What's hard to disrupt is their amazing logistics network for delivering impulse purchases on the cheap and fast.
But while I have nothing but contempt for sponsored products, enabling Amazon arbitrage like it's eBay, and for deleting keywords to increase the size of search results because that defeats my spearfishing for specific products, it doesn't seem like something someone else couldn't build.
Start by addressing all of the bad customer experiences with counterfeit products, badly shipped items, and lousy customer support. What makes this difficult is that (beyond the challenge of execution itself) I don't think traditional venture-capital would invest in this so you have a chicken and the egg problem that I suspect will never be solved until someone manages to come up with a business model better than Amazon's at a time when Amazon is too large and inefficient to respond properly.
And I think they're getting there. I worked there for six years and left recently. I'm never going back. The culture in my opinion is horrible. And that's a second area where you could do better than Amazon and brain drain them. I'm sure there's more.