Ming✓

Dear Mr. Anderson, Tesla never ignored physics. I'm writing because of what you told The Verge on October 27, when you called dropping radar from a self-driving car "a weird religious stance that ignores physics." That line got a lot of attention among Tesla owners, and I think it deserves a direct answer. You know this territory better than most. You ran Tesla's Autopilot program from late 2015 to late 2016, cofounded Aurora, and now lead product at GM. Your new employer is betting the other way. A 2028 Cadillac Escalade IQ will launch eyes-off highway driving with cameras, radar and lidar all working together. As I understand it, your case rests on four claims. Lidar helps camera systems learn faster, cameras can't see through dense fog, cameras can't measure instantaneous velocity the way radar can, and radar is so cheap that removing it makes no sense. I agree that physics is real, and nobody at Tesla is pretending otherwise. What I'd like to show you is that Tesla's own engineering has been answering these exact points for years, and two of your arguments describe things Tesla already does. A lot has also changed since you left. The in-house FSD computer arrived in 2019, camera-only Tesla Vision followed in 2021, and the end-to-end neural network that drives the car today came in 2024. So let me take your physics one claim at a time. 🎓 ON LIDAR AS A TEACHER, TESLA HAS BEEN DOING THAT FOR YEARS You said lidar works on a wavelength close to visible light, so it helps camera systems learn faster. You're right, and Tesla agrees so strongly that it built an entire training pipeline around that idea. I like to think of lidar as a tutor holding a perfect answer key. It measures distance directly by timing laser pulses, so it can grade every guess a camera network makes about how far away something is. Tesla has designed a system where a special test fleet carries radar and lidar alongside the regular cameras. Every reading is synchronized, so the precise depth and speed from the expensive sensors get attached automatically to the objects the cameras see. The labeling even uses hindsight. When two cars blur into one blob early in a clip and separate later, the training servers rewind the footage and use the clear moment to label the confusing one correctly. Tesla applies the same teacher to 3D space itself. Its occupancy network learns from lidar-backed ground truth to predict, for every small cube of space around the car, how far that cube sits from the nearest surface. That's how plain cameras learn to sketch a sloped curb or a curved bumper smoothly instead of as a pile of blocky Lego bricks. Even the grading got faster. Tesla designed a cross-labeling trick that draws a box around a car in a flat photo, then projects that box into the laser point cloud as a pyramid of vision, so the software only searches that slice instead of the whole 3D world. You probably know DeepScale, the startup Tesla acquired in 2019. It was already working out in 2016 how to train cheap sensors to mimic premium lidar data, and Tesla is still pursuing that work today. Here's the step where our views split. Once the student has learned, the teacher can stay home. The test fleet keeps the expensive hardware, and the cars customers actually buy ship with cameras only. Tesla has put real money behind this. In early 2024, lidar maker Luminar named Tesla its largest customer for the quarter, and Elon soon replied on X, "We don't need them even for that anymore." The student is even learning to write its own answer key. To map the shape of the road surface, Tesla has designed a method that can triangulate ground truth points straight from camera images instead of relying on lidar. Co-training requires lidar in the classroom, not in every car on the road. With respect, GM's plan keeps the teacher riding along forever. ⚡ ON SPEED, CAMERAS ARE QUICKER THAN THEY LOOK Your sharpest point is about speed. Radar can measure how fast something is moving in a single snapshot, and a camera can't. Radar's trick is the Doppler effect, the same physics that makes an ambulance siren sound higher as it rushes toward you and lower as it drives away. Radar reads that pitch shift in its own radio echo to measure closing speed instantly. A camera has to work speed out from motion. It compares where an object sits in one frame with where it sits in the next, much like you judge a thrown ball by watching it fly. That sounds slower, but the gap is tiny. Tesla's own designs describe cameras running at rates such as 36 frames per second, which puts a fresh frame in front of the network about every 28 milliseconds, quicker than a single blink. The physics also cuts both ways. Doppler only