Tesla FSD beta video

AIDRIVR video for April 1. I’m looking forward to watching his real video of V11 through the Berkeley/Oakland hills when it finally comes out. Take a look at some of his older stuff for excellent commentary on what works and what doesn’t.



That was quite clever. And he gave a pretty accurate summary of some of the criticisms against Tesla’s approach. Had he underlay the voice over with videos of Teslas struggling, rather than Wayne and Cruise vehicles, it would have almost been a solid critique.

In the end, though, the video doesn’t really address - or even mention - the main reason why those companies have chosen their approach. No one thinks that a Level 4 system with lots of expensive sensors relying on LiDAR maps is better than the mostly vision-based Level 5 system Tesla wants to build.

It’s just that the latter system doesn’t exist today, and the former does. Waymo and Cruise vehicles do fail from time to time - dramatically since they do t have drivers in the car. But Tesla’s system doesn’t yet have the capability to even be allowed to fail that way - they aren’t permitted to operate without drivers anywhere yet, so their failures are merely disengagements that don’t stop traffic.

The Waymo/Cruise theory of the space is that we are many, many years from having software capable of driving a vision-based system sufficiently better than an attentive, sober adult driver that it would be allowed on the roads. Musk has been claiming that Tesla is a year or two away from having such a system for many years - the rest of the competition thinks he’s completely wrong.

A person who can’t drive might find it frustrating to wait for autonomy until a company LiDAR maps their route - but if it takes much longer for Tesla to finally “solve” autonomy, it still might be much better for them.

It certainly was. So was the date, April 1. :laughing:

The Captain


Yes - it was pretty obvious about halfway through what the joke was - that the things that they were pretending to say about Tesla were actually descriptions of what the author thought were the failures of Cruise, Waymo and all the other competitors in the AV space.

But the joke only works if the criticisms actually work - if Cruise and Waymo really are performing badly relative to Tesla.

The joke works fine in the context of his video history. He is often quite critical of Tesla’s efforts.

That’s why I wrote:


1 Like

They are 3 different things.

  1. A highway only autodriver, uses maps and GPS and vision and other sensors. Works reasonably well on the roads it supports.
  2. A vision only semi-autodriver, tries to do it on all roads, in all conditions, with varying levels of success. It’s “semi” because hands must remain on the wheel and attention must remain on the road.
  3. A full autodriver, using many sensors. But only on a completely custom vehicle that will never be for sale to the public. Might someday be designed into a “real” car (real = for sale).
1 Like