I agree. They also tend to discount the value of an ecosystem around various high profile electronics goods. Apple and gopro are both amazing examples of that, there are hundreds of accessories for every iPhone and every gopro camera. Then there's the value of a known workflow, of positive word of mouth and, finally, a feeling of community. People would much rather ask a friend who owns a product about some issue than hit the Internet to find a solution.
GoPro and Apple both also tend to bring killer features to market first. Be it TouchID or 4K 30fps for $600.
Also, the no-name cheap brand can become a known one pretty quickly. Brand is not a long-term defence to being undercut. It can give you some time to adjust to the new market being created, though.
People do want to shop on price, but they want a discount on the brand they know. That's why they'll buy a laptop for $20 less on Amazon than at Best Buy, but they probably won't buy a non-name brand laptop.
I don't even think his examples are compelling. There were action cameras before GoPro, there are alternative cameras on the market now. Some have different form factors, some are cheaper, some are better. At the end of the day, GoPro is plastered all over every action sport competition and their daily video email gets millions of views. That's harder to replace than the actual tech - especially when the product itself needs to be redesigned every 18 months.
Also note that "come in under the incumbent's price" is a bad startup strategy. The incumbent has economy of scale & bulk pricing advantages the startup can't match. The only way to do it is to build a different product that is so much cheaper to make that you can squeeze the same margins in a lower price tier. If that's possible, then it means GoPro has shot over what the market needs. In that particular case, I don't think it is true today. It might be true someday.
In AWS's case, I actually think the competitive play is on the higher end. Everyone is trying to undercut AWS's price (generally by giving it away for free). I feel that the last 20 years has shown trying to get under Amazon or Walmat's price is a bad strategy. How can anyone get under their price at smaller scale without taking a "strategic" loss or having 0 margins? I'd rather compete on the high side by offering better docs / UI & UX / support and not be competing completely on the price of hardware.