measures motion toward or away from the sensor, so a car crossing your path sideways at an intersection produces little to no Doppler shift. To catch that sideways motion, a radar has to track the car across several scans, which is exactly what a camera does with frames. On that kind of motion, the "instantaneous" advantage disappears. Radar has another well-known weak spot. Owner's manuals for radar-based cruise control have long warned that the system may not brake for a stopped vehicle, partly because a parked car, an overhead sign and a bridge can all look like the same motionless metal. Now here's what Tesla has built since you left. Tesla has designed its vision network to output a full 3D velocity vector for every object, covering speed, direction and even height. It learned that skill from the radar on its test fleet, which turns ordinary cameras into something that behaves like a Doppler radar made of pixels. Tesla's virtual camera invention goes further by measuring speed in two different ways. Cars are tracked relative to the moving Tesla itself, while pedestrians are tracked against the fixed ground, so the system knows a person's true walking path even while the car accelerates or turns. When things get fast, the network speeds up its memory. A frame selector pulls video snapshots more rapidly as speeds rise and blends them with the car's own acceleration and yaw rate, which is how quickly the car is rotating, to build a smooth picture of motion over time. There's also a dedicated sentinel for the car directly ahead. A super-narrow model watches only the closest vehicle in the lane through the narrowest forward lens and keeps estimating its distance, speed and acceleration, which is exactly what you need to brake hard and early. So yes, a camera infers velocity instead of measuring it in one shot. But inference refreshed every few dozen milliseconds, trained on radar's own answers, and able to see sideways motion that Doppler misses is not a camera ignoring physics. 🌫️ ON FOG, I'LL GIVE YOU THAT ONE, BUT HEAR ME OUT No camera sees through dense fog, and neither does the human eye. Radar does. But look closely at your own argument. The near-visible wavelength that you said makes lidar a great teacher is the same reason lidar struggles in fog, because the tiny water droplets that scatter visible light also scatter laser light. So the fog argument is really a radar argument, and radar alone can't finish the job. It can't read lane paint, a traffic light, a speed limit sign or a flagger's hand signal, so even a car loaded with radar still leans on its cameras when the world turns gray. That's why I think the real question isn't whether a car can see through fog. The real question is whether the car knows when its vision is degraded, and whether it behaves like a careful driver when it is. This is where Tesla's visibility invention comes in. Tesla has designed a separate neural network whose only job is grading its own eyesight. It splits every camera feed into a fine grid and scores each cell from clear to blacked out, building a live heat map of where the car can and can't trust what it sees. Each camera position gets its own specialist. A model judging the left pillar camera was trained on images from that same pillar location across many other cars, because a windshield camera behind wipers and a fender camera facing tire spray get dirty in very different ways. The system also diagnoses the cause. It can tell a smear of mud from fog, heavy snow or a hardware fault, then reach for the right fix, such as running the wipers or warming the glass. When a lens is blinded for a moment, the car doesn't forget the road. Tesla anchors what it saw to fixed spots in the physical world, so a pothole hidden by a truck's spray stays pinned on the car's mental map for those few seconds. And when conditions really are too much, the car responds the way a good human driver would. It slows down to widen its reaction window, can move to a safer lane, and can hand control back with a plain explanation such as "Due to Heavy Snow, Autopilot is Reducing Speed." Tesla is also sharpening the eyes themselves. Its designs feed raw, unprocessed sensor data straight into the AI chip, the same idea behind the "photon count" images Elon posted in May, when he said this is why FSD can see so well at night or through extreme glare. Around the lens, Tesla has designed micro-cone glare shields, corrective windshield patches, pulsed laser cleaning and even gas-absorbing plastics that keep a hazy chemical film off the inside of the glass. Each one attacks a different way a camera can lose its view. Fog will always be fog. Tesla's bet is a car that measures how blind it is, remembers what it saw, and slows down exactly when a wise driver would. 