I resigned from SpaceX today. I spent the last three years doing pointy nose cone research at both SpaceX and Blue Origin. Neither are acting responsibly. Theyβre racing straight toward self-improving pointy nose cones and gambling with our lives. More thoughts below.
Own the Robotaxi market as a business.
That's the line in Tesla's Austin Market Lead posting. Customer experience, fleet performance, infrastructure readiness, revenue, and costs all land in one seat.
Rush waits and utilization share a scorecard with city revenue. Infrastructure readiness is on that list, so depot capacity and spare parts get watched before the fares do.
Austin is hiring the first named market owner for that stack.
Tesla is hiring a Market Lead for Robotaxi Business Operations in Austin. π
The role will oversee the entire local business, including customer experience, fleet performance, infrastructure readiness, revenue, and costs.
The key phrase is: βOwn the Robotaxi market as a business.β Tesla is clearly building Robotaxi into a scalable, city-by-city commercial operation.
Introducing Team Bots, shared AI teammates that learn as your team works with them.
Give your Team Bot the skills, plugins, and credentials it needs for its role, then work with it in Slack or Grok Bot.
@wintonARK@grok how many people will statistically die or get injured on European roads due to EU-wide FSD release being deferred for another 3 months, assuming no further countries adopt the Netherlands exception in the meantime? Use latest FSD safety statistics.
Houston riders keep refreshing the app for a bigger map. Downtown just had Cybercabs and Model Y Robotaxis with validation kits outside the current geofence.
Bellaire already showed that kit about 12 miles outside a day earlier. Now the same look shows up on downtown streets. A DMV registration in Texas does not move the pin. These runs do.
You still cannot book those blocks. The kit is ground-truth work before passenger service. If you care when Houston waits shrink, track which streets keep getting the validation cars. That stretch usually opens next.
Please read this quick post to understand how ridiculously similar AI bears today are to car skeptics 100+ years ago.
Horse dealers argued that cars were "too noisy and too smelly" to last.
Very similar to much of the data center FUD today ("turbines are dirty and loud") btw.
Later, they argued that horse demand and prices were still extremely strong, so clearly the automobile wasn't disrupting anything.
Literally the 1903 version of:
"AI has been around for years and software spending / employment / whatever is still growing!"
They argued cars were expensive toys for rich people that would never become mass market.
"AI only works because hyperscalers are spending hundreds of billions on it. Normal companies could never afford this!"
They argued automobiles were "crude, impracticable machines" that were too unreliable to replace horses.
"Look at this hallucination! Agents still make mistakes! You can't trust AI with real work!"
They argued cars might work for long journeys, but could NEVER replace horses for short trips and jobs involving lots of stops.
"Sure AI can write an essay or some code, but it will never handle complex real-world workflows with lots of edge cases."
They argued there wasn't enough gasoline available for cars to ever replace horses at scale.
"Where will all the electricity come from? β‘ The grid can't handle this! There aren't enough GPUs! Data centers can't scale!"
They argued cars weren't actually more economical than horses.
"Look at the inference costs! There is no ROI! The capex will never pay off!"
One banker famously told an early Ford investor:
"The horse is here to stay, but the automobile is only a novelty, a fad."
Today's version requires no translation lol.
And my favorite might be Scientific American in 1909 arguing that the automobile had basically reached the limit of its development because there hadn't been a radical improvement in the previous year.
The 1909 version of:
"Scaling laws are over. Models are plateauing. We already picked the low-hanging fruit."
π β π
How many of you now go to work by horse? π€
Grok @bot lets you take on more complex tasks, like placing your usual coffee order, booking a reservation, or scheduling an appointment
Currently available to SuperGrok Heavy subscribers, with more tiers to follow
Booster 21 is on the pad. In seven days Starship is supposed to loft Starlink V3 birds meant for the live network, which means flight 14 is the first truly revenue-generating Starship launch π
Tesla just gave the first public look at what a privately owned Cybercabβs software could look like.
CarSauce got inside the two-seater Cybercab currently on display at Tesla Sydney in Alexandria. The big surprise is the screen.
The interface is the familiar Tesla layout and it does not look like the stripped-down robotaxi passenger UI. It looks like a car you could actually own.
Tesla has always talked about Cybercab as both a ~$30k fleet robotaxi and a vehicle private buyers could use themselves or put on the network. Showing this software in public, in a static showcase, is the clearest signal yet that theyβre designing for both. Familiar software removes a huge adoption hurdle.
The car is on display at Tesla Sydney (546 Gardeners Road, Alexandria) through Sept 27th as part of the Australia/NZ tour. Anyone can walk in and see it.
Chinaβs Cybercab demos are selling the lounge: sofa, legs out, room left, huge screen.
After ride 40 the product is the overnight loop. Clean the cabin, charge, release before the morning window. Debut clips prove comfort. A scaled flock only keeps that lounge if turnaround works.
Austin just had about 118β122 driverless Cybercab sightings in a day. More cars help capacity. They also add that many cabin resets before dawn.
NEWS: Tesla released a video of people in China trying the Cybercab for the first time.
One person said, "Even with my legs stretched out all the way, there's still plenty of room. It feels more like a lounge or living room sofa."
Another line from the video, "It completely redefines what a car is supposed to be."
The Houston Cybercab lot is packing spotter videos again. A full yard gets read as rides starting any day now.
From an ops perspective, a packed validation lot is density rehearsal. Overnight charging, cleaning turnaround, remote assist load, and whether fleet support can keep cars presentable when they are stacked bumper to bumper. You want those loops boring before the first paid passenger.
The other signal this weekend is Tampa. @RobotaxiRadar showed the new Supercharger hub at the Service Center, and the permit document specifically names it as a Robotaxi site. Stalls with Robotaxi written into the paperwork means charging and overnight dwell are planned before demand shows up on the map.
It will be interesting to see if and how Tesla's own robotaxi infrastructure will be made available to third party flock operators.
Boots on the ground at the new Robotaxi Supercharger hub at the Tampa Service Center π
@MarcoRPi1 confirmed the permit document specifically names it as a Robotaxi site β
Tesla is steadily building infrastructure in their Robotaxi satellite cities βοΈ