Tesla FSD to Get Navigation Improvements and Pothole Detection

Tesla is preparing to address two of the most common complaints regarding its Full Self-Driving (Supervised) software: inconsistent navigation routing and the system's inability to reliably avoid potholes. A principal software engineer on Tesla’s AI team confirmed that improvements for both issues are in development.

What Happened

The confirmation came via a direct exchange on X between Shopify executive Farhan Thawar and Phil Duan, a principal software engineer at Tesla. While Thawar praised the current FSD v14.3.10 build, he highlighted broken asphalt and stubborn navigation routing as his primary friction points. Duan responded by stating that "improvements on both are coming."

Pothole detection has been a long-requested feature, with Tesla initially promising it as far back as 2019. More recently, CEO Elon Musk indicated that pothole avoidance is officially coming soon, and the feature has been listed as an upcoming improvement in release notes since FSD v14.3 earlier this spring. Current builds have shown occasional glimpses of this capability, with one tester recording a vehicle on FSD v14.2.2.5 gently centering over cracked asphalt to allow the crater to pass between the tires. However, this behavior has been inconsistent, and it remains unclear if the neural net specifically recognized the hazard or simply reacted to a general road anomaly.

On the navigation front, Tesla is overhauling route-picking logic to better align with driver preferences. The new system will learn from daily habits by analyzing driver interventions, such as manual lane choices and steering nudges. If a driver consistently ignores a mapped detour in favor of a preferred highway exit or surface avenue, FSD will remember that preference for future trips. This builds upon the recently introduced Preferred Routes feature. Additionally, the system will study parking habits, noting preferences for reversing into driveways or hugging garage walls, to mimic these behaviors during autonomous parking.

Why It Matters

For owners, these updates aim to reduce the need for manual disengagement, particularly when FSD plows through tire-damaging potholes or takes confusing side-street detours instead of sticking to efficient commute routes. Reliable pothole avoidance would enhance vehicle safety and occupant comfort, while smarter navigation logic could significantly improve the usability of the system for daily driving.

The integration of these features also signals a shift toward more personalized autonomous driving. By using driver interventions as training feedback, Tesla is moving FSD away from rigid, map-based routing toward a system that adapts to individual user behaviors. Beyond these core improvements, Tesla is working on integrating Grok directly into the FSD stack, allowing drivers to give natural voice instructions for turns, parking, and route changes.

The Bottom Line

It remains to be seen whether these upgrades will arrive in an imminent point release or debut with the upcoming FSD v15 architecture, which is expected to bring a major neural network upgrade to 10 billion parameters. Regardless of the rollout timing, the addition of reliable pothole avoidance and habit-based navigation is expected to make daily autonomous commuting smoother for Tesla drivers.