Connect with us

News

Tesla releases FSD Beta 10.4 with improved emergency vehicle detection network

(Credit: Whole Mars Blog/Twitter)

Published

on

Tesla released FSD Beta 10.4 this weekend. The latest FSD Beta release included improved emergency vehicle detection, which has caused tension between Tesla and the National Highway Traffic Safety Administration (NHTSA) in the past. As could be seen v10.4’s Release Notes, however, it appears that Tesla really worked on increasing its driver-assist system’s safety with the recent update. 

Below are the specific FSD V10.4 improvements that Tesla reported in its Release Notes. 

  • Improved handling when driving off navigation route by allowing better recovery, when safe to do so
  • Improved handling and detection of high speed objects when crossing high speed roads. Enabled faster acceleration across high speed roads. 
  • Improved speed through narrow spaces surrounded by high obstacles. 
  • Improved static obstacle control by upreving the generalized static object network with hyperparameter tuning and improvements for oversampling strategies (+1.5% precision, +7.0% recall)
  • Improved VRU detection (e.g. pedestrians, bicyclists, motorcycles) by adding data from next generation auto labeler (precision +35%, recall +20%)
  • Improved emergency vehicle detection network by adding new data and improved training regime (pass rate +5.8%)
  • Improved VRU control relevance attribute by adding navigation route as input to object detection network (accuracy +1.1%)
  • Press the “Video Record” button on the top bar UI to share your feedback. When pressed, your vehicle’s external cameras will share a short VIN-associated Autopilot Snapshot with the Tesla engineering team to help make improvements to FSD. You will not be able to view the clip.

One of the most notable updates in FSD v10.4 would probably be improvements to the emergency vehicle detection network. The improvement has drawn attention because of the NHTSA’s involvement in the feature’s release. The agency initiated a probe into Tesla Autopilot a few months ago, looking into accidents that reportedly involved the software. As the probe continued, Tesla released a free over-the-air update improving Autopilot’s capability to detect and respond to stationary emergency vehicles. 

The NHTSA criticized the update, asking why Tesla didn’t initiate a safety-related recall instead as per the agency’s protocols. However, Tesla did not release the update as a response to a defect. The emergency vehicle detection update seemed to be more of a proactive move on Tesla’s part.

Recently, FSD Beta’s emergency vehicle detection was put to work when a Tesla driver passed a police vehicle on the road. According to the Model 3 owner, FSD Beta detected the police cruiser pulled over at the side of the road and avoided the vehicle, giving the cop room on the road. His vehicle then safely merged back into the lane with a turn signal after giving the cop car ample space. 

Watch another demo of FSD Beta 10.4’s capabilities in the video below.

Advertisement
-->

The Teslarati team would appreciate hearing from you. If you have any tips, reach out to me at maria@teslarati.com or via Twitter @Writer_01001101.

Maria--aka "M"-- is an experienced writer and book editor. She's written about several topics including health, tech, and politics. As a book editor, she's worked with authors who write Sci-Fi, Romance, and Dark Fantasy. M loves hearing from TESLARATI readers. If you have any tips or article ideas, contact her at maria@teslarati.com or via X, @Writer_01001101.

Advertisement
Comments

Elon Musk

Elon Musk and Tesla AI Director share insights after empty driver seat Robotaxi rides

The executives’ unoccupied tests hint at the rapid progress of Tesla’s unsupervised Robotaxi efforts.

Published

on

Ashok Elluswamy

Tesla CEO Elon Musk and AI Director Ashok Elluswamy celebrated Christmas Eve by sharing personal experiences with Robotaxi vehicles that had no safety monitor or occupant in the driver’s seat. Musk described the system’s “perfect driving” around Austin, while Elluswamy posted video from the back seat, calling it “an amazing experience.”

The executives’ unoccupied tests hint at the rapid progress of Tesla’s unsupervised Robotaxi efforts.

Elon and Ashok’s firsthand Robotaxi insights

Prior to Musk and the Tesla AI Director’s posts, sightings of unmanned Teslas navigating public roads were widely shared on social media. One such vehicle was spotted in Austin, Texas, which Elon Musk acknowleged by stating that “Testing is underway with no occupants in the car.” 

Based on his Christmas Eve post, Musk seemed to have tested an unmanned Tesla himself. “A Tesla with no safety monitor in the car and me sitting in the passenger seat took me all around Austin on Sunday with perfect driving,” Musk wrote in his post.

Elluswamy responded with a 2-minute video showing himself in the rear of an unmanned Tesla. The video featured the vehicle’s empty front seats, as well as its smooth handling through real-world traffic. He captioned his video with the words, “It’s an amazing experience!”

