In order to test the difference in flavor/crust texture/handling between autolysing and not I ran the following test. Same recipe and rise conditions for each approach. Did the test twice.
Autolyse: Add 75% flour to water and mix. Allow to autolyse for 20 minutes. Add starter, mix, add rest of flour, mix, add salt.
Non-Autolyse: Dissolve salt in water, dissolve yeast in water, add 100% flour, mix.
Below are pictures. For the dough and the pizzas the autolysed one is on the right. The crust differences on the first trial were most pronounced so I'm showing those here. In the second trial the differences were smaller.
For the autolysed, I see a higher rise in the doughball and a more textured/bubbled/airy skin when handling. The autolysed crust exterior is more irregular than the non-autolysed. The non-autolysed crust interior shows more gumminess on the bottom than the autolysed. The flavor of both is the same.
My hypothesis of what's going on is simply that the autolysed dough is experiencing more advanced yeast activity. I'll bet if I let the non-autolysed dough rise another few hours it would look exactly like the autolysed dough.
Because starter cultures are so fickle, I'm not sure if this difference was caused by experimental error on my part, or if some aspect of the autolyse process is truly responsible.
After these tests, I didn't feel like autolysing was buying me anything special. I think Peter suggested it may be more helpful for those with regular home mixers that are less effective at forming gluten structure than the Santos mixer I am using.