In the rapidly advancing realm of artificial intelligence, Elon Musk is embroiled in a unique talent battle as Tesla and his latest enterprise, xAI, vie for the industry's leading minds. The shift of Tesla's computer vision chief, Ethan Knight, to xAI, illustrates a wider movement of key personnel between major tech players. This highlights the intense competition for AI's brightest stars.
While initially poised to join OpenAI, Knight's pivot to xAI adds a twist to the narrative, with Musk downplaying the notion of poaching yet acknowledging the fierce rivalry for top talent. To counter this, Tesla is introducing even more compelling incentives to retain its pioneering engineers."
In response to hundreds of queries and comments on the situation, Musk downplayed the impact of Knight's departure, emphasizing Tesla's robust team of over 200 engineers dedicated to AI and autonomy. Many comments on X suggest Musk is poaching from Tesla as a way to force the company to give him more compensation and voting power. “Musk responded: I (obviously) don’t need (the) money personally, but am concerned about not having voting power that to influence Tesla to do the right thing.” He currently has about 13% of the company but wants 25% control.
Tesla's Countermeasures
Musk himself noted that Knight's decision was influenced by competitive offers from both xAI and OpenAI, hinting at the fierce recruitment landscape Tesla is navigating. Musk's comments highlight a broader narrative, portraying Tesla's ongoing acceleration in autonomy advancements despite the competitive pressures. He acknowledges the AI talent war as unprecedented in intensity, underlining the strategic importance of AI talent for the future of both Tesla and the broader tech industry.
Tesla is increasing comp (contingent on progress milestones) of our AI engineering team
Musk revealed Tesla's strategy to counter-recruitment efforts. The company is not just matching the lucrative offers from rivals but also innovating its compensation structure. By tying compensation increases to progress milestones, Tesla aims to foster a culture of achievement and progress.
The Underlying Talent War Dynamics
The movement of high-profile engineers like Ethan Knight highlights a deeper competition for AI supremacy. With xAI successfully enticing several Tesla engineers, concerns arise over potential conflicts of interest and the implications for Tesla's Full Self-Driving (FSD) technology. The talent war is not just about compensation but also about the direction of AI development. It reflects a broader industry trend where the demand for AI expertise far outstrips supply, leading to aggressive recruitment strategies and significant shifts in talent.
Tesla's ability to continue its pace of innovation in AI and autonomy will depend on its success in navigating this talent war, balancing the need to retain key personnel with the drive to push the boundaries of technology. As the battle for AI talent rages on, Tesla has to balance stability and the pursuit of groundbreaking advancements with its strategies and responses, shaping the future of autonomous driving.
Subscribe
Subscribe to our newsletter to stay up to date on the latest Tesla news, upcoming features and software updates.
Following a period of radio silence from Tesla on FSD updates, Elon Musk has finally hinted that progress is continuing behind the scenes on FSD and that “Several major improvements are incoming.” We’re pretty excited - it has been over 100 days since the last FSD update, and we haven’t heard much since then.
The latest versions of FSD were V13.2.8 for AI4 vehicles and V12.6.2 for HW3 vehicles, both of which were released in January of this year — almost four months ago. While development has been ongoing internally, many have been wondering what the next public release will be. Will it be FSD V13.3, or will we jump straight to FSD V14.
Decoupled Releases & Spring Update
Tesla has now begun pushing Early Access users the 2025 Spring Update without an FSD version change. This means that we can expect the next FSD update to likely be based on the 2025.14 branch.
It’s worth noting that Tesla can add improvements to FSD at any time - and sometimes they do make minor changes without incrementing release numbers - small flag changes in FSD’s software to address how it does a specific task, or what data is uploaded.
With all that said, we expect the Spring Update to begin going out to more of the fleet in the coming days. We’re currently seeing about 58% of the fleet on the Spring Update, and only 30% of the fleet on the older 2025.8 January Update.
Not A Tesla App received information that an upcoming update was set to bring features from other AI4 vehicles to the Cybertruck, including Start FSD from Park, Unpark, Actually Smart Summon, and more. This update was intended to bring it closer to feature parity with the rest of Tesla’s AI4 fleet, but for now, Cybertruck remains the redheaded stepchild of the fleet.
We’re still confident that Tesla is working on this, and the continued delays on the release of an FSD update could point to the Cybertruck and a lack of data continuing to be a pain point for Tesla’s AI team. Cybertruck owners, including the author, have noted that FSD-equipped Cybertrucks continue to upload several hundred gigabytes or more of data per month. This topped out at nearly 1.9 TB of data uploaded in April 2025 for the author.
That’s a massive amount of data - and other users on social media have mentioned much the same for their own Cybertrucks. Tesla needs as much data as possible to tune the FSD models, and given the small fleet size for the Cybertruck, it requires a vast amount of data per user.
“Major Improvements”
While Elon didn’t mention what constitutes these major improvements, we have a lot of expectations besides what we know about the Cybertruck. Learning from the recent and successful FSD launch in China, Tesla is now able to utilize a more generalized model without specific local training data. This could potentially translate into better performance in North America as well, as the 7.7 million miles globally driven on FSD every day are feeding back into Tesla’s data loop.
