I rode in one of the UK's first self-driving cars (www.theverge.com)

I rode in one of Wayve’s autonomous cars in north London — a tangible preview of planned Level 4 robotaxi pilots with Uber (targeting 2026) and broader UK trials fast-tracked by government policy. The vehicle (a Ford Mustang Mach‑E with a modest roof sensor box) handled narrow streets, cyclists, buses, roadworks and even a blind pedestrian edging into traffic, though its driving felt noticeably cautious and intermittent. A safety driver, emergency-stop button and a buzzer signaling autonomous control underscored that these are still supervised demos, not fully driverless commercial services. For the AI/ML community the ride highlights an important technical divergence: Wayve uses an end‑to‑end, embodied learning approach that aims to generalize driving behavior across cities, rather than relying on dense maps and rule-based logic (Waymo’s common strategy). Wayve’s global “roadshow” across 500 unfamiliar cities and tests in the Scottish Highlands exemplify efforts to improve robustness and handle real-world edge cases without per-city mapping. That generalizability promises faster geographic scaling and simpler deployment, but brings tradeoffs — more hesitant, human‑like behavior, potential friction with local driving norms, and ongoing trust, safety and regulatory challenges as systems move from supervised pilots to unsupervised service.