Advertisement
-->

Towards Unsupervised operations

During an xAI Hackathon earlier this month, Elon Musk mentioned that Tesla owed be removing Safety Monitors from its Robotaxis in Austin in just three weeks. “Unsupervised is pretty much solved at this point. So there will be Tesla Robotaxis operating in Austin with no one in them. Not even anyone in the passenger seat in about three weeks,” he said. Musk echoed similar estimates at the 2025 Annual Shareholder Meeting and the Q3 2025 earnings call.

Considering the insights that were posted Musk and Elluswamy, it does appear that Tesla is working hard towards operating its Robotaxis with no safety monitors. This is quite impressive considering that the service was launched just earlier this year.

Continue Reading

Elon Musk

Starlink passes 9 million active customers just weeks after hitting 8 million

The milestone highlights the accelerating growth of Starlink, which has now been adding over 20,000 new users per day.

Published

on

Credit: Starlink/X

SpaceX’s Starlink satellite internet service has continued its rapid global expansion, surpassing 9 million active customers just weeks after crossing the 8 million mark. 

The milestone highlights the accelerating growth of Starlink, which has now been adding over 20,000 new users per day.

9 million customers

In a post on X, SpaceX stated that Starlink now serves over 9 million active users across 155 countries, territories, and markets. The company reached 8 million customers in early November, meaning it added roughly 1 million subscribers in under seven weeks, or about 21,275 new users on average per day. 

“Starlink is connecting more than 9M active customers with high-speed internet across 155 countries, territories, and many other markets,” Starlink wrote in a post on its official X account. SpaceX President Gwynne Shotwell also celebrated the milestone on X. “A huge thank you to all of our customers and congrats to the Starlink team for such an incredible product,” she wrote. 

That growth rate reflects both rising demand for broadband in underserved regions and Starlink’s expanding satellite constellation, which now includes more than 9,000 low-Earth-orbit satellites designed to deliver high-speed, low-latency internet worldwide.

Advertisement
-->

Starlink’s momentum

Starlink’s momentum has been building up. SpaceX reported 4.6 million Starlink customers in December 2024, followed by 7 million by August 2025, and 8 million customers in November. Independent data also suggests Starlink usage is rising sharply, with Cloudflare reporting that global web traffic from Starlink users more than doubled in 2025, as noted in an Insider report.

Starlink’s momentum is increasingly tied to SpaceX’s broader financial outlook. Elon Musk has said the satellite network is “by far” the company’s largest revenue driver, and reports suggest SpaceX may be positioning itself for an initial public offering as soon as next year, with valuations estimated as high as $1.5 trillion. Musk has also suggested in the past that Starlink could have its own IPO in the future. 

Continue Reading

News

NVIDIA Director of Robotics: Tesla FSD v14 is the first AI to pass the “Physical Turing Test”

After testing FSD v14, Fan stated that his experience with FSD felt magical at first, but it soon started to feel like a routine.

Published

on

Credit: Grok Imagine

NVIDIA Director of Robotics Jim Fan has praised Tesla’s Full Self-Driving (Supervised) v14 as the first AI to pass what he described as a “Physical Turing Test.”

After testing FSD v14, Fan stated that his experience with FSD felt magical at first, but it soon started to feel like a routine. And just like smartphones today, removing it now would “actively hurt.”

Jim Fan’s hands-on FSD v14 impressions

Fan, a leading researcher in embodied AI who is currently solving Physical AI at NVIDIA and spearheading the company’s Project GR00T initiative, noted that he actually was late to the Tesla game. He was, however, one of the first to try out FSD v14

“I was very late to own a Tesla but among the earliest to try out FSD v14. It’s perhaps the first time I experience an AI that passes the Physical Turing Test: after a long day at work, you press a button, lay back, and couldn’t tell if a neural net or a human drove you home,” Fan wrote in a post on X. 

Fan added: “Despite knowing exactly how robot learning works, I still find it magical watching the steering wheel turn by itself. First it feels surreal, next it becomes routine. Then, like the smartphone, taking it away actively hurts. This is how humanity gets rewired and glued to god-like technologies.”

Advertisement
-->

The Physical Turing Test

The original Turing Test was conceived by Alan Turing in 1950, and it was aimed at determining if a machine could exhibit behavior that is equivalent to or indistinguishable from a human. By focusing on text-based conversations, the original Turing Test set a high bar for natural language processing and machine learning. 

This test has been passed by today’s large language models. However, the capability to converse in a humanlike manner is a completely different challenge from performing real-world problem-solving or physical interactions. Thus, Fan introduced the Physical Turing Test, which challenges AI systems to demonstrate intelligence through physical actions.

Based on Fan’s comments, Tesla has demonstrated these intelligent physical actions with FSD v14. Elon Musk agreed with the NVIDIA executive, stating in a post on X that with FSD v14, “you can sense the sentience maturing.” Musk also praised Tesla AI, calling it the best “real-world AI” today.

Continue Reading