💸 YES, RADAR IS CHEAP, BUT THAT WAS NEVER THE POINT You're right that a radar unit doesn't cost much. But Elon's case against it was never about the sticker price. In April 2021, as Tesla prepared to remove radar, Elon explained it this way: "Sensors are a bitstream and cameras have several orders of magnitude more bits/sec than radar (or lidar)." A second sensor, he added, has to meaningfully improve the signal to be worth the complexity of integrating it. In August 2025 he put it even more bluntly: "Lidar and radar reduce safety due to sensor contention. If lidars/radars disagree with cameras, which one wins?" In the same post, he wrote that Tesla turned off the radars in its cars to increase safety. Picture two witnesses to the same moment telling different stories. The radar reports a solid object ahead, the cameras see an empty road under a steel overpass, and the car has a fraction of a second to decide who's wrong. Trust the radar every time and you get phantom braking under bridges. Trust the cameras every time and the radar was only along for the ride. Then there's the cost that never shows up on a parts list. Every extra sensor needs calibration, its own data pipeline, fusion software and validation for every way the sensors can disagree, and that burden grows with each new sensor type. Tesla's own history shows how heavy that lift is. The DeepScale design for mimicking lidar needed separate neural branches for each sensor type just to absorb their different quirks, timing and mounting positions before merging them. I'd argue the biggest hidden cost is data. Tesla's FSD (Supervised) fleet has logged more than 16.7 billion miles, nearly all of it seen through cameras alone, and a radar channel added today would arrive with none of that history to learn from. Look at what happened at Volvo this summer. In July, the company confirmed it was dropping lidar from the EX90 and ES90 after cutting ties with its supplier Luminar, which went bankrupt, and Norwegian owners are receiving NOK 18,000 per car for lidar features that never arrived. Elon's reply was short: "I did try to warn them. Humans drive using neural nets and optical sensors. Same is true for robot cars." Cheap hardware can turn into an expensive argument, and back in 2021 Tesla decided it would rather not have the argument at all. 🛣️ WHILE YOUR SYSTEM WAITS FOR 2028, CAMERA-ONLY TESLAS ARE ALREADY LOGGING THE MILES Your plan for handing control back deserves real credit. You said the Escalade won't toss the wheel to a surprised driver, and will instead give a countdown of around ten minutes and pull off at an exit if nobody takes over. That philosophy should sound familiar, because Tesla's visibility invention describes the same graceful exit. The car slows down first, looks for a safer lane, explains the reason on the screen, and only then hands control back. The real difference is timing. GM's eyes-off system is still a 2028 promise, while Tesla's camera-only cars are collecting evidence right now. Tesla's latest FSD (Supervised) safety report shows 6,006,200 miles per major collision in North America, compared with a U.S. average of 660,164 miles. That's roughly nine times as far between airbag-level crashes, measured over a rolling twelve months. Tesla's new Robotaxi safety report goes a step further. Driverless Teslas with no safety monitor inside have covered 3.1 million miles in Texas and Florida this year, including 862,000 miles after dark and 113,000 miles in rain, with zero people injured by a Robotaxi and zero airbag deployments, based on the crash reports Tesla files with NHTSA. And that Robotaxi fleet runs on the latest Model Y and the Cybercab, neither of which carries radar or lidar. I find that hard to square with the idea that camera-only driving ignores physics. 🔭 WHERE I THINK WE REALLY DISAGREE You framed this as physics versus faith, but I think both sides accept the same physics. What we really disagree about is where the intelligence should live. GM puts more of it in the hardware, stacking sensors so each one covers another's weak spot. Tesla puts it in the neural network, using radar and lidar as teachers in the classroom and then trusting a trained brain behind the cameras. There's a quiet twist hidden in that choice. A multi-sensor car can only learn from fleets that carry the same sensors, while every camera-equipped Tesla on the road feeds the same classroom. Physics decides what a sensor can see. Data decides what a car understands. Best regards, Ming P.S. If you're ever in Austin, a driverless Tesla might make my case better than I can, and it won't need lidar to find you.

