GM's big bet on driverless cars turns sour
GM’s troubled Cruise program is prompting renewed skepticism about the economics and feasibility of fully driverless cars, especially when compared to low-paid human drivers and existing transit options. Commenters argue that true Level 5 autonomy is an immensely hard, still-unsolved problem that was badly overhyped, with early promises outpacing both technical reality and regulatory readiness. Many suggest that proven measures like better public transport, stricter licensing, and urban design reforms may reduce traffic deaths and emissions more reliably than speculative self-driving fleets.
Economics and Business Models
- Many argue the unit economics of current robotaxis (expensive vehicles plus large support staff) cannot compete with low-paid human drivers; savings per ride are too small.
- Supporters respond that early-stage R&D is always uneconomic and costs should fall with scale and custom vehicles, but others counter that this reasoning has already fueled unsustainable burn rates.
- Debate over whether robotaxis can reduce the total vehicle fleet by enabling 24/7 shared use; skeptics highlight rush-hour peak demand and people’s desire for private, predictable space.
- Some think subscription or fleet-based robotaxis are plausible; peer‑to‑peer “rent out your own car” is widely seen as unrealistic.
Technical Difficulty and Timelines
- Strong consensus that full Level 5 autonomy in arbitrary environments is extremely hard and was badly underestimated.
- Several note that 90–95% of driving is “easy”, but the last few percent may be one of the hardest tech problems ever.
- Comparisons are made to Apollo-scale efforts; current staffing and investment may be mismatched to the ambition.
- Some think 30+ years away; others warn that past tech predictions have often been too pessimistic.
Safety, Urban Design, and Alternatives
- Some emphasize the potential to save tens of thousands of lives annually; others caution that we cannot assume self‑driving will outperform humans soon.
- Multiple comments argue that the fastest path to fewer deaths is better infrastructure: dense housing, transit, cycling, stricter licensing, and fewer cars overall.
- There is disagreement over whether self‑driving helps or hurts broader goals like congestion, emissions, and reduced car dependence.
Regulation, Testing, and Deployment Strategy
- Many see current systems (especially some high‑profile deployments) as clearly not ready for open roads, citing basic mistakes.
- Calls for dedicated, rigorous testing and certification regimes for autonomous systems, not just adapting human driving tests.
- Some suggest starting in easier environments (wide roads, campuses, resorts, low complexity cities) and progressively increasing difficulty.
Liability and Public Trust
- Legal exposure for automakers after fatal crashes is seen as a major barrier.
- Overpromising timelines and capabilities is viewed as undermining investor confidence and public acceptance.