Throwing out the fact that Southwest and jetBlue "pick and choose" their destinations is a bit of a red herring; after all, US Airways doesn't serve SBA or even PDX, while United flies to neither Jackson, MS nor Little Rock, AR. You'd expect any well-run airline to "pick-and-choose" its destinations so as to maximize the use of its available resources. Why would WN add service to ROA, for example, when they could serve RIC or GSO, for example? That's not to say they CAN'T make money in markets like SBA or ROA; after all, Southwest still serves smaller markets like MAF and AMA 25 years after deregulation allowed them to fly outside of Texas.
No one is going to fly extremely thin routes like Rochester, NY to Rochester, MN non-stop. Well, Legacy/Oneida/whatever might claim they're going to do it, but it just won't happen. There's only one major point-to-point carrier right now in the U.S. simply because the hub-and-spoke model works GREAT when the air travel market is essentially an oligopoly. But once competitors come in with deep enough pockets (and low enough costs) to break a hub carrier's pricing power, or if consumers simply refuse to fly for what a network carrier must charge to make money, the hub-and-spoke model starts to break down.
Take a non-stop route (using the above-mentioned ROA as an example) like ROA-CLT. The *lowest* round-trip fare on US (aside from the New Year's special) on the route is over $300! Most people would take the 3-4 hour drive instead. Meanwhile, on the heavily-traveled AUS-LBB route, Southwest wants a minimum of $85 round-trip or a maximum of $240. For SBA-SFO, the lowest fare on United is $170 round-trip. The lowest round-trip fare between capacity-limited SNA and OAK on WN is $72. Of course, if you don't have the flexibility of planning 14 days in advance, the UA fare goes up to $465 round-trip, so it's not surprising that someone might drive down to LAX or BUR to save a couple hundred dollars.
The funny thing is that routes like CLT-ROA would probably be a lot more popular with fares closer to $100 round-trip -- and that would be a yield of over 30 cents/mile.
I actually don't believe that WN, B6, FL, and the other low-fare carriers have gotten anywhere near their limits in size. There are dozens of cities which WN could serve profitably without ever having to enter a fortress hub or a congested/slot-limited airport like BOS, LGA, or DCA. It wouldn't surprise me to see WN at BOS in response to a "market opportunity," though -- delays at BOS have been way, way down with most airlines having reduced flights dramatically here. Or if United were to shutter its IAD hub, why not take the opportunity to bracket the DC area?
If Congress were to step in to "limit congestion," do you think that they would go after 737's or RJ's first? What would create more congestion -- 5 daily AA Eagle RJ's from ALB to ORD or 2 hypothetical WN 737's from ALB to MDW? 4 daily PVD-IAD roundtrips on RJ's or a couple of 737's? RJ's from EWR to BUF/ROC/SYR or A320's from JFK to the same cities? The gridlock at LGA stemmed from the influx of RJ's, not from low-fare carriers with 737's and A320's.
I'd argue that the response to further growth among the low-fare carriers will be continuing consolidation among the network carriers -- we're already seeing that with the code-share alliances. Smaller, more marginal hubs will continue to be scaled back or will lose hub status. I'd predict that in ten years, there will be three or four network carriers/alliances (possibly two) with relatively broad international reach and strong hubs in large cities in geographically suitable locations -- like ORD, ATL, DFW, IAH, DTW, etc. The travel market won't continue to support 6 large, high-cost, hub-and-spoke networks.