- Joined
- Jan 7, 2008
- Messages
- 5,790
- Reaction score
- 780
- Points
- 203
- Location
- Earth
- Website
- orbides.org
- Preferred Pronouns
- she/her
If you mean google cars, then that is plain wrong.Thats wrong - right now, current cars that are getting more and more production grade in their design (and do no longer look like early prototypes) are not worse than humans in their error rates. But also not better.
They only work in a relatively clean environment that is pre-scanned to absurd detail.
Ability for unconstrained navigation of the real world is one of the definitions of general artificial intelligence.
You'd have to get pretty close to that for planes to fly themselves with error rate similar to human pilots.
Depression have it's patterns. One of them is that depressed-suicidal people often try hard to avoid hurting other people's feelings, much less hurting or killing them.If you say "it doesn't make any sense why someone would bring down 150 people with them" you aren't thinking like a depressed person because you're applying logic to the situation.
The exception is if the depression is (or perceived to be) caused by other people. That can end up massively deadly.
---------- Post added at 23:38 ---------- Previous post was at 23:30 ----------
Computers as we know them suffer from GIGO problem - Garbage In, Garbage Out.Sure, failures will always happen, but a flight computer with multiple redundancy can decrease the probability of failure way lower than a human pilot can.
They don't make mistakes in their calculations, but they won't give you a correct answer from wrong figures.
But a human often can.
Because a human have a much greater map of the world than we currently can describe to a computer, and a pathfinding core that we don't really understand yet.
Summarily known as intelligence.
A human pilot can cross-reference multiple very noisy signal sources (eyes, ears, gauges n the plane, smell, "seat of your pants" and so on) and determine what is going on and how to get out of the situation.
A computer won't be able to solve an arbitrary sensor issue - if one gauge shows 10000 feet and the other shows 20000 feet, how would it know which is right?
The big problem is not to define the algorithm of flying a plane - that is trivial.
The problem is in determining where the plane is in the grand scheme of things, what state is it in, what works, what doesn't, what is going on, where to go, and so on.