Banning loss leaders (a.k.a. market dumping) seems like an inherently attractive fix to improve free markets, but it's fraught with difficulty.
The most obvious problem is R&D costs. I do market research and decide that people would be willing to pay $100 for a widget. But said widget does not yet exist, so I spend a million dollars to develop it, and then start selling it for $100 a pop. I calculate it will take several years to break even but that's OK, because I'm a businessman who thinks long term and we like those sorts of people don't we?
I think you can see where this is going - the business runs at a loss for several years, to build the market and spread out the development costs. Eventually I can reduce the price of my widget because I paid off the R&D costs. But until then I'm still in the red.
Amazon is no different. If they make no profit, it's because they choose to charge low prices, build the market and develop new products all at the same time, instead of cashing out. Though actually I think you're distorting history by saying they "muscled their way into the market". Amazon was one of the first online stores. There was no market to muscle in to, nobody else was doing what they were doing. Bezos pretty much created a new market from scratch.