3 ms·
> I agree, but I think their approach is flawed. A car is more than an AI with wheels and a seat. Tons and tons of engineering expertise goes into making a road
by oddevan 9y ago
> I agree, but I think their approach is flawed. A car is more than an AI with wheels and a seat. Tons and tons of engineering expertise goes into making a roadworthy car, and even more goes into making a truly good car. Apple quite simply does not have automotive engineering experience, and the jump from computers to iPhones pales in comparison to the jump from consumer electronics to automobiles.
I agree with you, and I think Cook and Apple do to! The article just says they're focusing on "an autonomous system." When asked if they were going to commit to a full automobile:
> "We’ll see where it takes us," Cook said. "We’re not really saying from a product point of view what we will do."
In other words, they're not making a car, just the AI for it.
Granted, Apple rarely makes 'components'; they make Products. It doesn't really jive with the Apple we know to make part of something without making the _whole_ thing. But I think these comments from Cook are telegraphing that they are--at least for right now--just making a component.
edit to clarify: Based on this and previous rumors, I think the "Apple Car" project (that totally existed, no joke!) has been refocused on the autonomous AI project. Cook and Mansfield realized that a car was big and complicated, and the market was inevitably moving toward self-driving cars. In realizing that without the self-driving component, they wouldn't have a car at all, they refocused the project to see if they could even build such a component. Once they have that, they'll go from there.
- TheOtherHobbes 9y agoIt's a seductive argument, but how much of the AI R&D in an autonomous car would be 1: truly unique 2: transferrable to other projects? It seems to me domestic/concierge/robotic/secure IoT AI is a much bigger and uncontested market. What specifically does building an automotive AI give Apple that a more general robotics and/or concierge AI project doesn't?