R1 Gen2 Camera Hardware meets or exceed Tesla HW3 & HW4 Vehicles so some level of turn-by-turn local autonomy and summon is capable.
Basically, The Gen 2 hardware would not be the limiting factor; in many areas (resolution, redundancy, compute) it is actually more capable on paper.
The main “bottlenecks” would shift from hardware to data scale, model training, and certification/regulatory rather than sensors or TOPS.
Tesla still ahead today because Tesla’s HW4 is paired with a much more mature behavior‑cloning‑style stack and enormous real‑world data volume, so even though Rivian’s sensors and nVidia Orin‑based compute are strong, Tesla’s software is the current performance ceiling.
The bottom line: I’ve heard that the ambition for Gen2 went from “Point to Point” in early 2025 to “on ramp to off ramp” in the last several weeks. But there’s no reason Rivian R1 Gen2 and early R2 (No LiDAR) vehicle owners should be limited from local, turn by turn, point to point autonomy.
Is it crazy to be annoyed by this change of tune by Rivian leadership?