[mani04]

From that NASA website on mission logs - anybody else noticed that the success rates have dramatically increased after 2000?

I am wondering if this has got anything to do with significantly enhanced computing power we have nowadays.

[HarryHirsch]

I prefer to think that this has to do with the fact that NASA got rid of Dan Goldin and his "faster, better, cheaper" approach to project management.

In space you can't afford to cut corners. You need three people to tighten a screw: one to set the torque wrench, one to tighten the screw, and one to write the protocol, saying that the screw has been tightened. NASA found out in the 1950s, they forgot and learned it again, and Elon Musk will take note eventually.

[dirtyaura]

Interesting perspective. Is there any analysis written about Goldin's era? I studied AI and robotics in late 90s / early 00s and Rodney Brooks' Behavior-based robotics and NASA's approach were in vogue and were given as examples how to do things (and I believed them too). There is some parallels to MVP thinking in tech community.

What I find interesting is that compared to Brooks it was actually Sebastian Thrun that managed to lead teams to build something remarkable (self-driving cars). IMO, his approach was never based on any single philosophical idea or AI technique, but instead his teams build robotic systems that combined many approaches, and the target was always to build something that worked in a real-world situations, like his museum robot.