Sensors capture light. Lenses can be entirely virtualized in software. Whenever very large sensors become cost effective I'm sure smartphone cameras will all be "lightfield" cameras and there will be no more lense bulge. Post processing will also become way more interesting: https://pictures.lytro.com/charmingtoad/pictures/795243/embe...
Lytro cameras use lenses, too. In fact they use a huge number of micro lenses (in addition to larger traditional lenses) to capture the "light field".
"Very large sensors" also seems to means something different to you than it does to me. I would not want a full frame sensor (which I would only consider large, not even "very large") on my phone even if it were cost effective and didn't require a large lens.
Once our sensors are dense/sensitive enough (it doesn't necessarily have to be very large) physical lenses will not be necessary. The Lytro link is about interesting post processing. Today's lensless cameras are very new and very limited but the concept is sound and the technology will get there.
Eventually we will be reconstructing high quality photos from raw sensor data, applying all manner of algorithms to stabilize and accumulate brightness, stylized blur that isn't even optically possible, no shutter or lense or any moving parts really, and the whole thing will happen on your smartphone.
It's been a long time since I looked into this stuff in depth so I may have some details wrong. You are incorrect about the importance of sensor size. For a high quality image I'd prefer a larger sensor over a small sensor with more megapixels. For one example where small dense sensors are worse look at noise in the image. As the pixel density increases each pixel is getting hit by fewer photons. This reduces the true signal so your signal to noise ratio goes up.
Based on the use of Foveon sensors and light field tech from Lytro I wouldn't count on lens less cameras showing up in a phone in the next five years. A decade wouldn't surprise me but it's hard to predict tech trends that far out.
As Fredphophile pointed out, you're fundamentally wrong about sensor size importance. There are only so many photons hitting a given surface area unit. Smaller sensor size means more noise. You can't fight physics.
I'm not particularly bullish on lensless cameras either. I hope I'm wrong, but I don't expect those to meet modern quality standards in a small form factor for a long time.
