[jaxbot]

If you read through Autonomy by Lawrence Burns, the author states that Levandowski felt the project was losing focus & getting too bloated, and wanted to disrupt it with a Team Macintosh style parallel project that he would head.

It could've just been part of the big ego'd power struggle that occurred on the project and saw the exit of many early stage employees, but I also wonder if he had a point, looking back today at the stages the project went through. It's hard not to feel like they thought they were close to launch, only to hit some roadblocks and need to rebuild major chunks of it. Speculation. But they're a huge company with a product that is at risk of getting usurped at this point.