Lyft's business model is betting that automation will happen at the OEM level, and thus has a slower corporate burn rate. So less of their balance sheet goes direct towards automation efforts. Contrast this with Uber's significant losses in which the company's long term survival is contingent on the automation developed in-house to be leveraged as their own. I'm not entirely sure how Uber's technology would be integrate with OEMs where most (GM/Honda's Cruise, Ford, and others) are developing their own level-3+ autonomy, and how their sensor arrays/networks would be physically installed on cars.

I personally wanted to see the patent lawsuit Google had brought against Kalanick to completion- much of the lidar and sensor technology at stake needs to be either licensed using FRAND terms or have each company using its own disparate systems. V2V may be a desired open standard but I'm less than optimistic about its targeted implementation date being a few years away because of this competition.