Apple dials back car's self-driving features and delays launch to 2028

State of self‑driving and the 2030 “Level 5” bet

• Thread frequently references a public bet about Level‑5 autonomy by 2030; many now think the “for” side is more plausible due to Waymo’s progress.
• Others argue Level‑5 everywhere/all conditions by 2030 is still extremely unlikely, especially for privately owned cars vs geofenced robo‑taxis.

Debating SAE Level 5 and real‑world conditions

• Multiple comments quote SAE J3016: Level 5 = operates anywhere a typical human can drive, with exceptions for extreme conditions (whiteouts, floods), where system should pull over safely.
• Major contention: what “everywhere in all conditions” means:
  • Strict view: must handle snow/ice, complex weather, ad‑hoc traffic control, evacuations, ferries, etc.
  • Relaxed view: needs to be as capable as a typical local driver; pedantic edge cases are goalpost‑moving.
• Several note that many human drivers themselves likely don’t meet such a high bar.

Waymo vs Tesla and rollout strategies

• Many riders report Waymo in Phoenix/SF/LA already feels like Level 4 robo‑taxis: no steering wheel use, blind passengers can ride.
• Disagreement whether remote human guidance and emergency‑responder override make it Level 3 or Level 4.
• Debate over Waymo’s geofenced, HD‑map approach vs Tesla’s “works anywhere” vision‑only strategy:
  • Pro‑Waymo: safer, more mature in limited domains, already operating commercially.
  • Pro‑Tesla: HD maps are a fragile crutch; Tesla’s approach scales better to arbitrary locations.

Boondoggle vs real progress

• Some call self‑driving a massive boondoggle driven by Uber‑era hype and valuation games; products are still niche and heavily constrained.
• Others counter with direct experience using Waymo regularly and see it as clearly real and rapidly improving, especially for fleets in good‑weather cities.

Apple’s car project, scope reduction, and timing

• Article’s claim that Apple is shifting from planned Level‑4/5 to “Level 2+” is seen as:
  • Evidence Apple is being realistic and trying to ship something closer to today’s mainstream driver‑assist.
  • Or evidence of repeated resets and a likely eventual quiet cancellation.
• Some argue delaying instead of canceling may be sunk‑cost behavior; others note Apple often iterates internally and only reveals products when confident.

What Apple could (or couldn’t) bring to cars

• Supportive takes:
  • Strong UX, integration, privacy stance, and logistics/manufacturing scale could matter as cars become “computers on wheels.”
  • Potential to offer a premium commuter EV or even a vertically integrated robo‑taxi service, leveraging mapping, sensors, silicon (M/R‑series), and CarPlay experience.
• Skeptical takes:
  • Car manufacturing is low‑margin vs Apple’s usual business; service networks, safety, and regulation are alien domains.
  • Fear of a “non‑repairable, subscription‑locked” Apple car with tight parts‑pairing and high repair costs, echoing some current EVs’ body‑repair horror stories.
  • Question whether Apple is too late vs Tesla, Chinese makers (e.g., BYD), and Waymo‑style fleets.

Vision Pro, VR, and Apple’s tech pipeline

• Vision Pro is discussed as a template: a moonshot, low‑volume, high‑price “step 1” to test spatial computing, sensors, and custom silicon.
• Some see parallels: Apple seeds enabling technologies (lidar, sensor fusion, R1‑style chips) across products years before a big play (as with M1 Macs).
• Others note Apple’s past moonshots (VR, cars, AGI) by high‑profile technologists often underestimate problem difficulty and may not reach mass‑market success.