That quote's taken from the tl;dr-equivalent at the start of the article. It seems to be summarising the following, from further down, which I think is perfectl

That quote's taken from the tl;dr-equivalent at the start of the article. It seems to be summarising the following, from further down, which I think is perfectly reasonable: > Karl Brauer, an analyst with the auto research firm Kelley Blue Book, said the accident served as a signal that the technology might not be as advanced and ready for the market as some proponents have suggested. > "This is a bit of a wake-up call," Mr. Brauer said. "People who were maybe too aggressive in taking the position that we're almost there, this technology is going to be in the market very soon, maybe need to reassess that." I think that's perfectly reasonable, e.g. it may take 5 or 10 years to reach some kind of approved minimum safety level, especially since such safety approval standards/processes are still being developed. Just because Tesla's cars already have an "auto-pilot mode" doesn't mean the technology's ready for widespread deployment. Regarding "as close to 100% as you can get", that's not stated in the article. Instead, it says: > The federal traffic safety agency is nearing the release of a new set of guidelines and regulations regarding the testing of self-driving vehicles on public roads. They are expected to be released in July. > At a recent technology conference in Novi, Mich., the agency's leader, Mark Rosekind, said self-driving cars should at least be twice as safe as human drivers to result in a significant reduction in roadway deaths. "We need to start with two times better," Mr. Rosekind said. "We need to set a higher bar if we expect safety to actually be a benefit here." Again, I think that's reasonable. While a fully-automated road system would probably be safer than human drivers, we're not going to get an overnight switch; the harder problem is the transition, with manually- and autonomously-driven cars coexisting. It's not just a case of swapping out more and more cars, since self-driving cars will behave and interact in different ways (as an obvious example, looking at the occupants won't give you any information about intentions/gestures/etc.); hence it's worth waiting until a threshold higher than break-even is reached before beginning the transition.