I do agree with the first part of your analysis and that Bitcoin will eventually lose dominance (and perhaps number 1 spot). However I don't agree with your assessment that any Joe can start a valuable coin. I do think that many projects have inherent value and that offer / demand for tokens will be based in utility of them. I also think that Bitcoin itself will continue evolving (albeit at slower pace).
The point I was trying to make was that in crypto, coins have value because they are scarce. The original bitcoin dream was that there would be 21 million coins, and that's it. And most probably, there will never be more than 21 million bitcoins. But there will be many, many, many more crypto coins, so the market of crypto can be infinitely diluted. As to what is a valuable coin, almost ANY alt coin out there is technically a more valuable coin than bitcoin. Bitcoin's only real use case is the transmission of tokens, on a limited block size, one every 10 minutes, that consumes more electricity than a small European country, with a clunky piece of software, and with your transactions visible to any Joe out there in the world for ever. That's something that about any alt coin does better than bitcoin. Most of the "currency" alt coins improved upon the problems of bitcoin. It is amazing that one of the best improvements, namely stopping the waste of electricity, was experimented by Peercoin, which is now "out of sight". DASH and monero, and later zcash, and many others, attacked the problem of privacy and the potential problem of non-fungibility that goes with it (although the non-fungibility problem hasn't manifested itself in bitcoin, only in ethereum).
Then there are the "useful" tokens, that buy you a ticket for a service nobody has been waiting for: all the ICO app tokens. All these tokens can act like a currency coin (can be transmitted) but on top of that, they do something on an application that is, most of the time, pretty useless, but sound as if they are going to change the world, from blogging (steam and co) to prediction markets (augur and co).
Yes, at the moment, it still takes some dev. effort to make a plausible coin, even though just copying bitcoin's code, and changing some parameters is something a reasonable dev can do in one week's time (like Doge coin). The essence is simply to have a token that can be transmitted. Making a block chain with a token that can be transmitted isn't that hard. You can take the open source code of many stuff out there, change a detail, and then hype it as the next revolution (that last part is probably the most delicate one). You can do that an unlimited number of times, which makes the supply of crypto unlimited. You can also fork existing chains an unlimited number of times.
So even if there's scarcity on a *given chain branch*, there's no scarcity in the crypto market. You simply need to make believe people that only a very limited number of chain branches is "real", or the whole market will be infinitely diluted. But there's nothing that distinguishes the "real" branches from others, apart from some brand name. All essential functionality (and the only real functionality is transactions) is present on all coins, on all branches of all forks of coins.
It still takes some dev effort, like it took some knowledge of HTML in the 90-ies to set up a web page. But I don't see why there won't be soon "tools to make a coin", or even better, tools to make a fork (say, of bitcoin), so that you don't even need to be a dev. Say, the "dreamweaver of crypto".