Google Has a Huge Advantage
Today was Tesla’s long-awaited “Investor Autonomy Day”. Man, what a snore-fest. Everybody on stage turned into Charlie Brown’s teacher squawking away about stuff that we all knew was a bunch of BS.
I got the distinct feeling that Tesla had something else planned but nixed it at the last minute. Where was the fancy video demonstration? Was it cancelled? Tesla was filming video with a Model 3 last week when they found themselves under surveillance from Twitter critic @skabooshka. Did his threat to expose any chicanery ruin a would-be demo reel? Live-streaming an autonomy event without video is like opening your steakhouse on a Saturday night without any beef.
Anyway, this event turned into a bunch of fluff with long boring hardware and software segments.
The big take-away for me was the fact that Tesla’s huge bottleneck in analyzing data is people. They literally pay people to identify objects in their gathered video. It’s expensive to pay people to circle objects and label photos one object at a time. For all of the tech that Musk brags about, they still do the important work by hand.
Google has a HUGE advantage over Tesla. Google literally has billions of people identifying objects for them 24/7 and they are doing it for free. Courtesy of captchas. All those password verification puzzles that you fill out asking you to identify busses, fire hydrants, bridges, store fronts, etc. There is a reason why you are seeing stuff on streets in those puzzles. It’s to help Google’s proprietary artificial intelligence grow. Google can probably identify every fire hydrant and bridge on every street in America in under a month.
After the 9/11 World Trade Center attack, the CIA started gathering data on a mass scale and found that the problem wasn’t in gathering the data. It was finding the time to sift through such a large mass. It was accumulating data far faster than they could read it. You can either reduce the time it takes to process the data or throw more people at it. Artificial intelligence isn’t good enough to replace humans in this regard yet.
To this day, the CIA hasn’t solved this problem. Tesla hasn’t either. Google has.
And it seems that Tesla's claims of imminent self-driving coming to a car near you is an annual event. Look at the dates of the headlines below. Geesh, what an embarrassment of a company. How can anyone take these guys seriously any more?
Anyway, this event turned into a bunch of fluff with long boring hardware and software segments.
The big take-away for me was the fact that Tesla’s huge bottleneck in analyzing data is people. They literally pay people to identify objects in their gathered video. It’s expensive to pay people to circle objects and label photos one object at a time. For all of the tech that Musk brags about, they still do the important work by hand.
Google has a HUGE advantage over Tesla. Google literally has billions of people identifying objects for them 24/7 and they are doing it for free. Courtesy of captchas. All those password verification puzzles that you fill out asking you to identify busses, fire hydrants, bridges, store fronts, etc. There is a reason why you are seeing stuff on streets in those puzzles. It’s to help Google’s proprietary artificial intelligence grow. Google can probably identify every fire hydrant and bridge on every street in America in under a month.
After the 9/11 World Trade Center attack, the CIA started gathering data on a mass scale and found that the problem wasn’t in gathering the data. It was finding the time to sift through such a large mass. It was accumulating data far faster than they could read it. You can either reduce the time it takes to process the data or throw more people at it. Artificial intelligence isn’t good enough to replace humans in this regard yet.
To this day, the CIA hasn’t solved this problem. Tesla hasn’t either. Google has.
And it seems that Tesla's claims of imminent self-driving coming to a car near you is an annual event. Look at the dates of the headlines below. Geesh, what an embarrassment of a company. How can anyone take these guys seriously any more?
RSS Feed