CYBERCAB AND NHTSA.

 








GROK:


I’ll check the current NHTSA rules, exemption path, and any 2026 Tesla/Cybercab or no-controls vehicle actions before giving a probability.High for limited fleet use. Lower for “sell anyone a car with no wheel and no pedals, no extra paperwork.” The hardware question is mostly settled in principle. The fight now is how it is legal, at what scale, and with what strings. Tesla FSD mileage is not what decides the steering-wheel/pedal issue.


NHTSA does not pre-approve cars the way the FDA approves a drug. Makers self-certify to FMVSS. A vehicle with no steering wheel and no pedals collides with standards written for a human driver—especially FMVSS 135 (“service brakes shall be activated by means of a foot control”), plus mirrors, telltales, and related 100-series rules. So “approval” really means one of three things: a Part 555 exemption, a rule rewrite, or NHTSA accepting self-certification that those rules do not apply.


What has already happened

Amazon’s Zoox already got the first commercial exemption for a purpose-built, no-wheel, no-pedal robotaxi on July 30, 2026. Cap: 2,500 vehicles per year for two years, extra crash reporting, and the vehicles are not for retail sale to private owners. That is the existence proof: NHTSA will allow the category when the maker asks, documents equivalent safety, and accepts limits.


Tesla took the other path with Cybercab: self-certify, start paid rides in Austin, no Part 555 petition. NHTSA opened an audit the same day and, on September 10, issued a Special Order demanding sworn answers by September 30 on how a car with no foot brake can be certified to a standard that still requires one. The June 2026 proposal to drop the pedal requirement for ADS-only vehicles is not finalized yet. That is why Tesla is under oath and Zoox is not.


The 2022 occupant-protection rule already treated “no driving controls” as a real vehicle class for crashworthiness. The current DOT is separately proposing to strip leftover driver-equipment rules (pedals, and earlier wipers/defoggers/placards). Direction of travel is not ambiguous.


Probabilities if you ignore FSD quality

These are judgments about controls and FMVSS, not about whether unsupervised FSD is good enough.


| Outcome | My probability | Why |

|---|---|---|

| NHTSA continues to allow some no-wheel, no-pedal vehicles in commercial service | ~95% | Already granted to Zoox. Walking that back would take a political and legal shock. |

| Final rule removing the foot-brake / manual-control mandate for ADS-only vehicles within 18 months | 75–85% | NPRM is out, comment period closed July 27, 2026, current DOT wants the barrier gone. Final rules still slip. |

| Tesla’s self-certification without an exemption stands as written | ~25–40% | Agency text still says you cannot certify to FMVSS 135 with no foot control. Special Order is the tell. More likely: exemption, or wait for the final rule. |

| Uncapped fleet deployment of no-controls vehicles once rules are rewritten | ~60–75% over 2–3 years | That is the point of dropping 135 and similar standards. Oversight and state permits still exist. |

| Unrestricted retail sale to consumers of a no-wheel, no-pedal car | ~30–45% by end of 2027 | Zoox’s grant explicitly blocked consumer sale. Crash seating, emergency stop, kids, accessibility, and “who is the driver in a recall” are still open. |


So: probability NHTSA blesses the vehicle *type* is already high. Probability it blesses Tesla’s current legal theory is much lower. Probability it becomes a normal showroom product with no cap is the uncertain one.


Why FSD data is a separate question

You asked to set FSD aside, and for equipment rules that is correct. Pedals and a wheel are about whether the car as built can be certified. They are not a score on disengagements.


NHTSA is still building a different thing: an ADS performance standard (SAE ASCEND work was just announced). That will care about driving data, incidents, and operational design domain. A pedal-free Cybercab can be legal as hardware and still be restricted as a driver. Those are two gates.


Bottom line

If the question is “will the U.S. allow a car with no steering wheel and no pedals?” — yes, it already has, in limited form, and the remaining equipment rules are being rewritten. I’d put ~80% on that becoming ordinary for fleet robotaxis once the pending FMVSS changes land.


If the question is “will NHTSA rubber-stamp unlimited consumer sale of a wheel-less, pedal-less vehicle just because the hardware exists, with no exemption and no performance standard?” — closer to one in three in the next year or two. The agency is opening the door. It is not taking the hinges off.


I’ll check how NHTSA actually treats Tesla FSD incidents versus viral X clips, and whether those videos are showing up in current probes.

