3 ms·
> It was genuinely good. You lack data to draw this conclusion. The most important factor is deaths per mile, which is sparse, so it requires aggregating data
by energy123 1y ago
> It was genuinely good.
You lack data to draw this conclusion. The most important factor is deaths per mile, which is sparse, so it requires aggregating data from many drivers before you have enough statistical power.
- terminalshort 1y ago[flagged]
- energy123 1y agoData is needed to know if it's safe. For self-driving software, safety is the #1 criteria for whether it's a good product.
- ryandrake 1y ago"Trust me, bro, it really works" is these guys' slogan. It's like when you hear of a fad diet, and all the evidence for the diet's effectiveness is anecdotal "It totally worked for me!" testimony.
- estearum 1y agoWhere "good" is pretty much identical to having a specific rate of accidents over a specific set of conditions, then in fact data is absolutely necessary to determine that. Lol.
- vessenes 1y agoQualitative reactions matter. I didn't say it was genuinely safe. I can say that it was smooth, reacted better than most adult drivers to the situations I witnessed, and was significantly better at those things than the prior version. Safety is sparse, but steering, acceleration, decision making happen at like 60hz, and I would say I can make a qualitative judgment on it.