The best is the last
The Lockheed Constellation was a lovely thing. TWA got the first one in late 1945, and started transatlantic service in 1946. It was the first pressurised airliner in widespread use, and Lockheed sold over 800 of them.
It was also pretty much the last great piston-powered airliner, because the future was jets.
The de Havilland Comet, the first production jet airliner, first flew just 4 years later, in 1949, followed by the Caravelle in 1955 and of course the Boeing 707 in 1957, which was the one that really changed the world, and capped a decade or so of frantic innovation and improvement.
During the same period, though, the US Airforce was building and deploying the Convair B36 'Peacemaker', which first flew in 1946 and was retired in 1959.
With six piston engines and four jets, the official slogan was 'six turning, four burning', but apparently the crews preferred 'two turning, two burning, two smoking, two choking, and two more unaccounted for'. Every time it was serviced the ground crews had to replace 336 spark plugs (which wasn't much fun given that it didn't fit in most hangers and the bases were generally as far north as possible, to be close to Russia). Undeterred, Convair then used this platform to design this truly remarkable thing.
Rather unsurprisingly, no-one bought it. Meanwhile, the airforce was shifting to B52s.
The point of this excursion into tech history is that a technology often produces its best results just when it's ready to be replaced - it's the best it's ever been, but it's also the best it could ever be. There's no room for more optimisation - the technology has run its course and it's time for something new, and any further attempts at optimisation produce something that doesn't make much sense. The Constellation was perfect - the XC-99 or the Bristol Brabazon just showed that pistons had run their course.
You can see another example of this trend a few generations earlier with these three ships, the Cutty Sark, the Agamemnon and the Thomas Lawson.
The Cutty Sark was one of the best of the 'clippers' - high-speed long-distance freight-carriers - and was just about as good as commercial sailing ships could ever be.
One day we sighted a vessel, a mere speck on the horizon, astern of us, and the way she came into view it was evident she was travelling much faster than ourselves. 'Bringing the wind up with her' was remarked on board, and that seemed the only feasible conclusion to arrive at and account for the manner in which she overhauled us. In a few hours she was alongside us, and proved to be the famous British clipper Cutty Sark, one of the fastest ships afloat. She passed us going two feet to our one, and in a short time was hull down ahead of us."
—Wool clipper crewman, 1879
But it was launched in 1869, the same year as the Suez Canal was opened (which was impractical for sail and especially clippers), and 4 years after SS Agamemnon, the first steam ship that was practical for trade between Britain and China. The Agamemnon was faster than clippers and carried three times more cargo, and as one captain said, 'steam ships carry the wind in their holds'. People stopped building clippers. Then, three decades later, the Thomas Lawson, like the Peacemaker, showed what happened if you took sail just a little bit too far. It went past optimisation into unscalable impracticality.
(Just for comparison, while the Cutty Sark achieved 17.5 knots with the right wind, the Emma Maersk can make 25.5, with half the crew, and (officially) carries around 150,000 tons of cargo where the Cutty Sark carried 1,700 and the Thomas Lawson 11,000.)
The development of technologies tends to follow an S-Curve: they improve slowly, then quickly, and then slowly again. And at that last stage, they're really, really good. Everything has been optimised and worked out and understood, and they're fast, cheap and reliable. That's also often the point that a new architecture comes to replace them. You can see this very clearly today in devices such as Apple's new Macbook or Windows 'ultrabooks' - they've taken Intel's x86 and the mouse and window-based GUI model as far as they can go, and reached the point that everything possible has been optimised. Smartphones are probably at the point that the curve is starting to flatten - a lot has been optimised but there's still work to do, especially around cameras and battery life, and of course GPUs for VR. That curve will probably flatten out just at the point that AR starts to start shipping.