GM exits robotaxi market, will bring Cruise operations in house

Original Article ↗ Hacker News Discussion ↗

GM’s decision to stop funding its Cruise robotaxi unit and fold the technology into in‑house driver‑assist features is seen as both a cost-cutting move and an admission of how hard and expensive urban robotaxis are to scale. Commenters contrast Cruise’s retreat with Waymo’s growing paid services and Tesla’s camera‑only FSD approach, debating which technical and business models, if any, can reach safe, profitable autonomy. The thread also highlights broader skepticism around consumer demand for paid “self‑driving” options, the drag of the dealership model, and the likelihood that only a few deep‑pocketed players will survive in this space.

GM’s decision on Cruise

  • GM will stop funding Cruise as a standalone robotaxi business and fold its tech into in‑house driver‑assist systems.
  • Reported spend was on the order of $1–2B/year; GM stock rose after the cut, implying investors didn’t value Cruise highly.
  • Many see this as finance-driven short‑termism and an admission GM is “not cut out for the future”; others note the huge cost, slow scaling, and regulatory backlash after Cruise’s pedestrian-drag incident and alleged misrepresentation to regulators.

Robotaxi business viability

  • Several comments question robotaxi unit economics: expensive sensor stacks (~$100k historically), large fleets, and long timelines versus uncertain fares and margins.
  • Waymo is viewed as the only US robotaxi at scale (multiple cities, paying rides), sustained by Alphabet’s deep pockets.
  • Some argue the market (taxis, trucking, aging drivers) is ultimately huge; others say big automakers exiting (Ford/Argo, GM/Cruise) signals the business case is weak or distant.

Tesla, Waymo, and other AV approaches

  • Large, polarized debate:
    • Pro‑Tesla side: FSD v12/13 is described as “remarkably good,” end‑to‑end, vision‑only, improving rapidly with massive data and compute; claim Tesla is structurally better positioned to scale globally and may soon challenge or surpass Waymo.
    • Skeptical side: FSD is still Level 2, requires constant supervision, makes dangerous mistakes (red lights, pedestrians, odd intersections), and is far from reliable enough for unsupervised service; Waymo’s Level 4 robotaxis are already operating safely for the public in several cities.
  • Mercedes’ limited Level 3 system is cited as legally “hands‑off but very constrained”; arguments over whether narrow, certified L3 is better than broad but supervised L2.
  • Some expect only a few winners (often named: Tesla, comma.ai, Waymo); others think the entire full‑robotaxi vision may be overhyped.

Dealerships, subscriptions, and adoption

  • Multiple anecdotes of GM/Ford dealers:
    • Salespeople poorly trained on Super Cruise / BlueCruise and other tech.
    • Features locked behind paid subscriptions and apps; sometimes disabled on demo cars.
  • Many criticize the US dealership model as rent‑seeking, hostile to EVs and advanced features, and misaligned with automakers and customers.
  • Tesla’s direct sales, fixed pricing, and app‑centric service are frequently praised, though some report long repair waits in earlier years.

Safety, regulation, and ethics

  • Strong disagreement about acceptable risk:
    • One camp emphasizes that humans already cause massive road carnage; AVs don’t need perfection, just to beat human rates.
    • Another insists AVs must be much safer than average humans and are currently far from proven, especially outside narrow geofences and ideal conditions.
  • Concerns about “beta‑testing on public roads,” deceptive incident reporting (especially for Cruise), and Tesla shifting blame to drivers versus Mercedes assuming liability within L3.

Broader attitudes toward autonomy and cars

  • Some think interest in “self‑driving” add‑ons is cooling: expensive L2.5 features deliver little value if drivers must stay fully alert; many cancel subscriptions.
  • Others say true unsupervised autonomy (read a book, sleep, send kids alone, drunk rides home) would be transformative and still worth pursuing.
  • Side discussions touch on:
    • Whether more autonomy means more cars and traffic versus more efficient fleets.
    • Environmental trade‑offs of keeping old ICE cars vs buying new EVs; lifecycle CO₂ math is debated and left unresolved.