Spotted this Cadillac Escalade IQ southbound toward Naples this afternoon. Marked as a GM Automated Research Vehicle with a full roof rack of LiDAR and sensors. Looks like one of the development vehicles GM is running for the eyes-off highway system they plan to put on the production Escalade IQ in 2028. Competition on the way?
1
1
17
8 procurement crashes. 37 dead. All since the first kidney transplant. We're still chartering jets to move organs that fit in a cooler the size of a lunchbox.
A drone doing the same route costs about a tenth of a King Air charter, roughly 1,200 versus 12,000 for a 2-pound blood sample kit.
NASA's modeling puts a fully loaded drone hour at 500 to 1,500 dollars against 933 to over 6,000 for air ambulance. The tech works. Vibration testing across 100 miles showed drone G-forces lower than both manned flight and ground, with the tightest temperature stability of any transport mode.
What's missing isn't engineering. It's the full cost model that includes chase vehicles, backup aircraft, and hospital landing pads. The redundancy layer that quietly eats the savings. The real currency isn't dollars, It's cold ischemia time. A drone that cuts an hour off a kidney's clock can turn a marginal organ into a usable one. That's a $400,000 graft against a $4,000 flight.
The FAA's drone rule still isn't finalized. 26 transplant organizations asked for organ transport to be baked in. It wasn't.
1
2
32
The interesting crypto move for Uber is not letting riders flex a balance.
It is letting drivers take the fare the same way.
Robotaxi has no driver to pay. Uber still does. Every extra day that payouts sit in cards and weekly batches, the human network looks like the slower product even when the car itself is fine.
Cryptocurrency on both ends of the trip is how a platform with people in it competes with a platform that does not. Rider pays in. Driver takes it out. No bank in the middle making them wait.
They already said this was coming.
Robotaxi is what turns “coming” into a clock.
26
Treating multi-operator robotaxi fleets as isolated businesses creates an adverse selection problem.
Cameras handle the mess in one car. If operators do not share incident data, the rider who trashes Fleet A is a new customer on Fleet B. The operators with the highest standards absorb the worst passengers.
Privacy arguments against a shared rider record will lag the fleet math. Rotating the same passenger’s mess across operators does not scale.
36
Finally saw a 3D printer on a Michaels pegboard.
Been waiting for Cricut to give up a bay. Cutters and printers solve different halves of the same “make it at home” problem.
Machine is here. Filament still isn’t priced like it is online.
26
You cannot buy Club Juke or an Ozzy house off a pallet. What you can buy is repeatability: a witch with 30 response lines, a 9'5" popup, a clearance jumpscare that does not care what night it is.
One is a ticketed set. The other is hardware you own.
I want both.
46
If you're parking under King Kong or Jurassic Park on the ground floor, your park strategy is already in the top 5%.
The moving walkway is still empty and the hub hasn't bottlenecked yet.
Best feeling in Central Florida infrastructure.
52
A veto just ended the program that helped low-income cancer patients in Pinellas get rides across county lines for treatment.
No more free transport. No easy way to reach chemo or radiation appointments.
These are people already fighting for their lives. Transportation shouldn’t be another battle.
@elonmusk — this is exactly what Robotaxi was built for.
Door to door. On demand. No drivers. No fixed routes.
Cancer patients shouldn’t have to choose between treatment and getting there.
Would you put Robotaxi to work here first?
@RidePSTA
44
Most specialized kitchen appliances are terrible tenants.
They take up permanent counter space, get used a few times a month, and contribute nothing the rest of the time. A machine that can make soft serve or protein ice cream is impressive, but it still has to justify the real estate it occupies in a kitchen that only has so much usable surface.
The best tools earn their spot. The rest are just expensive squatters.
35
The best security practices are the ones you don’t have to remember.
If you’re using Grok (or any AI) to help you code, make this a default rule:
Any HTML form that inserts user input into a live preview or generated document should escape HTML entities automatically. Don’t leave XSS protection as something you have to remember each time.
Most of these issues happen because the protection was optional, not because people didn’t know about it. Make it a standing project rule so it’s handled in the background.
28
A lot of the most valuable land at successful retail centers is still given over to surface parking.
These lots work fine today, but they represent a growing opportunity cost. As shared autonomous vehicles reduce the need for one car per person, large amounts of this space will look increasingly expensive to leave empty most of the day.
1
25
“Men wanted for hazardous journey, low wages, bitter cold, long hours of complete darkness. Safe return doubtful. Honor and recognition in event of success.”
In 1914 that ad ran in a newspaper. Today the equivalent call for Mars would almost certainly appear on X.
The medium changed. The type of person it attracts did not. Some people are still looking for difficulty and purpose over comfort and a predictable 9 to 5.
