View attachment 990958
A solid update. Improved latency at no cost. A few 10% improvements, a few of these a year and it adds up.
Imo it seems that Tesla have been very silent about V12. No AI day announced yet, they have been held in September and October in the past. Don't think it's because they have run into some wall and cannot improve and have nothing to show. Clearly Optimus was moving forward fast and V12 demo was impressive even with one intervention(that should be fixable with some data). And Dojo should have been ramping up and H100s added to seriously improve compute. So I think that they are intentionally choosing to be quiet at this point. The question is why? Maybe it's because they are very close to actually solving FSD? The demo showed almost solving it and since then the V12-branch has likely gotten a lot more compute and priority of the developers. The last release notes seems more like project vacation improvement to the less beta version that customers use right now, that they need to maintain until the next V12 release goes from beta to public wide beta in a few months.
And if V12 actually is FSD(not robotaxi, but A->B without interventions 99% of the time and safer than humans), then the value of this will dwarf 2023 earnings.