Musk Shares Details on FSD Beta v11: Neural Nets to Be Used for Vehicle Control

Elon Musk says we may get FSD Beta v11.3 as early as this week

Not a Tesla App

Elon Musk has provided further guidance on Tesla's next major release of FSD Beta, v11.3. Last week Musk said that v11.3 would be ready in about 'two weeks' and would contain many major improvements.

We've heard these two-week estimates before, but it's reassuring that Musk is being specific with details on this upcoming release.

Now just seven days after his last tweet, Elon Musk is giving us more details on this significant upgrade.

Timeline

Musk said he expects FSD Beta v11.3 to start rolling out to some customers later this week, or next week at the latest. This matches up with his initial two weeks estimate a week ago.

Elon Musk's estimates are known to be overly optimistic, but given the number of details he's releasing, it sounds like Tesla may be close to releasing this next build.

FSD Beta v11.3 will likely roll out to Tesla employees first and then go out to select FSD Beta testers, possibly the original OG group. Since this is a major milestone and includes major improvements, expect a slow and gradual rollout. Although we may get our first glimpse into FSD Beta v11 this week, it may be several weeks or more before the majority of customers have access to this beta.

Neural Nets for Vehicle Behavior

A week ago Musk said this upgrade will include 'many major improvements.' Last night Musk revealed some additional details. He said there will be "many small things," one of which will be that Tesla will begin to use neural nets for vehicle navigation and control, instead of just vision.

Today Tesla uses neural networks to determine the vehicle's surroundings, where objects are, what they are, and their distances from the vehicle to create a 3D environment known as 'vector space.' With this information, the vehicle can then plan a path and navigate around these objects toward its destination.

However, based on Musk's comment, it sounds like Tesla is currently only using neural nets to determine its environment and not for controlling the vehicle. This means that how the vehicle behaves, how it finds a path, and how it moves is still a process that is coded traditionally.

In the same way that Tesla uses millions of images to determine what a stop sign or traffic cone is, it sounds like Tesla will now use a large number of examples to determine how to best control the vehicle in various situations.

That could mean that Tesla will take millions of quality examples of how to gradually accelerate or slow down, based on real driving behavior to determine how the vehicle should accelerate in different situations.

This could be applied to every driving characteristic such as turning, slow downing, driving around a parked vehicle, etc.

If Tesla starts leveraging neural nets to aid vehicle control we may soon see drastic improvements to vehicle behavior, making it much smoother and human-like.

New Features

Although Elon Musk didn't specifically mention new features coming to FSD Beta v11.3, there are several that have been talked about in the past that could show up in this major full self-driving update.

Reverse Creep

Reverse Creep has been a feature that has been talked about as far back as FSD Beta 10.13. This feature would allow the vehicle to go into reverse to move out of the way of danger or adjust its trajectory. Right now FSD Beta will only ever move forward, so this improvement would be a giant step toward achieving human-like behavior.

A good use case for this is when the vehicle moves forward for better visibility. There may be times when the vehicle sees a car coming after moving creeping forward. In these cases, it'd be smart to let the vehicle reverse back to its previous position if it is now in the path of traffic.

Navigating Without Map Data & GPS

In the past Musk also alluded to the fact that Tesla is working on the neural networks' ability to complete 'dead reckoning' navigation, which is navigating based only on inertial measurements such as speed, direction and wheel movement.

He gave underground parking garages as an example of where FSD would need the ability to navigate without GPS or map data.

The car will be able to do this by using its last known GPS location and then determine its future location using only a compass, wheel movement and speed.

FSD Beta v11

FSD Beta v11 has always been expected to be a big leap forward, and as we get closer, that hasn't changed. This update is expected to be a huge improvement to what is currently available to customers. Although Musk's timelines have usually shifted and features have typically taken longer than initially planned, it looks like we may be getting close to the next major release for FSD Beta.

Although we're still years away from true full self-driving, Tesla's mission inches closer with every update. 

Recently there was also a leak revealing some details of Tesla's upcoming cameras in hardware 4.0, which are expected to include a fan and heater for select cameras.

Tesla’s Optimus Robot Learns to Walk Without Vision [VIDEO]

Optimus Falls - but catches itself!

Not a Tesla App

Tesla recently showed off a demo of Optimus, its humanoid robot, walking around in moderately challenging terrain—not on a flat surface but on dirt and slopes. These things can be difficult for a humanoid robot, especially during the training cycle.

Daily walks help clear your mind pic.twitter.com/dUsW58trS6

— Tesla Optimus (@Tesla_Optimus) December 9, 2024

A Look Behind the Curtain

Most interestingly, Milan Kovac, VP of Engineering for Optimus, clarified what it takes to get Optimus to this stage. Let’s break down what he said.

Optimus is Blind

Optimus is getting seriously good at walking now - it can keep its balance over uneven ground - even while walking blind. Tesla is currently using just the sensors, all powered by a neural net running on the embedded computer. 