We’re hopeful that future improvements continue to focus on improving tracking and decision-making, as well as lane handling. FSD users on X continue to point out issues with lane selection and lane keeping in the latest versions of FSD. On the flipside, Tesla has greatly increased the comfort and smoothness of FSD - and V13 is a prime example of that.
While V13.2.8 is also capable of pulling into parking stalls both forwards and in reverse (thanks to one of those flag changes recently), it does an oddly poor job of parking. Tesla’s Vision Autopark, on the other hand, is exceptionally accurate, even with big vehicles like the Cybertruck. It feels like Tesla is working on the parking lot stack to prepare for the upcoming launch of Robotaxi in June.
What About FSD V14
Back in the Q4 2024 Earnings Call, we heard about FSD V14, and just learned a little bit of what will make it unique. In this case, it's auto-regressive transformers that will improve FSD’s already powerful perception system and help it to predict better how other vehicles and road users will behave around it.
That, alongside a larger model and increased context size, will help FSD manage edge cases and make better decisions. The larger model and context size increases are likely another challenge for Tesla, which is already pushing the hardware limits of AI4 with FSD V13.
We did a deep dive into what we know about V14, which you can read here.
The non-Launch Edition version of the refreshed Model Y began shipping just a few weeks ago. One of the few things you couldn’t get that the Launch Edition came with was the Acceleration Boost. Tesla wasted no time with this vehicle and is now offering Long Range AWD owners the option to purchase Acceleration Boost.
Tesla didn’t offer a Launch Edition version of the new Model 3 when it launched in 2024. While Acceleration Boost is now available for the new Model 3, it remains restricted to the Asian-Pacific market. It’s not clear why it’s not available in other regions, but it’s possible it could be due to battery chemistries.
Boost Now
Acceleration Boost is an option that has generally been offered for most Model 3 and Model Y Long Range Dual Motor variants, providing excellent performance for the price. A vehicle with Acceleration Boost places the vehicle somewhere between the regular Long Range variant and the Performance version.
In the Refreshed Model Y’s case, Acceleration Boost cuts the official 0-60 mph (0-100 km/h) time from 4.6 seconds down to 4.2 seconds, adding a hefty boost. This is achieved by unlocking approximately 100 additional horsepower from the drivetrain through software.
That 4.2-second time is also 0.1 seconds faster than the rated 0-60 on the original Model Y with Acceleration Boost, so it’s another generational improvement right there.
Cost & Availability
Acceleration Boost is priced at the standard $2,000 USD - and will eventually become available in most global markets for that equivalent price. The upgrade is available now in the United States and Canada ($2,700 CAD), and should hopefully be available globally within the next few days.
Owners can purchase Acceleration Boost through the Tesla mobile app from Upgrades > Software Upgrades > Acceleration Boost, or through the vehicle touchscreen from Controls > Upgrades, provided that In-Car Upgrades are enabled (set in the Tesla App).
The vehicle will need a network connection to process the configuration update, which only takes a few minutes. During the configuration update procedure, the vehicle will not be drivable and will reboot.
Eligibility
It is important to note that the Long Range Rear-Wheel Drive (RWD) variant, which is available in Europe and Asia-Pacific, is not eligible for this upgrade as it lacks the necessary dual-motor configuration.
Owners of the Launch Edition already have this feature by default and won’t need to purchase it.
No Track Mode
If you’re hungering for Track Mode and all the functionality it offers, we recommend waiting for Tesla to release the Performance variant of the Refreshed Model Y. While Elon did mention in the past that Acceleration Boost would grant access to Track Mode, it hasn’t happened quite yet.
Is It Worth It?
Acceleration Boost helps to bridge the gap between the Performance variant and the Long Range AWD variant, offering quicker acceleration without the need to opt for the dedicated and more expensive Performance variant.
This is especially so today, where the Refreshed Model Y Performance hasn’t been announced yet - so if you’ve already purchased a Long Range AWD and want to go just a bit faster, this is a fantastic option at a good value.
The difference in acceleration is noticeably faster, just like Acceleration Boost with the original Model Y. If you really want to take off a little faster or want to take the vehicle to the track, this will definitely be a worthwhile upgrade.
Tesla hasn’t noted a reduced range with this upgrade, similar to the refreshed Model 3’s Acceleration Boost. Older models had about a 5% range loss due to the additional always-available acceleration and cooling requirements.
This is one of Tesla’s most popular upgrades due to its relatively low cost while also adding some additional value to the vehicle.
![Tesla’s B-Pillar Recording Requires HW4 [Update: New Dashcam Viewer Included on HW3 / Ryzen Vehicles]](https://www.notateslaapp.com/img/containers/article_images/tesla-screen/dashcam-viewer-hw3.jpg/ba8281c93679d2cf4a405a771e96b167/dashcam-viewer-hw3.jpg)
![Tesla Semi Keynote: New Features, 46 Charging Sites, Upgraded Battery & More [VIDEO]](https://www.notateslaapp.com/img/containers/article_images/semi/semi-new-features.jpg/c7df4e0d32f8cb0beee080ad201e943b/semi-new-features.jpg)
