Raw LLM Responses

Those tests done by Don O'Dawd were fudged. He never provided a non tampered video evidence and it was apparent from what could be deciphered from the video that the driver in the car overrode the FSD in those problematic situations. He's also not a reliable source is information on this topic for many other reasons. Also, if you claim that the main difference between Tesla and Waymo is that Waymo uses lidar and Tesla doesn't, then you clearly don't have a complete understanding of those systems. The fundamental difference is how those systems "learn" to drive and then the secondary one is what infrastructure apart from the car and its sensors you need to operate them. Tesla could add those sensors to their cars and it still would be fundamentally different than Waymo because their whole approach to "learning" and how the car makes decisions is completely different and on completely another level than Waymo. And those people who say that you obviously need w lidar, they aren't in charge of the biggest project of this type in the world and don't work with hundreds of world class engineers and researchers on this topic. Is easy to say that lidar has it's advantages so obviously you should use it too - it's completely not obvious what you need in order to actually effectively use this new sensor and what you do when the sensors disagree for example. Also, lidar does have serious problems on its own. It does struggle in rain, fog and dust as it produces a lot of false positives in those conditions as all those little particles do reflect the laser. Also, yeah, version 14 is much better than 13 and yes it still makes mistakes but nobody claims it's a finished product or is ideal. The question is not of it makes mistakes (it will always do), the question is how many and how dangerous they are and what is the trend. Saying "version 14 still makes mistakes" is showing that you (unfortunately) don't understand how the progress works. It's the direction that matters and not that it makes some mistakes.