John Krafcik, Waymo founder and former CEO, clapped again at Tesla CEO Elon Musk after his feedback on Waymo at Tesla’s earnings earlier this week.
He introduced just a little dose of actuality to Musk’s delusions.
Throughout Tesla’s earnings name, CEO Elon Musk made a bunch of recent claims about up to date timelines for Tesla to ship on long-promised self-driving milestones.
Musk even said that he doesn’t see anybody competing with Tesla on autonomous driving and that he believes Tesla will take “90-something p.c market share.”
When explicitly requested about Waymo, which is already seen because the market chief by most consultants, Musk once more claimed that it received’t be aggressive due to value:
Effectively, okay. The difficulty with Waymo’s automobiles is it prices far more cash, however that’s the situation. The automobile may be very costly, made in low quantity. Teslas are in all probability value 25% or 20% of what a Waymo prices and made in very excessive quantity. So, satirically, like, we’re those to make the wager {that a} pure AI answer with cameras and what do you have got? The automobile really will hear for sirens and that type of factor. It’s the best transfer. And Waymo determined that an costly sensor suite is the way in which to go, despite the fact that Google is superb at AI. So I’m questioning. And it’s price noting that Tesla has each an unbelievable AI software program staff and AI {hardware} chip design staff.
Whereas it’s true that Waymo’s autos are dearer, this has little to do with an “costly sensor suite”, as Musk claimed.
Tesla’s sensor suite, which consists solely of cameras, is definitely inexpensive than Waymo’s, which incorporates cameras, radar, and lidar sensors. Nonetheless, a lot of the value distinction is because of Tesla constructing autos in excessive volumes for customers whereas Waymo buys present autos and retrofits them with its sensor suite in a lot decrease volumes.
Speaking to Enterprise Insider this week, John Krafcik, who helped launch Google’s self-driving automobile program and was Waymo’s first CEO till 2021, doesn’t see it as an issue:
In the long term, the price of sensors has a “trivial cost-per-mile influence over the helpful lifetime of a robotaxi,” he advised BI,” whereas additionally offering large quantifiable security advantages.”
Krafcik has been skeptical of Tesla’s efforts for some time and even steered that what Tesla plans to launch in Austin in June might quantity to “faking” self-driving.
Now, the previous Waymo CEO highlights the way it’s not the primary time Musk claimed that Tesla would compete with Waymo, however it’s but to occur:
“Tesla has by no means competed with Waymo — they’ve by no means offered a robotaxi journey to a public rider, however they’ve offered loads of automobiles. And though Tesla hopes to compete with Waymo sometime, they’ve failed totally and utterly at this for every of the ten years they’ve been speaking about it.”
This week, Waymo introduced that it’s now finishing 250,000 self-driving rides per week.
In June, Tesla plans to launch a restricted “10 to 20-car” fleet to offer geo-fenced and teleoperation-assisted rides in Austin.
At this fee, it could take some time for Tesla to catch as much as Waymo’s quantity and that’s if it could actually.
Krafcik added:
Effectively, after 10 years of undelivered guarantees, it appears fairly rational for these watching to be data- and evidence-driven. There’s nonetheless loads of guarantees, nonetheless no Tesla legal responsibility for FSD driving efficiency, and nonetheless no common robotaxi service.”
After Tesla’s earnings, the automaker introduced that its “FSD Supervised ride-hailing service” working for workers in Austin and San Francisco accomplished “over 1,500 journeys and 15,000 miles of driving.”
Nonetheless, these had been accomplished with drivers able to take management, and even when that they had no interventions in these 15,000 miles, it wouldn’t be sufficient to show a better security than human drivers, who crash on common each 700,000 miles.
Electrek’s Take
He has a degree. Irrespective of how a lot Tesla shareholders wish to imagine Musk when he says that Tesla is forward, the reality is that when it launches its restricted 10-20-car fleet in Austin, it is going to solely now begin to compete with Waymo.
By that point, Waymo could be doing over 300,000 rides per week, and there’s no proof that Tesla will be capable of scale quicker but.
Similar to Waymo, Tesla’s fleet might be geo-fenced and teleoperation-assisted as there’s no proof that Tesla’s generalized answer is wherever near working.
Briefly, Tesla has the identical scaling limitation as Waymo, apart from doubtlessly mapping, however even that’s a query mark.
