Former Tesla Self-Driving Leaders Challenge Elon Musk's Optimism On Autonomy

Two ex-Tesla executives who spearheaded the company's self-driving efforts have publicly contradicted CEO Elon Musk's bold claims of imminent full autonomy, underscoring persistent challenges in achieving reliable robotaxi technology. 

In separate interviews, Andrej Karpathy and Sterling Anderson, key figures in Tesla's Autopilot and AI programs, highlighted the "march of nines" in reliability gains and safety gaps, painting a more cautious picture than Musk's repeated end-of-year promises.

Karpathy, who led Tesla's AI and Autopilot teams from 2017 to 2022, described autonomy as an exhaustive process requiring exponential effort for each additional "nine" of reliability (e.g., from 90% to 99% success). 

He praised Tesla's scalable vision-based approach but warned against overhyping demos, noting that early progress masks the grind ahead for unsupervised Full Self-Driving (FSD).

Anderson, Tesla's inaugural Autopilot director (2015-2016) and now GM's global products head, took a sharper jab during GM's Level 3 autonomy reveal for 2028. He advocated building trust through incremental safe deliveries, implicitly critiquing Tesla's hype-driven timeline.

Musk has forecasted unsupervised FSD in consumer vehicles and robotaxis by late 2025 since 2016, with hardware purportedly ready in all Teslas since then. Yet, Tesla faces lawsuits over Autopilot crashes, and even Musk's lawyers have dismissed his predictions as corporate puffery.

As Tesla eyes Q2 2026 Cybercab production, former leaders' grim assessments suggest years more for Level 4-5 viability.

Also read: Euro NCAP 2026 Protocol to Penalise Cars Using Touchscreens for Essential Functions