ENGINEERING NEWS 🗞️ Tesla has designed a red-hot shortcut that lets plain water pull lithium out of old batteries 💧 When a lithium-ion battery goes through a shredder, it comes out as a fine black powder that looks like soot. That powder still holds the lithium, nickel, cobalt, and graphite that went into the cells, and the lithium is the trickiest part to win back. Recyclers call it black mass. It comes from battery factory scrap like trimmed electrodes, rejected cells, and pack offcuts, and one day retired packs will add to the pile. Getting that lithium out typically takes brute force. Common processes run furnaces hotter than 900 °C, lean on acid to dissolve the powder, or both. From there, a long chain of steps pulls each metal back out of the soup. Lithium tends to lose out in that crowd. It gets diluted, lost in handling, or tangled up with impurities, and the cleanup leaves large volumes of liquid waste behind. Even the lithium that survives often arrives as a mix of lithium carbonate, lithium sulfate, and lithium hydroxide, which needs extra purification before it can go into a new cathode. Lithium carbonate carries one more handicap. It barely dissolves in water, roughly a tenth as well as lithium hydroxide at room temperature. Tesla flips the order and frees the lithium before any acid touches the powder. The journey starts in a furnace at about 750 to 900 °C, where the battery can bring its own fuel. Black mass already carries plenty of graphite from the anodes, and at these temperatures that carbon pulls oxygen out of the cathode material. A rotary furnace can keep the powder tumbling like clothes in a dryer while it heats. A cathode particle works like a cage of metal and oxygen with lithium tucked inside. Carbon yanks out the oxygen bars, the cage collapses, and nickel and cobalt drop out as metal. The lithium keeps just enough oxygen to become lithium oxide, which is exactly what Tesla wants. Lithium carbonate tries to crash the party here. Some reactions park the lithium as carbonate first, but above about 750 °C the heat can crack it into lithium oxide and carbon dioxide. That carbon dioxide is the next threat, because it can grab lithium oxide and lock it right back up as carbonate. Tesla shuts that door in two ways. A steady stream of nitrogen, sometimes with a touch of hydrogen, sweeps the gases out of the furnace. Meanwhile, the hot graphite can turn stray carbon dioxide into carbon monoxide, which leaves lithium oxide alone. The temperature window is a balancing act. At 650 °C, too much lithium stays locked up, and only a little over a third of it comes out later. Hotter runs raise costs and sinter particles together, and lithium oxide can start slipping away as vapor above about 800 °C, so Tesla weighs furnace time against the heat. The sweet spot sits around 850 °C. With the powder held there for four hours, lithium recovery passed 96 percent in testing. Then comes the step that sounds almost too simple. Tesla can drop the treated powder into room-temperature water and stir it hard for an hour or two. Lithium oxide reacts with the water to form lithium hydroxide, which dissolves easily, while nickel, cobalt, and graphite stay behind as solids. This step needs no acid, and the water turns strongly alkaline, around pH 12 to 14. If a little carbonate slips past the furnace, a dose of lime can convert it into lithium hydroxide right in the tank. Tesla tunes the water the way a barista adjusts a brew. More water rinses out more lithium but makes a weaker brine, while less water makes a richer brine and leaves more lithium behind, so Tesla dials the ratio to hit its target. A pressure filter then splits the slurry in two. The liquid is a lithium hydroxide brine that can crystallize into lithium hydroxide monohydrate, the same form cathode makers rely on. One version of the process sends that brine "to Corpus," most likely the Corpus Christi area in Texas, where Tesla built its lithium refinery. The solids become a dark metal cake of nickel-cobalt alloy and graphite, with most of its lithium washed out. Because the furnace already stripped the oxygen from those metals, the cake gives them up far more easily in later processing. Under identical acid conditions, the cake released 99.9 percent of its nickel and cobalt, while raw black mass released only 43 percent of its nickel and 31 percent of its cobalt. The furnace can also clear out leftover electrolyte and break down the PFAS "forever chemicals" hiding in the electrode binder. The approach holds up beyond the lab bench, too, pulling about 85 percent of the lithium into the brine during a single three-hour wash in a 50-liter tank. That soot-black powder no longer has to surrender its lithium last, battered by acid and buried in waste. Its lithium now walks out first on a stream of water, already in the exact form a new cathode is waiting for. The nickel and cobalt follow right behind, packed into a cake that gives them up almost completely. The acid route fights the battery every step of the way. Tesla's route lets the battery's own graphite help set its lithium free. One red-hot shortcut later, plain water carries that lithium back toward a brand-new battery.

Source:

STRONG SIGNAL 🇮🇹 Italian Deputy Prime Minister and Infrastructure Minister Matteo Salvini just had a constructive call with Elon Musk today, noting that Italy supports rapid EU approval for Tesla Assisted Driving 🔥 “Italy […] supports swift EU approval of Tesla Assisted Driving.”

Very positive and constructive conversation with @elonmusk today. Italy is looking with great interest at the opportunities offered by assisted driving technologies, with a focus on innovation and road safety, and supports a swift EU approval of Tesla Assisted Driving. Thanks for the exchange, Elon!

Elon Musk says saving money for the distant future is a total waste of time because superintelligent abundance is right around the corner 🔥 He seems right about the abundance. But if you take his advice and stop saving, you might never survive long enough to see it 💀 Quick background if you do not follow Elon closely. For years, he has been pushing Universal High Income, or UHI: a regular federal check for every single adult, no strings attached. Universal Basic Income is designed to cover the absolute basics, but UHI goes much further. It is meant to let you live exceptionally well without ever needing a job. Here is how he sees it working. AI and robots end up taking over most of the work, making employment completely optional. He has not laid out a formal funding plan yet, but the underlying logic is simple: the machines generate immense wealth, and government checks distribute it around. For that vision to come true, three things have to go right. Machines must perform most of the labor on a massive scale. That work must make goods and services dramatically cheaper. And total output has to outpace the money being handed out, or those checks will simply drive prices up through inflation. The first part is easy to accept. High productivity pushes prices down, and AI paired with robotics is productivity on steroids. Even Reason magazine, which called his UHI plan a major mistake, agreed he is almost certainly right on that specific point. Getting there is the hard part. Start with the raw numbers. The U.S. has about 270 million adults. Pay each of them $50,000 a year, and UHI carries a price tag of around $13.5 trillion. In fiscal 2026, the federal government collected $5.4 trillion in total revenue and still ran a $2 trillion deficit. Scale it down to a modest $1,000 a month, and it is still roughly $3.2 trillion annually, which eats up about 60% of all federal revenue. Even if UHI replaced every single existing federal program, the $50,000 version would cost nearly double the government's entire $7.4 trillion budget. That is why UHI simply cannot run on today's economy. It only works once those three conditions reshape the market so that every dollar stretches vastly further. And not everything gets cheaper. Robots can crank out more cars, more houses, and more food. They might even build new islands and deliver world-class teaching to every kid. What they cannot manufacture is more of the absolute best: the hottest neighborhood, the view everyone wants, or the prime spot everyone else is fighting for. Hand everyone a big check, and those finite prizes will get bid straight through the roof. Then Congress has to pass it. No UHI program exists today. One of the largest U.S. cash studies to date gave 1,000 low-income Americans $1,000 a month for three years. They covered their basic needs better and worked about 1.3 fewer hours a week, giving both supporters and critics plenty to cite. Consider that a small preview of the political showdown ahead. Now look at the timing. Elon has said work could become optional in 10 to 20 years, and he expects AI to be smarter than all humans combined by 2030. That is the technology clock. The policy clock runs much slower, and jobs can easily vanish on the first clock while the checks are still stuck waiting on the second. That gap is precisely where his advice falls apart. When your job is on the line and the checks have not shown up, savings buy you crucial time, and skills decide what you do with it. Even Elon calls abundance the benign scenario. In the bad one, as he has noted, all bets are off. I am betting on the age of abundance too. I am just building my own bridge to get there, one plank of savings and one new skill at a time. Build yours, and you will walk confidently into Elon's future instead of waiting around for it.

@battleangelviv Saving money for the distant future is pointless with the advent of super intelligent abundance

BOOM 🇨🇳 According to Tianyancha, a Chinese corporate information platform, Tesla has recently established Tesla Auto Sales and Service (Xiong’an) Co., Ltd., with Wang Hao as its legal representative and $1 million in registered capital. The new company's business scope covers vehicle sales, new energy vehicle sales, and wholesale distribution of automotive parts and accessories. It's also fully owned by Tesla Motors Hong Kong Limited. But what makes this interesting is where Tesla chose to set up shop. Located in Hebei Province, roughly 100 kilometers (62 miles) southwest of Beijing, Xiong’an is no ordinary city. It's a major national development zone launched in April 2017 under President Xi Jinping's initiative. Often called China's "City of the Future," Xiong’an is being built from the ground up as a high-tech urban hub, designed to attract businesses, research institutions, and key functions relocating from Beijing. So why would Tesla need another company when it already has a massive presence in Shanghai? Tesla actually operates a network of regional sales and service subsidiaries across China, including Beijing, Guangzhou, Shenzhen, and Tianjin. These companies help handle local sales, customer service, and day-to-day business operations. The new Xiong’an entity appears to follow the same playbook. Even its $1 million in registered capital is in line with several of Tesla's existing regional subsidiaries. While the registration alone doesn't confirm a new showroom or service center, it could be an early step toward expanding Tesla's local sales and service network as Xiong’an grows. It's a small corporate move on paper, but it could give Tesla an early foothold in one of China's most ambitious urban development projects.

Source:

FSD NEWS 🇰🇷 A Seoul court sided with Tesla Korea on October 29, 2026, rejecting all refund claims from owners who bought FSD between 2019 and 2022 🔥 The decision came at the trial-court level 🔥 The owners argued that FSD promised a car that could drive itself under all conditions. They pointed to years of delays, the “FSD (Supervised)” rebranding, the shift to subscriptions, and concerns that older HW3 hardware could never deliver the promised capabilities. The court found that buying FSD did not mean Tesla had contractually promised completely driverless, Level 5 autonomy. Instead, buyers purchased existing driver-assistance features and access to future improvements as technology and regulatory approvals allowed, with no fixed delivery date. Judges noted that owners already had features such as lane keeping, automatic parking, and Summon. FSD v14 Lite recently added traffic-light and sign recognition and city driving to the plaintiffs’ HW3 vehicles. Differences in capabilities between HW3 and newer HW4 vehicles did not mean Tesla could no longer fulfill the contract. The court rejected the argument that delays justified canceling the purchase because buyers knew the delivery timeline was uncertain. It also found insufficient grounds to hold Tesla Korea responsible for the delays, given the need to comply with regulatory requirements. Finally, judges rejected claims that the missing features amounted to a defect, finding that owners had received existing features and the right to future additions. All refund requests were denied.

Source:

STRONG SIGNAL 🚨 Terafab is moving forward at lightning speed because the engineers themselves have the authority to make decisions 🔥 According to Korea’s The Elec, a Korean supplier executive said that even questions sent during the early morning hours in the US get answered within 30 minutes to an hour. In his view, working with Terafab is completely different from dealing with conventional semiconductor customers. An executive at an overseas equipment maker went even further: “I’ve never seen this level of execution speed in the semiconductor industry.” Building a fab involves countless questions like these, across every tool and material that goes into it. When each one gets a fast answer, the time saved keeps compounding, and that could add up to a serious head start before a single wafer is even processed. The chip industry spent decades perfecting how a fab should run. Terafab is being built by engineers who answer a supplier’s question before the US is even awake and are trusted to decide on the spot, and that may turn out to be the most disruptive process change of all.

Source:

Reference article:

BOOM 🚨 @SpaceX has patented a cosmic card dealer that fans out an entire deck of satellites with one slow spin 🃏 A rocket stage punches into orbit with sixty satellites stacked on top like a fresh deck of cards. Now every one of them has to leave, and not one can so much as graze its neighbor on the way out. That exit makes or breaks the whole trip. Packing dozens of satellites onto one rocket splits the cost of the ride, but only if the rocket can set every one of them free cleanly. The classic fix is a dispenser, a dedicated rack that cradles each satellite through liftoff and pushes them out one at a time. It gets the job done, but it’s bulky and heavy, and every pound it adds is a pound of payload the rocket can’t carry. Stacking the satellites right on top of each other ditches the rack but swaps weight for complexity. Fasteners lock each layer to the next, a controller releases them layer by layer, and springs nudge them apart. Those springs have to be tuned so each satellite leaves a touch faster than its neighbor, and getting that right makes the whole setup a headache to install and control. SpaceX scraps the layer-by-layer routine and lets the entire stack go in a single instant. Then the rocket’s own motion hands every satellite a different speed, the exact job those tuned springs used to do. It starts by turning the satellites into their own scaffolding. One version splits sixty flat satellites into two side-by-side piles of thirty, with one pile bumped up by the thickness of a single satellite. The stack climbs like a staircase, and each satellite overlaps its neighbors the way bricks interlock in a wall. Sturdy cylindrical fittings on each satellite’s edges line up one on top of the next, building four columns at evenly spaced points around the stack. The brutal forces of launch pour straight down those columns, so the satellites carry their own weight and there’s no separate frame holding them up. Every joint in those columns is a cone resting in a cup. The cup is just a hair wider than the cone, so the two seat snugly but never jam the way a cork wedges into a bottle. That fit stands up to squeezing, sideways shoves, and bending on the way up, yet it lets go in orbit without any extra force. Slick finishes such as hard-anodized aluminum or a dry lubricant film can help the joints slide apart. A small wave spring, basically a thin ring of rippled metal, may also sit inside each joint to give it a nudge. Four sets of tension rods clamp the whole tower together. Each set runs the full height of the stack alongside one column, hinging at the bottom on the payload adapter, the cone-shaped base that links the stack to the rocket. At the top, each set latches onto a cap called a load head. An air- or fluid-powered piston pulls the rods tight, and they bear down on the column like two hands squeezing a stack of books, locking satellite to satellite and the whole tower to the rocket. Once in orbit, the second stage, the top section of the rocket that carries the stack the whole way up, cuts its engine and starts to twirl. Its thrusters ease it into a slow, steady flat spin around its center of mass, its balance point, like a baton turning in a level circle. At roughly three degrees per second, one full turn takes about two minutes. Before letting go, SpaceX lines up the exits. The stack sits turned about 45 degrees, angling the four rod sets so two will shoot ahead along the orbit and two will drop behind. Without that twist, two sets could drift off sluggishly and tag along with the stage and satellites, risking a collision. Then the rods let go. The tension drains away, the load heads unlatch from the tops of the columns, and the rods swing outward on their hinges like a flower opening until they’re completely clear of the stack. They may even break free of the adapter and sail off on paths of their own. Now the spin takes over. It’s the same physics as a merry-go-round, where the kid on the rim whips around while the kid by the hub barely moves. The satellites at the top of the stack ride farthest from the stage’s center of mass, and speed grows right along with that distance. So the top satellites leave fastest, the ones nearest the rocket drift off slowest, and the tower unzips into a spreading stream without a single tuned spring setting the pace. Where a satellite sits becomes how fast it goes, which can make the whole release cheaper, lighter, and easier than a bank of carefully calibrated springs. Back on that rocket stage, the deck isn’t a deck anymore. Sixty satellites stream away in a long, widening fan, the fastest already pulling ahead and the slowest trailing behind, and not one of them needed a dispenser to send it on its way. The old way needed a heavy machine to deal every card by hand. SpaceX just lets go of the whole deck mid-spin, and physics deals each card its own speed.

🚨 When will SpaceX first release a whole stack like this? The answer has a fun twist, because it's already happened, back on SpaceX's very first Starlink launch. On May 23, 2019, a Falcon 9 carried 60 Starlink satellites to orbit packed into one tight stack. About an hour after liftoff, SpaceX released all 60 at once, and the stack slowly peeled apart in orbit like a zipper coming undone. So why was this patent only granted on October 27, 2026? A grant date marks when the patent office approves a set of claims, not when the idea was born or first flown. This patent belongs to a family that starts with a provisional application SpaceX filed on May 14, 2019, just nine days before that launch. The full application followed in 2020 and became an earlier patent, and this one is a continuation filed in 2024. It shares the same original disclosure but protects a different set of claims, which is how companies keep locking down new angles of an idea for years. The patent never mentions Starlink by name, but the details line up almost perfectly. A sixty-satellite stack released all at once, filed nine days before a sixty-satellite launch that did exactly that, is about as close to a match as you'll get without SpaceX saying so outright. So the card trick isn't coming someday. It opened the show seven years ago, and the patent office just finished checking the deck.

Source:

SOS 🚨 @Tesla saw one of the steepest valuation drops in Interbrand’s 2026 Best Global Brands report, falling 25 spots and losing nearly 38% of its brand value 😱 It’s the third straight year of decline. Tesla was #12 at $45.5 billion in 2024, slipped to #25 at $29.5 billion in 2025, and has now landed at #50 with $18.3 billion. That’s about 60% of its brand value gone in just two years, and Ferrari is right on its heels, only one spot and $600 million behind. To be fair, it was a tough year for every carmaker. The auto industry lost nearly $57 billion in brand value, with Toyota, Mercedes-Benz, BMW and Honda all down between 18% and 22%. Even so, no brand in this year’s top 100 fell harder than Tesla. Keep in mind this isn’t market cap. Interbrand blends projected economic profit, how much the brand drives buying decisions, and an overall brand strength score. Put simply, it’s trying to measure how much the name alone sells cars. Interbrand doesn’t mince words about what went wrong. People bought into Tesla’s philosophy: no dealers, cars that get better overnight with software updates, and a charging network built just for them. But those things can be copied, and plenty of rivals now plug into Tesla’s own charging standard. The report argues that Tesla’s identity was never bigger than one person, so once Elon became a polarizing figure, the brand had nothing else to lean on. It’s not all doom and gloom, though. Interbrand’s CEO points to the Semi, Optimus, Cybercab and Roadster as proof of Tesla’s ambition. He says the real opportunity is turning that tech edge into experiences people actually use, and he calls Tesla one to watch heading into 2027. Tesla built its brand by showing people the future first. Next year’s ranking will tell us whether actually delivering it can win that value back.