Curious how many would still answer that call today.
33
Most people subscribing to Grok through the App Store are paying more than necessary.
The higher price exists mainly to cover Apple and Google’s 30% commission on in-app purchases. Subscribing directly on grok.com gives you the same access and features at the actual rate xAI charges, without that platform tax built in.
If you’re planning to use it regularly, the web option is the simpler and cheaper path.
50
The next phase of AI isn’t about better chatbots. It’s about physical systems that can act in the real world.
Nvidia’s alignment with Japan’s industrial giants (Fujitsu, Fanuc, Kawasaki) on “Physical AI” marks a real shift. We’ve spent years training models to reason over text and code. The harder, more valuable problem is getting those models to safely control heavy machinery, coordinate supply chains, and operate in environments where a hallucination can cause real damage.
This changes the game in three ways:
Error tolerance drops to near zero. Software bugs are annoying. Physical mistakes are catastrophic.
The bottleneck moves from raw compute to tight hardware-software integration, sensors, edge silicon, and mechanical engineering working together in real time.
It targets actual labor shortages in manufacturing and industrial work rather than just displacing white-collar tasks.
The teams that can make digital intelligence produce reliable physical force will have a very different kind of advantage.
32
Caught this Aerostar stratospheric balloon (HBAL791) cruising at 64,600 feet over the Gulf, one of the cleaner rare spawns on Flightradar24.
These platforms sit in a useful middle layer: high enough to stay above weather and commercial traffic for weeks at a time, yet far cheaper and faster to deploy than satellites. They’re used for persistent communications relay, atmospheric research, and surveillance where drones can’t match the endurance and satellites can’t match the flexibility or cost.
It’s a reminder that the aerial infrastructure stack is getting more layered, not just higher or faster.
50
Hyundai buying Lucid would give them some of the strongest EV powertrain and efficiency engineering currently available, along with a U.S. factory and an existing robotaxi pipeline through Uber.
The bigger issue for Hyundai right now isn’t a lack of technology, it’s scaling profitably in a cooling EV market. Lucid has shown the opposite pattern: excellent hardware that has never reached sustainable volume or margins without repeated external funding. Taking on that dynamic now would add cost, complexity, and distraction at the exact moment when manufacturing discipline and cost per vehicle focus matter most.
1
1
106
The 50 million kilometers (over 31 million miles) of FSD Supervised driving logged across five European countries in under three months shows the tech clearing a harder test than most headlines capture.
The car is already generalizing across different rules, signage, cyclists, and construction without per-country rewrites. What’s slowing wider rollout now is mostly regulatory pace, not the vehicle’s ability to adapt.
Every one of those miles also feeds the same global training loop, so the edge cases European drivers are hitting are quietly sharpening FSD for owners everywhere.
30
Starship Flight 13 is targeting July 16 and will carry the first Starlink V3 satellites.
The interesting part is how this connects to the recent 10 Gbps result in Utqiagvik (roughly 4,500 people located 320 miles above the Arctic Circle with no road connection to the rest of the state). That came from bonded gateways in one of the most difficult environments on the planet. V3 satellites are built for much higher per satellite capacity, and Starship can put a lot more of them up per flight.
Together, these change the slope: instead of peak performance staying rare and location-specific, it starts becoming more normal across the network faster than simple linear growth would suggest. The July 14 launch from FL tomorrow is another step in the same direction.
The real variable now is how quickly that higher ceiling turns into more consistent real world performance for more people.
43
What if Florida’s Road Rangers ran Cybertrucks instead of the usual white pickups?
I looked at it like a small team of experts: capabilities, cost, real-world fit for highway assistance.
The Cybertruck actually lines up surprisingly well: massive towing, secure vault bed, built-in high-power export (powershare), and that tough exoskeleton. Operating costs could drop significantly in the long-term.
Not perfect (charging infrastructure is the main hurdle), but worth exploring as a pilot.
Would love thoughts from anyone in fleet ops or FDOT circles.
58
This open water area right outside Mote Marine in Sarasota is beautiful… but it feels like something important is missing.
Right now it’s mostly grass, a path, and one lone palm tree next to the water. Most people drive past without a second thought.
Here’s what a few practical additions could turn it into for locals, visitors, and the bay itself. 👇
3
56
3) With IKEA coming in right next door, this spot feels like the perfect time to do something.
That development is going to bring a ton of new traffic and attention to the area. Right now this waterfront is basically empty, except for the grass and palm tree.
Adding a clean path and a couple simple places to sit would turn it into an actual public waterfront moment instead of leftover land next to a giant store.
18












