Gary Zhou 
Elon Musk recently provided some updates on the wide rollout of FSD Beta 10.13. The latest iteration of the company’s advanced driver-assist system began its initial release earlier this month, but its wide release has not been initiated yet.
In a recent post on Twitter, Elon Musk admitted that while Tesla is working extremely hard on FSD Beta 10.13, the system itself is not ready for wide release just yet. He estimated that FSD Beta 10.13’s wide release would likely be a week or so away, but when it does, drivers outside California will experience some very notable improvements.
As noted by the Tesla CEO, users “outside of California will notice improvements the most.” This is an interesting comment, especially considering that Musk has admitted in the past that FSD Beta “seems to work better in California” than in other areas such as Rhode Island. Musk admitted to this last year, noting that FSD Beta has been overfitted to the Bay Area.
As noted in a Benzinga report, overfitting in statistics suggests that a model relates to a specific set of data points far too closely, which may not translate well with different data points. In the case of FSD Beta, this could result in the advanced driver assist system working very well in areas like San Francisco but not as well in other areas of the United States.
Release notes of FSD Beta 10.13 leaked earlier this month have revealed that the update includes a number of key performance improvements.
Following are the partial release notes of FSD Beta 10.13 that have been shared thus far:
- Improved decision making for unprotected left turns using better estimation of ego’s interaction with other objects through the maneuver.
- Improved stopping pose while yielding for crossing objects at “Chuck Cook style” unprotected left turns by utilizing the median safety regions.
- Made speed profile more comfortable when creeping for visibility, to allow for smoother stops when protecting for potentially occluded objects.
- Enabled creeping for visibility at any intersection where objects might cross ego’s path, regardless of presence of traffic controls.
- Improved lane position error by 5% and lane recall by 12%…
- Improved lane position error of crossing and merging lanes by 22% by adding long-range skip connections and a more powerful trunk to the network architecture.
- Improved pedestrian and bicyclist velocity error by 17%, especially when ego is making a turn, by improving the onboard trajectory estimation used as input to the neural network.
- Improved animal detection recall by 34% and decreased false positives by 8% by doubling the size of the auto-labeled training set.
- Improved detection recall of far away crossing vehicles by 4% by tuning the loss function used during training and improving label quality.
- Improved the “is parked” attribute for vehicles by 5% by adding 20% more examples to the training set.
- Upgraded the occupancy network to detect dynamic objects and improved performance by adding a video module, tuning the loss function, and adding 37k new clips to the training set.
- Reduced false slowdowns around crosswalks by better classification of pedestrians and bicyclists as not intending to interact with ego.
- Reduced false lane changes for cones or blockages by preferring gentle offsetting in-lane where appropriate.
- Improved in-lane positioning on wide residential roads.
- Improved object future path prediction in scenarios with high yaw rate.
- Improved speed limit sign accuracy on digital speed limits by 29%, on signs with difficult relevance by 23%, on 3-digit speeds by 39%, and on speed limit end signs by 62%. Neural network was trained with 84% more examples in the training set and with architectural changes which allocated more compute in the network head.
