Validating Mistakes

How do we typically validate or rationalize our mistakes?  I've been thinking about this recently, as I've wanted to learn more about what I'm doing right and wrong in work/life.

Our human brains are wired to naturally blame outside factors: I was deceived, something changed unexpectedly, I had different constraints... or we even rationalize: my life is better because I failed, or the failure wasn't that painful.  But haven't we still failed?  Shouldn't we still learn from it?

Many of the decisions we have to make in life and work are "51/49" -- the choices are extremely close and we just have to choose one.  Either by factor of time, or others making the choice for us, or making the choice ourselves, the choice gets made.  But do we ever go back and evaluate what led us to that decision?  For very close choices like these, if they don't work out, it's almost easy to simply say, "it was such a close choice, we had to make a decision and move on".  But I almost always revisit these decisions and wonder what would have happened had I made the other choice -- perhaps a twisted and romanticized fantasy of what could have been.

I've taken to rationalizing my choices by blaming the data quality and availability.  I can say, "I made the best decision I could at the time with the data I had".  The fault of the wrong choice is no longer my own deficiency; the fault lies in the data being incomplete, or misleading.  This approach probably has its own faults: we all have human biases that make us overlook obvious data looking us in the face, as Nate Silver wrote in The Signal and The Noise.  But it at least forces me during each choice to consider whether I have enough data to reasonably rationalize later that I had all the [good-quality] data I could and made a reasonable choice given that data.

How do you validate/rationalize your mistakes?

Self-Driving Trucks and Beacon Navigation

If self-driving trucks are indeed going to be a mainstay of our shipping in the future, as is widely predicted, there have to be some rough edges in replacing human drivers.  A human driver can do things like take shortcuts, memorize a route taken many times before, and ask for directions if lost (human pride nonwithstanding.)

So how would a self-driving vehicle have the same adaptability?  It can 'see' where the road is and use built-in, offline navigation to navigate alternate routes, and if it has connectivity, phone home and ask for help or directions.  

But what if it doesn't have connection, is taking a road it doesn't really know or is on any map, and gets lost?  There are certainly many places like this even in the US, and certainly in other localities less documented.

What if, in these cases, major routes used for shipping that are outside of a good connection had small bluetooth beacons next to the road that communicated basic location/map information?  The vehicle could use it to find its way to the next beacon, kind of like crumbs through a forest.  These beacons could be powered by small solar panels, which would be more than enough for the small amount of juice needed to keep them running.

It could be a relatively expensive solution in the short term, as each beacon can run $10-20 per, and the range is maybe 10-20 meters at best.  But if they become significantly cheaper or have much wider range, you could run them along major delivery routes not really on any map (and be used by navigational devices).

To illustrate how this could work, I give you... a double rainbow: