Tesla has been improving its FSD features since introducing it in 2016
Tesla
Previously, the Tesla X account shared a video where FSD V12 appeared to consider a pedestrian’s hand gestures when navigating. Elon Musk also confirmed on X that hand gesture recognition would be improved in V12.4, and even better by V12.5.
Tesla has had plans for gesture recognition for many years, with Musk mentioning hand gesture recognition in 2021. FSD is supposed to be able to adapt to new and unique circumstances in a safe manner, whether following instructions from a police officer, or understanding the hand signals from a cyclist.
Boris Johnson and FSD
Boris Johnson, the ex-Prime Minister of the UK, recently got to experience FSD V12 in the crowded streets of LA, with his wife and child in the backseat. Tesla provided a vehicle and assistant for his self-driven experience through 5-road intersections, heavy traffic, and pedestrians.
At the end of about 45 minutes I feel like a driving test examiner – except that I want to tell the car that it has passed, with flying colours. -Boris Johnson
Gestures coming soon
Johnson had a single experience with hand gestures in his drive – being waved at a hotel. While the car didn’t recognize the gesture – even on FSD V12.3.6 – the Tesla official in the vehicle confirmed that “‘It doesn’t yet recognise that gesture, but we are fixing it for the next iteration. It should be done next month.”.
It’s not immediately clear whether the Tesla employee is referring to FSD v12.4, which is expected to start rolling out to the public later this week, or Tesla’s next major FSD release, V12.5. According to Musk, V12.5 is expected to handle much more complicated situations and even start vehicle-to-fleet communication.
Yeah, I wouldn’t count on this in 12.3.x, but 12.4 is a big improvement and 12.5 might nail it
When you’re tackling a problem like FSD, which has never been solved before, you sometimes have to pivot and realize there’s a better approach. That’s what Tesla has done with FSD and AI. Just a few years ago where people thought cars would need to be able to read signs to achieve autonomy, but with the latest approach, there’s no need to. As long as the vehicle has seen a specific sign before and has enough examples, it’s already trained on what humans do when such a sign is found. The same will be true for hand gestures, the vehicle won’t need to know what each hand gesture means, it’ll just recognize that when a current gesture is made, this is how the vehicle should react.
Tesla does this by feeding it millions of examples of very specific situations and AI recognizes patterns. This is an example where Tesla leverages fleet data. They can easily capture millions of examples of hand gestures and how humans reacted, and feed that to their AI training model.
As Tesla processes more video, FSD will continue to improve. Tesla is expected to spend $10 billion on AI this year alone, most of it going toward improving FSD.
Subscribe
Subscribe to our newsletter to stay up to date on the latest Tesla news, upcoming features and software updates.
With Q1 2025 now behind us, another round of Tesla’s Supercharger Voting has wrapped up. The previous vote took place back in December, and we now get to see which locations came out on top.
Let’s take a look at the Q1 2025 winners of the Supercharger Vote:
North America
🇺🇸 Lake Wales, FL
🇺🇸 Henderson, NV
🇺🇸 Clarksville, TN
🇺🇸 Roswell, NM
🇺🇸 Long Beach, CA
🇺🇸 Decatur, IL
🇺🇸 McAlester, OK
Europe
🇬🇧 Bournemouth, United Kingdom
🇮🇹 Bari, Italy
🇵🇱 Wrocław, Poland
🇨🇿 Hradec Králové, Czech Republic
Asia-Pacific
🇰🇷 Namyangju - South Korea
🇦🇺 Nicholls, Australia
Supercharger Voting Q2 2025
Another quarter means another round of Tesla's Supercharger Voting, giving Tesla owners the chance to influence where new Supercharger sites will be built. Following the Q1 2025 voting round, Tesla is now opening the polls for Q2 2025.
How to Vote
To participate in the Q2 2025 Supercharger Voting, visit the Tesla Supercharger Voting page and sign in to your Tesla account. You can vote for up to five different locations, with a limit of one vote per location, every three months. The most popular Superchargers are displayed on the leaderboard, and you can also suggest new locations for future voting cycles.
We’ve known for a while now that Tesla has been using a custom build of FSD to allow its newly produced vehicles at Giga Texas and Fremont to autonomously navigate themselves from the production line to the outbound delivery lot.
While we knew they were using a custom build of what was likely FSD Unsupervised, thanks to a recent post from Tesla AI on X, this has now been confirmed. Tesla has also confirmed it has accrued over 50,000 driverless miles, totalled from vehicles autonomously driving themselves to delivery lots.
Giga Texas production now uses FSD Unsupervised to deliver cars from end of line to the outbound logistics lot.
Over 50,000 driverless miles have been accrued between California and Texas factories so far pic.twitter.com/79zKY0U6Ox
For most of Tesla’s vehicles - that’s a 1.4-mile trip that is shared with pedestrians, cars, trucks, and construction equipment. You can see in the video that the Teslas are navigating public roadways and encountering real human drivers.
That’s great news, especially since many were wondering whether Tesla would secure the necessary approvals in time to launch their Robotaxi network in June.
Increased Confidence
Following Tesla’s post to X, Musk followed up by saying that when Tesla launches FSD Unsupervised soon, it will be the first time there will be a generalized, pure AI solution to autonomy. Tesla and several executives continue to post more content about autonomy and the Robotaxi network, leading us to believe they’re feeling confident in the June launch.
Model Ys autonomously navigate a 1.4 mile trip on a road shared with pedestrians, cars, semi trucks, construction equipment & more pic.twitter.com/iPx2fs78v2
Back when Tesla initially announced the autonomous travel of its vehicles from production to loading lots, they mentioned that the Cybertruck was the only vehicle at Giga Texas not receiving that same treatment.
Now, that’s changed - Cybertruck is now also navigating through the Cybertunnel to make its way from the factory directly to the loading docks, all on Unsupervised FSD. And that’s making us even more excited, especially because Tesla was supposed to launch an FSD Update for the Cybertruck a little while ago - but it hasn’t made it to production yet. This update is set to bring Start FSD from Park, as well as Actually Smart Summon - bringing the Cybertruck to feature parity with Tesla’s other AI4 vehicles.
Cybertrucks autonomously navigate a 0.6 mile route traversing beneath one of America's fastest highways, emerging through a steep 17% grade to reach their destination pic.twitter.com/3ZMYCRPhIj
Now that Tesla is confidently using a build of Unsupervised FSD to navigate the tight confines of the tunnel and park, we’re pretty sure that Tesla will likely launch the expected FSD update in the near future.
FSD Update Soon?
It’s been a while since any FSD hardware variant has received an FSD update. It appears that Tesla has been focused on Unsupervised FSD and launching FSD outside of North America.
It seems like we may get a new FSD update soon, and we don’t believe it’ll be just for the Cybertruck. Those who have the opportunity to sign up for Tesla’s new Early Access program will likely be some of the first recipients to receive the FSD update, so stay tuned.
![Tesla’s Missing Voice: Why a PR Team Matters More Than Ever [Opinion]](https://www.notateslaapp.com/img/containers/article_images/multiple-models/group_93.jpg/4e1056961f0480c7b9eff43dd2ec288e/group_93.jpg)