Yes, but not the way the timeline implies. Viral FSD failure clips on X affect pressure, scope, and politics. They almost never decide the legal question of whether a no-wheel, no-pedal car can be certified. NHTSA does not tally likes.

How NHTSA actually uses those videos

The agency’s own process lists social media and forum posts as a screening source, alongside complaints, Standing General Order crash reports, news, and manufacturer logs. When a probe is open, NHTSA does not treat a clip as proof. It treats it as a lead: then it asks Tesla for logs, EDR, cabin/center-display video, and a second-by-second timeline. That is already happening in the FSD files.

That channel is live right now:

  • PE25012 / EA26002 (about 3.2 million FSD cars) was built in part from complaints plus media reports, including intersection and visibility cases. It is at engineering analysis — last step before close or a recall push.
  • NHTSA has separately grilled Tesla on Tesla’s own and Musk’s videos of people making espresso, dancing, reclining, etc., while FSD Supervised is on — i.e., the agency treats promotional clips as evidence of how the product is marketed, not just how it drives.
  • This month, Rep. Krishnamoorthi wrote DOT after an NBC tally of 43 sleeping-driver videos. That is the political transmission belt: X → news package → congressional letter → NHTSA has to answer on the record.

So the videos matter. They are not the docket.

What they will and will not move

Will influence (moderately to strongly)

  • Whether NHTSA keeps EA26002 open, widens it, or demands a software/driver-monitoring fix on consumer FSD.
  • How hard the agency leans on Tesla over foreseeable misuse. A stack of “asleep on the highway” clips is useful to NHTSA even if each clip is not a crash. It undercuts “the driver is responsible” if Tesla’s public content taught the opposite.
  • State and local comfort with unsupervised Cybercab expansion. Austin PD and Texas DMV watch the same feed Congress does.
  • Speed and tone of an ADS performance rule. Ugly clips make “just trust miles” a harder sell.

Will barely influence (the decision you asked about last)

  • Cybercab FMVSS certification — the Special Order on pedals, mirrors, and self-certification. That fight is about the text of FMVSS 135 and Part 555, not a dashcam of a Model Y rolling a stop sign. NHTSA has already said, in public, that a vehicle with no foot control cannot certify to that standard as written. A thousand X videos do not change the regulation. A final rule or an exemption does.

Those two tracks are legally separate. Consumer FSD is still Level 2 with a human required. Cybercab is being fielded as an ADS vehicle on a different software branch (v15). Same cameras, different legal box. NHTSA can hammer one and still let the other proceed under a cap, the way it did with Zoox.

The bias problem, which NHTSA knows

X is a failure-highlight reel. People do not post “uneventful left turn in light rain.” NHTSA has already shown it will close a noisy owner-video issue when the crash record is thin — phantom braking is the example: years of clips, investigation shut with “low demonstrated hazard.” Failure videos without a matching SGO injury/fatality pattern are ammunition, not a verdict.

The opposite is also true. A single verified fatality with logs plus a viral clip of the same failure mode (glare, red light, wrong-way) is much heavier than 200 “FSD hesitated at a merge” posts.

My read

  • On the no-wheel / no-pedal Cybercab ruling: influence of X failure videos is real but second-order — they make NHTSA less willing to look sloppy, not more willing to invent a new legal theory. I’d put that at maybe 15–25% of the outcome weight. Statute, the Zoox exemption precedent, and the pending FMVSS rewrite dominate.
  • On consumer FSD (recall, geofence, driver-monitoring mandate, marketing clampdown): influence is material, say 30–50% of the political and investigatory momentum, because that case is about behavior, attention, and whether the system plus Tesla’s messaging creates a defect. The actual close/recall decision will still rest on logs and SGO crashes, not quote-tweets.
  • On public trust and city-by-city robotaxi permission: the videos punch above their statistical weight. Regulators can ignore a ratio. Mayors cannot ignore a compilation package on the evening news.


Net: 


the clips will not decide whether a pedal-less car is legal. They will decide how tightly Tesla is watched while that question is open, and they raise the odds of a supervised-FSD restriction even if Cybercab keeps rolling in a geofence.

Comments

Popular posts from this blog

TESLA MODEL 2 vs ROBOTAXI

SENSOR FUSION ORTHODOXY.

THE CAVEAT FACTOR.