Points Beneath The Floor With Tesla FSD vs. Waymo Driver

Help CleanTechnica’s work via a Substack subscription or on Stripe.

Following my newest article on the long-running Waymo vs. Tesla FSD debate, just a few readers tore into me (and Tesla) somewhat bit. One remark jumped out at me as a extremely necessary and fascinating level, after which one other did, after which one other. So, I’m publishing them right here for extra consideration and dialogue.

Redundancy vs. …

Reader “fm123” wrote: “The largest concern is the dearth of redundancy within the Tesla system. They designed it to be low price, however which means leaving out redundant options for security. Waymo has separate steering and braking management even when the automotive drivetrain fails. There’s additionally a separate battery system for the driving force pc, so it ought to stay operational even when the automotive electrical system fails. That’s along with the sensor redundancy and washer/wipers for cameras. Sensors that bodily detect objects and measure velocity and dimensions/distance on prime of digital camera AI techniques is best for poor lighting circumstances.”

It is a crucial level. So long as every part is working because it’s speculated to, particularly the cameras, Tesla FSD performs exceptionally nicely. However, if Tesla goes to quickly set automobiles unfastened with out human supervisors (one thing that’s hotly debated), when one thing goes mistaken with the system, what occurs from not having a backup system? Sure, the automotive can determine that it could actually’t function correctly and cease driving, or maybe pull over shortly and cease driving. However that would create fairly just a few issues. Tesla followers prefer to criticize Waymo for having extra {hardware} and being dearer on that finish, however there’s something to be mentioned for higher redundancy and backup in such a vital machine.

Incentivized Too A lot?

Reader “Michael” wrote: “People are likely to react to the incentives placed on them. The regarding factor for me about Tesla’s autonomy efforts is that from Musk on down, everybody at Tesla is nicely conscious that their sky-high inventory value relies on the promise of autonomous driving, and if their buyers stop believing in Tesla turning into the dominant supplier of autonomous driving, the inventory value will plummet, and wipe out an enormous quantity of Tesla worker’s internet price. As such, I don’t belief them to prioritize the protection of the general public above all else. My lack of religion is exacerbated by the best way Tesla does every part attainable to cover their security information. Their newest crash report out of Austin is redacted to the purpose of absurdity. Significantly with Waymo accelerating its enlargement, we’re vulnerable to Tesla throwing warning to the wind in an try and show to buyers that they can’t be written off as a competitor.”

Including onto the primary level above, this appears vital. When there’s a lot incentive to getting robotaxis deployed final month quite than 3 years from now, are corners being minimize, are folks being overly constructive in regards to the outcomes, are information being hidden or manipulated, are steps being skipped in an effort to look even with or about to surpass Waymo? What number of Tesla staff have shares of the corporate that they know are primarily based closely on speedy, important FSD progress? Maybe issues associated to this robust incentivization aren’t even evident but, however can be in a 12 months or two as extra robotaxis are deployed.

Crowdsourced FSD Knowledge Is Not Uplifting

Reader “Phil11” provides the next:

“Do you not imagine the crowdsourced information from the Tesla FSD tracker?

“In case you merely use that information to calculate what would occur if Waymo used the newest FSD (14.2.1) to journey its 2 Million+ miles per week, it could crash about 4,000 instances per week and require about 57,000 human interventions which is neither sensible or secure.

“The above numbers proves that Tesla FSD shouldn’t be remotely near ever being autonomous and but you proceed to fake that the confirmed liar — like traitor trump — deserves a wait and see each time he lies to pump his inventory.”

Yikes! Sure, if these numbers are even near consultant, you’ll be able to’t take a look at the info and assume Tesla FSD is prepared for mass deployment. And it’s wild to think about what the response can be if Tesla robotaxis did drive greater than 2 million miles per week and get into ~4,000 accidents in that point.