3 ms·

It’s really just a matter of time. Those that started out optimistic (when it was hyped) but are now saying that they don’t see self driving cars can do the com

by krm01 7y ago

It’s really just a matter of time. Those that started out optimistic (when it was hyped) but are now saying that they don’t see self driving cars can do the complex situations we humans supposedly can should have some patience. As the tech improves, and more data is used to train models, it’ll surely surpass humans in driving ability.

jacquesm 7y ago
If only it were that simple.
mark-r 7y ago
I have this argument constantly. Humans make mistakes too, so we will eventually have tech that makes fewer mistakes than humans. The problem is that humans have a unique ability to make sense of situations they've never specifically encountered before. No amount of training data will be able to make up for that, because there will always be situations so rare that they never made it into the data.
- pasttense01 7y ago
  But what really matters is the overall statistics: The self-driving doesn't know how to handle this one in a 10,000 situation. But in the other 9,999 of 10,000 situations the self-driving vehicle is equal or better. Thus the self-driving vehicle averages out to be a lot safer.
  - mark-r 7y ago
    Everybody says this, but they simply take it as a matter of faith. I don't see any statistics to back it up. The article itself points out that there simply aren't enough self-driving miles at this point to make a valid comparison.
wutbrodo 7y ago
This is what makes hype cycles so dangerous. Most people are too simple-minded to understand nuance, uncertainty, or cautious optimism. So you end up with either "this is terrible and worthless and impossible" or "this is incredible and game-changing and imminent". I've seen it happen for everything from SDC to politicians that my social milieu pin their wildest dreams on that end up disappointing them by being decent, competent and non-messianic. It's probably worth ignoring what most people think about topics like this, except to the extent that their insane thrashing affects the situation (votes, funding, etc).

by krm01 7y ago

jacquesm 7y ago

If only it were that simple.

mark-r 7y ago

I have this argument constantly. Humans make mistakes too, so we will eventually have tech that makes fewer mistakes than humans. The problem is that humans have a unique ability to make sense of situations they've never specifically encountered before. No amount of training data will be able to make up for that, because there will always be situations so rare that they never made it into the data.

pasttense01 7y ago
But what really matters is the overall statistics: The self-driving doesn't know how to handle this one in a 10,000 situation. But in the other 9,999 of 10,000 situations the self-driving vehicle is equal or better. Thus the self-driving vehicle averages out to be a lot safer.
- mark-r 7y ago
  Everybody says this, but they simply take it as a matter of faith. I don't see any statistics to back it up. The article itself points out that there simply aren't enough self-driving miles at this point to make a valid comparison.

wutbrodo 7y ago

This is what makes hype cycles so dangerous. Most people are too simple-minded to understand nuance, uncertainty, or cautious optimism. So you end up with either "this is terrible and worthless and impossible" or "this is incredible and game-changing and imminent". I've seen it happen for everything from SDC to politicians that my social milieu pin their wildest dreams on that end up disappointing them by being decent, competent and non-messianic. It's probably worth ignoring what most people think about topics like this, except to the extent that their insane thrashing affects the situation (votes, funding, etc).