Hacker News new | past | comments | ask | show | jobs | submit login

Is it really such a moat? Probably other people are trying to build autonomous cars but have the sense not to talk about it.



The moat comes from the fact that I'm pretty sure Google's success with autonomous cars comes from their integration of AI and their maps infrastructure. The cars drive well because they have the data from cars that have previously driven the same route. This synergy between big data, infrastructure, algorithms, and hardware is extremely hard to replicate and is exactly the sweet spot at which Google is the best in the world.

Basically, I think the reason you hear about Google's cars and not other peoples' cars is because Google's happen to actually work. The constant thesis of Google itself has been that dumb algorithms and lots of data works better than smart algorithms. They've cracked many, many long-unsolved problems by tilting their thinking in this direction (examples: search, NLP, voice recognition, translation.) Autonomous car navigation is the next one, and in this case the data is extremely difficult to get. Who else has had cars driving around for the last 5 years collecting street-level data? Who else is even capable of building the systems to collect, process, and organize such data?

It's a long play though. I don't think you'll start to see GOOG react tangibly to the autonomous cars project for another 4-5 years or so. But my guess is that their success will be on-par with the iPhone in terms of generational leaps ahead and barriers to entry, if not more. If Google manages to get their technology into most major car manufaturers' vehicles there will be massive switching costs due to integration expense. As soon as one manufacturer has Google technology in their cars, provided consumer reaction is positive, they will all want it. Once they are in, they are in, and there will be recurring revenue via software updates and next-generation capabilities for newer vehicles. Hell, they could even have a service model, where manufacturers or consumers themselves pay a monthly fee for autodrive service.

Of course, their current valuation may or may not justify this, if you presume their EPS growth due to autonomous cars will be offset by low growth in their current cash cow, adwords. I honestly think this is a bigger risk than the risk that autonomous cars will not work out though. It's going to happen, and it appears they are very far ahead of everyone else.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: