Elon Says Tesla FSD 12.4 Roughly 5-10X Times Better on Miles Per Interventions

by

Elon Musk said that Tesla FSD 12.4 goes to internal release this weekend and limited external beta next week.

Roughly 5X to 10X improvement in miles per intervention vs 12.3.

12.5 will be out in late June. Will also see a major improvement in mpi and is single stack – no

Tesla Cybertruck FSD will come out with FSD 12.5.

13 thoughts on “Elon Says Tesla FSD 12.4 Roughly 5-10X Times Better on Miles Per Interventions”

  1. Went to a drone conference in San Diego. A developer of another autonomous driving system stated that the Tesla system was deeply flawed and therefore dangerous.

    His reasoning was that they focused only on visible wavelength optics. The obvious problems are snow, rain, dust. His assertion was that lidar, thermal, etc. composite systems were a necessity.

    • Whereas Tesla’s position is essentially that humans don’t come equipped with IR or Lidar, so it is categorically possible to drive at least as safely as a human without them, and safer, because computers don’t get tired. Which I find pretty persuasive, having driven all my life without access to anything beyond the visible spectrum.

      Once they get the self driving system up to full human equivalence, they can ADD non-human sensory modes, and take it up to superhuman.

      • Yes “version 2.0” could add thermal as well as a third visible light camera for the front for superior depth perception. Absolutely no reason why this all can’t be done at a later date and not necessary right now.

  2. Elon says a lot of things which just aren’t so. It’s his MO, so it is irrelevant to report what he said. Report what has been independently tested.

  3. If/when Robotaxi is near commercial, Tesla will stop selling cars.

    A 3 or Y that can make 50c/mi of profit and run a million miles is worth nearly $500k.

    Why would Tesla sell them for $50k?

    If Tesla is still selling cars, they don’t think FSD is close to ready.

    • Unlike everybody else besides BYD, they actually make money on every car they sell. There’s no reason to even slow down selling cars until they’re actually rolling out robotaxis.

      And even then, I don’t think it’s realistic to expect it to all happen at once. They’ll convert one production line, roll out in a couple markets, get approval in more markets, convert another line, etc.

      If they stopped selling cars and started converting all lines now they’d eat an enormous amount of capital, crush the stock price to dust, and have a lot less money available for AI compute. FSD does look close but that presumes they keep spending lots of money on compute, so their rapid progress continues.

  4. US Dollar 2,000 in a Single Online Day Due to its position, the United States offers a plethora bx04 of opportunities for those seeking employment. With so many options accessible, it might be difficult to know where to start. You may choose the ideal online housekeeping strategy with the vs-80 help of this post.

    Begin here>>>>>>>>>>>>>> https://richesdays21.blogspot.com/

  5. US Dollar 2,000 in a Single Online Day Due to its position, the United States nv-03 offers a plethora of opportunities for those seeking employment. With so many options accessible, it might be difficult to know where to start. You may choose the ideal online housekeeping strategy with the vs-90 help of this post.

    Begin here>>>>>>>>>>>>>> https://buildrichness04.blogspot.com/

  6. Let’s assume that indeed FSD will reach 5000 miles per intervention very soon. The kind of mistakes that FSD is making is way more severe than what a conscious drivers does. Therefore, full autonomy is still very far. The Musk and his followers are not aware of the full implication of a FSD strategy failure. That is said even before we ignore the reality that the market is not ready for a working FSD system.

    https://finance.yahoo.com/news/tesla-drop-70-robotaxi-ai-030107002.html

    • That used to be the case but most reviewers these days using 12.3 are saying their interventions mostly aren’t safety-related anymore. It’s more for preference, or to avoid annoying other drivers.

  7. Seems like FSD in Vegas Loop tunnels in this sequence will be able to pretty quickly transition from unsupervised in the tunnels to unsupervised Robotaxi ops near the tunnels extending the functionality of the system above ground. For example at Harry Reid airport, stadium, arenas, downtown. This would begin to demonstrate the potential of the system.

    • If Tesla can really do FSD in Las Vegas tunnels, they should replace their cars with the long-promised busses, or even small trains. Otherwise, the volume will never be more than a carnival ride amount.
      I believe FSD ought to at least work in a predictable enclosed tunnel. If FSD can’t even do that much, there’s no hope for it on complicated streets.

Comments are closed.