[rjf72]

Space is like programming in that mistakes don't tend to be glorious, but just simple and mundane. The most famous example of this is almost certainly the Mars Climate Orbiter. [2] This probe was successfully launched and made it all the way to Mars. It was supposed to be an orbiter but ended up getting too close and most likely burnt up in Mars atmosphere. The culprit? Ostensibly some of the most qualified scientists with extensive oversight, regulations, and all that. And they were brought down by one operational system using imperial units, the other using metric.

A more recent failure has been on Curiosity. While it's been hailed a great success, it's had some major issues fulfilling with what was one if it's primary objectives - drilling into surface level targets and starting to measure things from the inside. It became inoperable after a few superficial utilizations. The most likely culprit? A piece of debris got caught in a drill feed mechanism. It could've been fixed by a man on the ground in a matter of minutes, but there is no man on the ground of course. NASA did manage to get a 2 inch drill without the stabilizers in the mid part of last year [1], thought there have been no updates since.

In many ways this is why I think people tend to rather overestimate the value of probes and remote robotics in general. Robotics, even to do relatively simple things, is incredibly hard. Trying to do not so simple things on planets tens to hundreds of millions of miles away in conditions unlike anything you can perfectly replicate on Earth? Well, you better send a few copies and just hope that what goes wrong on one probe/lander/etc (because something will go wrong) works out on one of the others!

[1] - https://spaceflightnow.com/2018/06/05/scientists-resume-use-...

[2] - https://en.wikipedia.org/wiki/Mars_Climate_Orbiter