Apologies for using a metaphor. Let me try to clarify. Batches allow you to look at a bunch of companies at roughly the same point in time. So you now get to look at the companies against the backdrop of all the other companies in the same batch. Rather than having to make a one-shot decision about a single start-up you make a decision about which N of the 100's of applicants you like best which will give you a statistical advantage.

But it is possible that YC now has crunched enough data and has enough people in the alumni network that they can abandon the batch model and get the same kind of quality without having to use the batch model.

The 'late' applications were a hint in this direction, if they get rid of the batch model entirely they will have potentially a huge benefit both in load (batches will create temporary overload) and in total throughput.

So instead of having two 'strokes' per year they'd be moving at an ever accelerating pace until they run at the maximum rate at which viable applicants can apply and be processed.

To explain a bit of the metaphor, a piston has very discrete stages to produce work (at minimum, up and down) whereas a turbine is continuously spinning. YC's history has relied on 'stages' where all the companies are on the same path, they present at demo day, and the investors know exactly what to expect. If they turned into a 'turbine', the companies would all be at random stages, Demo Day would be much different, and YC would come to resemble more typical VC funds.