@Tesla_AI is the most hardcore AI team on earth.
A decade of hard work on to building best AI software and hardware to start @robotaxi service for the public. Today establishes the new era of physical AI dominance for Tesla.
So grateful to be part of this mission!
Super congratulations to the @Tesla_AI software & chip design teams on a successful @Robotaxi launch!!
Culmination of a decade of hard work.
Both the AI chip and software teams were built from scratch within Tesla.
Eight years ago, I joined the @Tesla_AI team as an engineer. Infact, my bond with the company started even earlier, back at NVIDIA, working on chip bringup for Tesla's early hardware computers. This decade-long journey has been the adventure of a lifetime. Thrilling,
FSD v13: gradually, then suddenly.
Hope you all like this new brain update! One step closer to ‘all human input is error’ with a button to start drive autonomously and experience the magic.
BREAKING: First ever Tesla Model Y robotaxi with no-one in the drivers seat spotted testing on public roads in Austin, Texas!
Tesla's new "Robotaxi" wordmark/logo is on the side of the vehicle.
Humbled to play a part in transforming the future of humanity with Tesla AI's Optimus robots! 🤖
We are building an end-to-end AI stack that enables these robots to interact with the real world. Join us in shaping the future at tesla.com/AI!
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
Dancing is a hard problem not just for humans, even for robots. It stresses timing, torque and latency budgets. @Tesla_Optimus nails it like a pro dancer with all of the AI inference running on-board in real-time!
FSD Supervised ride-hailing service is live for an early set of employees in Austin & San Francisco Bay Area.
We've completed over 1.5k trips & 15k miles of driving.
This service helps us develop & validate FSD networks, the mobile app, vehicle allocation, mission control &
This event was a great forcing function for the team to come together, build core features that’s required for safe driverless operation, can’t wait to ship them in V13 and future releases.
Reflecting on the We, Robot event, we had:
- 19 Cybercabs and 29 Model Ys driving themselves
- 1,300 trips transporting over 2,000 guests
- Continuous operation of over the 3.5 hours
- And every trip was perfectly safe!
Proud to be working with the insanely talented @Tesla_AI team pushing the frontier of physical intelligence, where you not only need to be able to drive the car smoothly and safely but also do that within extreme latency constraints, where every millisecond counts.
We have
the delicacy FSD 12.5 has with the steering wheel is seriously next level
hard to get across on video, but this is already smoother than any Uber I've ridden in
Super proud of the @Tesla_AI team, these releases are the result of hardcore engineering and major innovations in both training and inference stack.
Happy to see positive reception on all the features from the customers, this will only get better from here!
Eventful September,
I gotta hand it to the @Tesla FSD/AI team. When I first saw this list of goals for September, it looked like it would be tough to meet them all. But with the launch of FSD on the Cybertruck tonight, they've officially hit every single September target.
While some Sept items
The @Tesla_AI team is hustling hardcore every day and night to get all these features out in the shortest time possible!
It's going to be nutty next couple of months!
Due to popular demand, Tesla AI team release roadmap:
September 2024
- v12.5.2 with ~3x improved miles between necessary interventions
- v12.5.2 on AI3 computer (unified models for AI3 and AI4)
- Actually Smart Summon
- Cybertruck Autopark 📐
- Eye-tracking with sunglasses 🕶️
-
We are trying to solve probably the hardest combination of challenges in order to achieve autonomy.
- using just 8 cameras and low-cost autonomy cars
- running real-time at 36 FPS and sub-human photon-to-control latency
- consuming just 100W of power
- real-world AI inference on
An accurate assessment.
What is also mindblowing is that the inference compute power needed for 8 cameras running at 36FPS is only about 100W on the Tesla-designed AI computer. This puny amount of power is enough to achieve superhuman driving!
It makes a big difference that we
