For Tesla Inc, the distance between a disastrous robotaxi launch and a merely disappointing one was just a foot or two. The “safety monitor” who was sat in the front passenger seat of the handful of Tesla robotaxis that began operating in a limited part of Austin on Sunday was the contradiction of Elon Musk’s autonomous ambitions made flesh.

To be clear, it is sensible to have a human driver ready to intervene when robotaxis first get out on the road; just as Waymo LLC, owned by Alphabet Inc., did. Whether they sit in the actual driving seat or a short distance away on the passenger side makes little practical difference, provided they can still stop the vehicle or reach across and adjust the wheel if necessary.

It does make a difference, however, when you have spent years saying that your company is ready to unleash swarms of self-driving vehicles that are safer than humans, and that company’s trillion-dollar valuation rests largely on the assertion being true.

In that case — let’s call it the Tesla case for convenience — it is important that there be no driver behind the wheel. This allows for initial riders to post backseat videos of the wheel turning itself uncannily, implicitly playing down the back-up sitting a foot or two away.

For any other company, Tesla’s robotaxi launch would have been a big success. It would have demonstrated the crossing of a significant threshold — vehicles driving themselves on public roads with passengers in the back — even if only in a limited number of vehicles, in a limited area and with humans ready to intervene. The latter include the remote support workers back at robotaxi HQ.

For Tesla, though, any judgment of success must be set against the pitch. On that basis, it was, under any reasonable standard, an admission of failure. This is where Waymo, a rival that Musk mocks frequently, was years ago. Had Tesla sat the safety monitor in the driving seat instead, it would have been an outright disaster in PR terms, lacking even the appearance of a meaningful advance.

Tesla’s core proposition with regards to autonomy is that it is building a “generalised solution”, where artificial intelligence learns to handle any situation the world can throw at it. This supposedly reduces the need for expensive sensor suites incorporating things such as LiDAR; Musk says cameras are enough. It also means, in theory, that an autonomous Tesla can adapt to and work pretty much anywhere. As recently as last summer, Musk was saying:

“If you see, like Waymo and whatnot, they have a very localised solution that requires high-density mapping, and it’s not — it’s quite fragile. So their ability to expand rapidly is limited. Our solution is a general solution that works anywhere. It would even work on a different Earth.”

Less than a year later, the actual Tesla robotaxi operates for invite-only riders (no children allowed), in “limited areas of Austin”, with a safety monitor, between the hours of 6am and midnight, and perhaps not at all in the event of “inclement weather”. It’s not quite ready for a downtown school run in the rain, then, let alone extraterrestrial habitats.

Tesla is right to be cautious in practice. Even Sunday’s highly controlled launch appeared to have some instances of robotaxis violating road rules. Any accident, particularly one resulting in injury or fatality, would be potentially catastrophic for the company’s autonomy efforts and already-struggling brand. The problem is that the rhetoric around it is so often anything but cautious. Despite this minimal launch, and unscalable use of an in-car monitor, Musk said in April that he was “confident” Tesla robotaxis would be available in “many cities in the US” by the end of the year.

That confidence is notably less pronounced than with his talk of a “different Earth” last year, but I suspect may also have to be downgraded over the next six months. Wedbush Securities analyst Dan Ives wrote that his invite-only experience in Tesla’s robotaxi on Sunday “exceeded our expectations” — and this is someone whose expectations run to a $500 price target, more than 40 per cent higher than today’s level and beyond the all-time peak.

For now, the footage of self-turning steering wheels allows the narrative to roll on, unimpeded by real-world conditions.

• Liam Denning is a Bloomberg Opinion columnist covering energy. A former banker, he edited the Wall