One of the things that help Tesla's electric cars stand out from the competition is its Full Self-Driving software (which, to be clear, isn't truly autonomous and does require driver supervision). On Monday, Tesla launched an update to that software, FSD v14, which had launched earlier as part of Tesla's Early Access program and is now rolling out to all Tesla owners in the US.
What's new in Tesla's software?
Teslas can park themselves when they arrive at your destination -- and you can pick whether that's on the street, in a driveway or in a parking garage. The car will remember your preference for next time. While the software already allowed for a more basic level of parking, v14 improves on this.
Tesla also revamped the Speed Profiles feature. There's a new Sloth mode drivers can choose that slows the car down for ultra-conservative lane changes and leisurely cruising. More importantly, Speed Profiles now control your actual velocity, not just how aggressive your Tesla feels on the road. It's a major behavioral shift for FSD, and one that seems to have made driving feel a lot more intentional -- more like a person driving the car.
The navigation interface has gotten an update, too. It's now easier to swap driving modes and parking preferences on the fly, and you'll find a permanent Start Self-Driving button even when you're already moving. That's especially handy in Tesla models that don't have the turn-signal stalk -- though last week a retrofit option became available to fix that.
Some of the visual UI updates for the navigation interface included in Tesla's FSD v14.
Other improvements include better reactions to emergency vehicles, construction zones and even rogue road debris -- FSD v14 is supposedly a lot more decisive. Navigation routing is now baked into the AI-powered neural net as well, which means the software has learned to navigate around traffic. This should make the car more adaptable to real-time detours and general road chaos.
Early testers such as X user AI DRIVR say the car feels snappier, more decisive and more human with the update. But some testers flagged overly sensitive braking, odd decisions triggered by approaching headlights and one rather futile attempt to reach a destination that was behind a locked fence.
The elephant in the room (or not in the room): lidar
While Tesla is charging ahead with a camera-only approach to awareness of the world around its cars, not everyone in the EV business is on board with ditching lidar sensors. Lidar, short for "light detection and ranging," is a sensor technology that uses laser pulses to measure distances and create detailed 3D maps of the environment.
The CEO of competitor electric car brand Rivian, RJ Scaringe, shared some very pointed thoughts on lidar recently. In an interview with Business Insider, he called lidar it a "really great sensor" that "can do things that cameras can't." He also noted how affordable it's become for car manufacturers -- dropping from tens of thousands to just a couple hundred bucks. Rivian uses lidar -- like many other electric vehicle brands -- on its vehicles for a lower level of self-driving capability.
It's the latest in an ongoing debate about the benefits of lidar. Some, like Scaringe, believe more data produces better results. But Tesla and CEO Elon Musk argue that too much data can confuse the neural system, with lidar often conflicting with vision results, and FSD v14 makes a solid case for sticking with just cameras.