Essentially, Tesla is building Optimus from the ground up, relying on as much additional data as possible while it trains vision. This is similar to how they train FSD on vehicles, using LiDAR rigs to validate the vision system’s accuracy. While Optimus doesn’t have LiDAR, it relies on all those other sensors on board, many of which will likely become simplified as vision takes over as the primary sensor.

Today, Optimus is walking blind, but it’s able to react almost instantly to changes in the terrain underneath it, even if it falls or slips. 

What’s Next?

Next up, Tesla AI will be adding vision to Optimus - helping complete the neural net. Remember, Optimus runs on the same overall AI stack as FSD - in fact, Optimus uses an FSD computer and an offshoot of the FSD stack for vision-based tasks.

Milan mentions they’re planning on adding vision to help the robot plan ahead and improve its walking gait. While the zombie shuffle is iconic and a little bit amusing, getting humanoid robots to walk like humans is actually difficult.

There’s plenty more, too - including better responsiveness to velocity and direction commands and learning to fall and stand back up. Falling while protecting yourself to minimize damage is something natural to humans - but not exactly natural to something like a robot. Training it to do so is essential in keeping the robot, the environment around it, and the people it is interacting with safe.

We’re excited to see what’s coming with Optimus next because it is already getting started in some fashion in Tesla’s factories.

Follow

Follow

Follow

Is Tesla Close to Licensing FSD? GM Quits Cruise, BMW Praises Tesla

Not a Tesla App

In a relatively surprising move, GM announced that it is realigning its autonomy strategy and prioritizing advanced driver assistance systems (ADAS) over fully autonomous vehicles.

GM is effectively closing Cruise (autonomous) and focusing on its Super Cruise (ADAS) feature. The engineering teams at Cruise will join the GM teams working on Super Cruise, effectively shuttering the fully autonomous vehicle business.

End of Cruise

GM cites that “an increasingly competitive robotaxi market” and “considerable time and resources” are required for scaling the business to a profitable level. Essentially - they’re unable to keep up with competitors at current funding and research levels, putting them further and further behind.

Cruise has been offering driverless rides in several cities, using HD mapping of cities alongside vehicles equipped with a dazzling array of over 40 sensors. That means that each cruise vehicle is essentially a massive investment and does not turn a profit while collecting data to work towards Autonomy.

Cruise has definitely been on the back burner for a while, and a quick glance at their website - since it's still up for now - shows the last time they officially released any sort of major news packet was back in 2019. 

Competition is Killer

Their current direct competitor - Waymo, is funded by Google, which maintains a direct interest in ensuring they have a play in the AI and autonomy space.

Interestingly, this news comes just a month after Tesla’s We, Robot event, where they showed off the Cybercab and the Robotaxi network, as well as plans to begin deployment of the network and Unsupervised FSD sometime in 2025. Tesla is already in talks with some cities in California and Texas to launch Robotaxi in 2025.

GM Admits Tesla Has the Right Strategy

As part of the business call following the announcement, GM admitted that Tesla’s end-to-end and Vision-based approach towards autonomy is the right strategy. While they say Cruise started down that path, they’re putting aside their goals towards fully autonomous vehicles for now and focusing on introducing that tech in Super Cruise instead.

NEWS: GM just admitted that @Tesla’s end-to-end approach to autonomy is the right strategy.

“That’s where the industry is pivoting. Cruise had already started making headway down that path. We are moving to a foundation model and end-to-end approach going forward.” pic.twitter.com/ACs5SFKUc3

— Sawyer Merritt (@SawyerMerritt) December 10, 2024

With GM now focusing on Super Cruise, they’ll put aside autonomy and instead focus solely on ADAS features to relieve driver stress and improve safety. While those are positive goals that will benefit all road users, full autonomy is really the key to removing the massive impact that vehicle accidents have on society today.

In addition, Super Cruise is extremely limited, cannot brake for traffic controls, and doesn’t work in adverse conditions - even rain. It can only function when lane markings are clear, there are no construction zones, and there is a functional web connection. 

The final key to the picture is that the vehicle has to be on an HD-mapped and compatible highway - essentially locking Super Cruise to wherever GM has time to spend mapping, rather than being functional anywhere in a general sense, like FSD or Autopilot.

Others Impressed - Licensing FSD

Interestingly, some other manufacturers have also weighed into the demise of Cruise. BMW, in a now-deleted post, said that a demo of Tesla’s FSD is “very impressive.” There’s a distinct chance that BMW and other manufacturers are looking to see what Tesla does next. 

BMW chimes in on a now-deleted post. The Internet is forever, BMW!

Not a Tesla App

It seems that FSD has caught their eyes after We, Robot - and that the demonstrations of FSD V13.2 online seem to be the pivot point. At the 2024 Shareholder Meeting earlier in the year, Elon shared the fact that several manufacturers had reached out, looking to understand what was required to license FSD from Tesla.

There is a good chance 2025 will be the year we’ll see announcements of the adoption of FSD by legacy manufacturers - similar to how we saw the surprise announcements of the adoption of the NACS charging standard.

Follow

Follow

Follow