"The book" says that once you have an operational configuration you freeze it. I guess we've been reading different books then. The Shuttle you're referring to had been constantly modified and tuned (noise suppression to prevent STS-1-like body flap damage, autopilot, the anti-geyser line, LWT, SLWT etc etc etc). The other manned vehicle, Soyuz, is on its 8th major iteration already, and got many minor modifications in between. They looked for old CPUs instead of upgrading because it wasn't worth it. Shuttle program got cancelled in 2004, and it was known well in advance that it will stop flying. The difference is, they don't make headlines nearly as much. (Shuttle did early on, but it was too long ago)
bane 6y ago
I agree with this. It seems pretty clear to me that SpaceX's key innovation is one of management -- namely rocket engineering using software-like iterative methods. The driving force of course is Musk, who comes from software but moved into hardware where we see the exact same iterative development at Tesla. There's no reason that other launch providers can't be doing what SpaceX is doing, but they aren't. They continue to design and build using more traditional and conservative aerospace management approaches. Blue Origin is run by somebody with a background in finance and retail. ULA by a former mechanical engineer. Arianespace is run by a former policy guy turned exec. Rocket Lab is run by an engineer. Scaled Composites (I know a different kind of beast) is run by an aerospace engineer. and so on.
- GlennS 6y ago
  Software people didn't invent the idea of doing quick, cheap iterations. It predates software really. We have embraced it strongly as an industry though. Except for the C++ people of course.
- redis_mlc 6y ago
  > Scaled Composites (I know a different kind of beast) is run by an aerospace engineer. Scaled Composites made junk (one-off subsonic airplane prototypes), and was sold to Northrop Grumman in 2007. The current name is "Scaled." https://en.wikipedia.org/wiki/Scaled_Composites https://en.wikipedia.org/wiki/Scaled_Composites
- kiba 6y ago
  I agree with this. It seems pretty clear to me that SpaceX's key innovation is one of management -- namely rocket engineering using software-like iterative methods. The driving force of course is Musk, who comes from software but moved into hardware where we see the exact same iterative development at Tesla. This is old stuff. They called it "skunkwork".
  - mcv 6y ago
    And Skunkworks was very successful too. Are they still using the same approach, though? I'm getting the impression they're not what they used to be.
paranoidrobot 6y ago
I think it's a bit unfair to look at Blue Origin like that. While it's true that Blue Origin hasn't done anything more than a few test flights, they've definitely got a different view of the way to develop rockets. Blue Origin effectively has no requirements to make a commercial return. Until Jeff Bezos runs out of cash (unlikely), gets bored (seemingly unlikely), or pissed off (who knows) they can keep going doing whatever they like. Certainly making money is an eventual goal, but it's not like a few delays or failures are going to result in them being unable to make rent. For this reason, they've developed from the other end, first - perfecting their launch and landing with a smaller-scale rocket and iterating from that perspective. SpaceX on the other hand is definitely marked by their early days where they were down to their last few $ and Elon being out of cash to keep running the business if it didn't work. So they've been working mostly on the "get something that customers will buy" side of things, with the addition of landing/re-use coming secondary.
- Spare_account 6y ago
  > Blue Origin effectively has no requirements to make a commercial return. But do they have any requirement to make a rocket? When will New Glenn fly? Will they ever produce something that does something useful? I hope they do, we need competition in commercial space flight.

baybal2 6y ago

I don't think so. Musk has hired pretty much the same people who were making the same rockets for LM, Boeing, and NASA. Their experience surely contributed to their speed. To me, they give a feeling that they have some very experienced, and conservative type managing their engineering who does everything by the book.

