5 ms·
A friend of mine wanted to show off his Tesla by making it come to the front of the restaurant from where he parked it. Like he hit a button and it was to drive
by seibelj 5y ago
A friend of mine wanted to show off his Tesla by making it come to the front of the restaurant from where he parked it. Like he hit a button and it was to drive up. It got stuck somehow and was diagonal in the row. He was like “ehh sometimes it doesn’t work.”
AI in general is very over stated. When it works it’s great, when it doesn’t (which is often) then you lose all trust in it.
- hiptobecubic 5y agoI think that's the wrong conclusion from this story. The conclusion I draw is that some companies (and Tesla in particular) don't appreciate the "last mile" when trying to apply AI breakthroughs to consumer products. "The plane lands safely 99% of the time" is an impressive demonstration and a completely worthless product, but if that's all you have then what can you do other than launch it?
- deltaonefour 5y agoExcept all self driving AI companies have failed to achieve reliable AI. Every single one. So it says something about AI more than companies.
- hiptobecubic 5y agoIt says that (driving) AI is a rapidly changing technology, pushed along by these companies. The definition of "reliable" here is also important. Usually this devolves into a series of moving goal posts ending with "if it can't do everything perfectly all the time then it's not real!"
- deltaonefour 5y agoIt's a fuzzy thing yes. But it is still clearly not safe to drive everywhere with AI. A claim was made by experts and leaders about AI how by now self driving cars would be ubiquitous. That claim was wrong. That does say something about AI. It is a counter claim against the initial claim.
- microtherion 5y agoThat's NOT a product you should be launching with planes and cars, considering what would be happening the other 1% of the time.