4 ms·

> Pity they didn’t learn the lesson and maybe Columbia in 2003 could have been saved. But again, mission controllers and engineers ignored the information as “i

> Pity they didn’t learn the lesson and maybe Columbia in 2003 could have been saved. But again, mission controllers and engineers ignored the information as “insignificant". That was in the comments. That is the best part of the article. We have a long road to go about learning from things that worked ok by pure luck. But, as usual, we barely learn from mistakes...

JustSomeNobody 8y ago
Which should definitely make one worry about our ability to produce self-driving cars that are safe[0]. [0] Yes, safe is relative. Are they safer than humans is one measure. But should we stop there? No. Can we truly even get there? ...
ubertakter 8y ago
I don't think any engineers deemed any available information about Columbia as "insignificant", not that there was very much information to go on due to lax execution by camera tracker teams. Management on the other hand was... not effective. I think that's really the crux of your statement. Humans seem to be terrible at working together in large organizations, or at least running them. Organizations seem to learn from mistakes only temporarily, or at all (as you pointed out). Consider the self-driving car group at Uber as a recent example.
- woliveirajr 8y ago
  Yes, and I agree with you. In my experience, even using some kind of protocol to prioritize which problem should be taken care first, it's easy to miss really important things because someone thought that "time to repair" should give the same score as "severity of consequences", and then we begin to fix things that are "easy and fast", not that are "relevant"
8bitsrule 8y ago
"But, as usual, we barely learn from mistakes... " As in, for example, public warnings by countless actual hackers back in the 80s and 90s about how vulnerable most US systems were. (For which, in return, they were persecuted and prosecuted. With no apologies as said warnings came to deserved fruition.)

JustSomeNobody 8y ago

Which should definitely make one worry about our ability to produce self-driving cars that are safe[0]. [0] Yes, safe is relative. Are they safer than humans is one measure. But should we stop there? No. Can we truly even get there? ...

ubertakter 8y ago

I don't think any engineers deemed any available information about Columbia as "insignificant", not that there was very much information to go on due to lax execution by camera tracker teams. Management on the other hand was... not effective. I think that's really the crux of your statement. Humans seem to be terrible at working together in large organizations, or at least running them. Organizations seem to learn from mistakes only temporarily, or at all (as you pointed out). Consider the self-driving car group at Uber as a recent example.

woliveirajr 8y ago
Yes, and I agree with you. In my experience, even using some kind of protocol to prioritize which problem should be taken care first, it's easy to miss really important things because someone thought that "time to repair" should give the same score as "severity of consequences", and then we begin to fix things that are "easy and fast", not that are "relevant"

8bitsrule 8y ago

"But, as usual, we barely learn from mistakes... " As in, for example, public warnings by countless actual hackers back in the 80s and 90s about how vulnerable most US systems were. (For which, in return, they were persecuted and prosecuted. With no apologies as said warnings came to deserved fruition.)