Self-Driving Cars
A prepared speech on why autonomous vehicles still make the same kinds of mistakes they were built to eliminate.
Photo by Roberto Nickson on Unsplash*
If Apple came out with a self-driving car, would it have Windows?
What is a self-driving car? A vehicle designed to be self-aware and make decisions on its own. You tell it "take me to work," and in theory it handles the rest.
So why do we need them? Because government data shows driver error accounts for 94% of crashes. The whole point of an autonomous vehicle is to take human mistakes out of the equation.
Except right now, they're making their own. A self-driving vehicle abruptly stopped on a highway once and caused eight rear-end collisions. Over an 11-month span, there were 400 crashes involving partially automated driver-assist systems, 273 of them involving Tesla.
The problem is that the algorithms underneath still aren't sophisticated enough. Roads are extremely complex, and every driver on them behaves differently. A computer has to make a decision in milliseconds, with no personality and no deep model of how humans actually communicate on the road. It's only ever as good as the data it was trained on.
Sergey Brin put it well: "We don't claim that the cars are going to be perfect. Our goal is to beat human drivers."
We're not there yet. There's a future where this technology is genuinely trustworthy, but it needs a lot more training, and a lot more testing, before it's ready for all of us to depend on it.
Feedback I got: strong opening joke, confident and composed delivery, clear speech, good video quality. Areas to work on: more vocal variety, and a few word-finding hesitations to iron out (recording myself on my phone was the suggested fix).