My understanding is that having to manufacture a different rib for each station along the structure is a significant challenge. It's possibly less of an issue with composites?
The differences in ribs across the span of a wing are much less of an issue than compound curvature in the wing skins—this is where composite manufacturing really shines, because you can mold them into arbitrary shapes.
The expensive part of composite aerostructures isn't making them, per se, it's the molds and tooling you have to build first, and you need a lot more of those when contours are constantly changing. (This isn't new though - continously changing contours have been the norm since the first high supersonic wasp-waisted "area rule" fighters and bombers appeared in the 50s and 60s.)
That said, I was working on B-2 aerostructures in the late 80s, and I can tell you that most all the parts on that plane have no symmetry in any direction other than centerline bilateral. My group figured we could save over $10 million each on a single B2 duct, if we could change the bizarre geometry to simplify the scary complex tooling it required. That was enough potential savings to provoke a design review, but the answer came back, "Nope. It has to be that way (we presumed for stealth). Go figure out how to build it..."
Because the fuselage is pressurized. And it is convenient to have pressure vessels as cylinders for even distribution of of the pressure forces across it. Ideally you want to have it capped off with two hemispheres (think propane tank shaped) but obviously you can't do that for other reasons. So there is likely more design and manufacturing for the front and rear of the plane than the main section.
Ah, gotcha. Yeah, there’s zero chance that you can make a pressurized pancake using rib-and-stringer architecture, probably even out of unobtanium. The best candidate I can think of is making it like a submarine, although that would definitely mess with the open concept floor plan and probably impose a huge weight penalty (although it might be helpful from a fatigue life standpoint). The only other remotely viable solution I can think of is drop stitching, but that has some fairly obvious drawbacks that would probably make it impracticable.
Another problem is how planes are designed for different sizes.
Say you design your plane with the base / most ordered as the dash 9.
Now someone wants a smaller longer range one.
just take some fuselage barrel sections out. Boom dash 8.
Want an even smaller one. Take more sections out. Dash 7.
Need a huge massive super long range one with loads of seats.
Add sections and you get the dash 10 or 11.
Current planes use a uniform fuselage structure for this reason. You're going to make different models within the same model family with a flying wing.