bpodgursky 6y ago
I think there's some nuance here -- my take is that SpaceX has a very good balance of knowing when going slowly and carefully is critical (launching important customer vehicles, launching crewed missions), and when it's critical to move fast and iterate (eg, every other time). SpaceX has blown up a LOT of vehicles via mistakes, by iterating fast and learning from them. They use these as learning experiences, and if they can get a few satellites up in the interim, that's great (see: the starlink launches on re-used rockets). But when it comes to launching actual astronauts, they checked every one of NASA's million boxes. NASA and Boeing are unfortunately incapable of operating in fast-and-loose mode, even when it would be better to iterate and break a few rockets. The two-mode operation is why SpaceX dominates, and likely will continue to crush.
- baybal2 6y ago
  I believe they are differentiated by not doing a lot of random iterations in comparison to the competition. Blue Origin for example managed to completely redo pretty much everything about their design few times over, while SpaceX had something resembling Falcon 1, and Falcon 9 on their drawing boards pretty much since the beginning. Too much pivots, too much iterations, and close to no straight advancements.
  - bpodgursky 6y ago
    I mean, I wouldn't call that a real "iteration", since Blue Origin didn't actually build or launch any of those designs.
  - Shorel 6y ago
    IMO a restart is not an iteration. Iteration is "improve the current model". A restart throws the current model away.
nordsieck 6y ago
> To me, they give a feeling that they have some very experienced, and conservative type managing their engineering who does everything by the book. I think this is pretty clearly not true. "The book" says that once you have an operational configuration you freeze it. I remember hearing that the shuttle program was having to scrounge for old CPUs from warehouses because someone of the processors it used were no longer manufactured, and they didn't want to make changes to the hardware. In contrast to this, SpaceX is famous for continuously modifying their rockets - so much so that one of the NASA requirements for certification was that they freeze the design of their F9.
- leoc 6y ago
  > I remember hearing that the shuttle program was having to scrounge for old CPUs from warehouses because someone of the processors it used were no longer manufactured, and they didn't want to make changes to the hardware. Indeed they were literally buying Shuttle replacement parts from eBay: https://www.nytimes.com/2002/05/12/us/for-parts-nasa-boldly-goes-on-ebay.html https://www.nytimes.com/2002/05/12/us/for-parts-nasa-boldly-...
- orbital-decay 6y ago
  >"The book" says that once you have an operational configuration you freeze it. I guess we've been reading different books then. The Shuttle you're referring to had been constantly modified and tuned (noise suppression to prevent STS-1-like body flap damage, autopilot, the anti-geyser line, LWT, SLWT etc etc etc). The other manned vehicle, Soyuz, is on its 8th major iteration already, and got many minor modifications in between. They looked for old CPUs instead of upgrading because it wasn't worth it. Shuttle program got cancelled in 2004, and it was known well in advance that it will stop flying. The difference is, they don't make headlines nearly as much. (Shuttle did early on, but it was too long ago)

bane 6y ago

I agree with this. It seems pretty clear to me that SpaceX's key innovation is one of management -- namely rocket engineering using software-like iterative methods. The driving force of course is Musk, who comes from software but moved into hardware where we see the exact same iterative development at Tesla. There's no reason that other launch providers can't be doing what SpaceX is doing, but they aren't. They continue to design and build using more traditional and conservative aerospace management approaches. Blue Origin is run by somebody with a background in finance and retail. ULA by a former mechanical engineer. Arianespace is run by a former policy guy turned exec. Rocket Lab is run by an engineer. Scaled Composites (I know a different kind of beast) is run by an aerospace engineer. and so on.

GlennS 6y ago
Software people didn't invent the idea of doing quick, cheap iterations. It predates software really. We have embraced it strongly as an industry though. Except for the C++ people of course.
redis_mlc 6y ago
> Scaled Composites (I know a different kind of beast) is run by an aerospace engineer. Scaled Composites made junk (one-off subsonic airplane prototypes), and was sold to Northrop Grumman in 2007. The current name is "Scaled." https://en.wikipedia.org/wiki/Scaled_Composites https://en.wikipedia.org/wiki/Scaled_Composites
kiba 6y ago
I agree with this. It seems pretty clear to me that SpaceX's key innovation is one of management -- namely rocket engineering using software-like iterative methods. The driving force of course is Musk, who comes from software but moved into hardware where we see the exact same iterative development at Tesla. This is old stuff. They called it "skunkwork".
- mcv 6y ago
  And Skunkworks was very successful too. Are they still using the same approach, though? I'm getting the impression they're not what they used to be.

paranoidrobot 6y ago

I think it's a bit unfair to look at Blue Origin like that. While it's true that Blue Origin hasn't done anything more than a few test flights, they've definitely got a different view of the way to develop rockets. Blue Origin effectively has no requirements to make a commercial return. Until Jeff Bezos runs out of cash (unlikely), gets bored (seemingly unlikely), or pissed off (who knows) they can keep going doing whatever they like. Certainly making money is an eventual goal, but it's not like a few delays or failures are going to result in them being unable to make rent. For this reason, they've developed from the other end, first - perfecting their launch and landing with a smaller-scale rocket and iterating from that perspective. SpaceX on the other hand is definitely marked by their early days where they were down to their last few $ and Elon being out of cash to keep running the business if it didn't work. So they've been working mostly on the "get something that customers will buy" side of things, with the addition of landing/re-use coming secondary.

Spare_account 6y ago
> Blue Origin effectively has no requirements to make a commercial return. But do they have any requirement to make a rocket? When will New Glenn fly? Will they ever produce something that does something useful? I hope they do, we need competition in commercial space flight.