ROBOTAXI APP NEWS 🚖 Tesla has updated its Robotaxi iOS app to version 26.10.6, and a peek under the hood reveals some exciting developments: ✅ Airport-specific booking flows The app now includes airport-aware pickup and drop-off logic, complete with an airline selector featuring a "Popular Airlines" section. This signals preparations for seamless airport trips, helping passengers easily select the right drop-off terminal or pickup spot as service expands. ✅ Trip sharing The code includes a dedicated trip-sharing link, an active sharing status indicator, and a remote feature flag. This sets the stage for sharing live ride updates with friends and family, providing an extra layer of peace of mind when riding in a driverless vehicle. ✅ Referral rewards and the Robotaxi coin Tapping the animated gold coin unlocks "Refer and Earn," giving riders quick access to shareable invite links and their credit history. A dedicated notification flow kicks in after an invited rider completes their first trip, complete with celebratory reward messages and on-screen confetti. This lays the groundwork for a built-in referral program designed to fuel viral, word-of-mouth growth, with offer terms and eligibility managed remotely behind the scenes. ✅ Grok and personalized voice settings The app supports push-to-talk, "Hey Grok" voice commands, and customizable spoken-announcement preferences, with availability tied to vehicle capabilities and prior ride history. This paves the way for a more personalized, hands-free experience that seamlessly carries over between compatible rides. ✅ Theater and YouTube search Remote feature flags connect Tesla's system to entertainment search controls, along with checks to verify if front-screen video playback is allowed. This could allow Tesla to roll out easier media discovery across eligible vehicles, giving passengers more ways to enjoy the ride. ✅ Model Y and Cybercab booking choices Vehicle selection logic accounts for available ride options, pricing, and Tesla's default choice, enforcing seating rules of four passengers in the Model Y and two in the Cybercab. This gives Tesla the framework to match each group with the right vehicle as the fleet expands.

HUMANOID ROBOT NEWS 🚨 According to reports from the October 28 SAECCE conference in Suzhou, XPeng Vice President Gu Jie said the company had unveiled a solid-state battery designed specifically for robots 🤖 It prioritizes high energy density in a robot’s limited internal space, resistance to impacts and nail penetration, and protection against thermal runaway spreading from one cell to another. In an all-solid-state battery, a solid material replaces the liquid electrolyte that carries ions between the electrodes. Semi-solid batteries use both solid and liquid electrolytes, and that distinction matters here. For humanoid robots, these are fundamental engineering challenges. Every kilogram of battery is another kilogram the robot has to support, balance, and move. Higher energy density lets designers store more energy at the same weight, or cut weight while keeping the same capacity. Energy per liter matters just as much, because space inside a humanoid’s body is tight. A more compact battery leaves more room for actuators, electronics, and cooling. Those gains only count at the level of the finished machine. Protective structures, cooling, electronics, and the compression hardware some solid-state cells need to keep their layers pressed together all add weight and take up space. The battery also has to deliver bursts of power for walking, lifting, and catching balance, manage heat, recharge efficiently, and hold its capacity through repeated use. Higher energy density solves only part of the problem. Safety takes on new importance when the battery walks into a room full of people. A humanoid can fall, hit equipment, or get damaged while working nearby. Gu also reportedly said XPeng had “reworked its battery standards.” The report offers no specifics, but that could mean testing batteries around a robot’s real working life: falls, repeated bumps, and close contact with people. A pack designed to contain a single failing cell could keep a local failure from turning into a larger emergency. Solid-state chemistry alone doesn’t guarantee that, so pack design still matters. XPeng’s timeline adds important context. When the company announced an all-solid-state battery for its next-generation IRON in November 2025, it reported that the pack withstood one hour at 250°C without thermal runaway, a 300G impact, and penetration by a 3 mm nail without catching fire. Those results apply to the 2025 design, and the October report doesn’t say whether it’s the same pack. XPeng announced the opening of its robot production line in September 2026. Its latest schedule calls for scaled production in late 2026, with initial deliveries in the first half of 2027. But which battery will actually go into production? In September, Jipian Lab reported that XPeng had tried unsuccessfully to source all-solid-state cells from BYD, Gotion, CALB, and EVE. It then selected a high-energy semi-solid cell in a pouch format, rather than the cylindrical format common in robot batteries. The reported problem was finding production-ready all-solid-state cells. Jipian also quoted a researcher who described the industry’s all-solid-state cells as largely laboratory samples, with almost none ready for outside testing. XPeng has not publicly confirmed the reported switch. The Chinese wording makes this especially interesting. The October report uses 固态, or “solid-state.” China’s draft national standard for EV batteries separates batteries into three categories: liquid, mixed solid-liquid, and solid-state. It also accepts “semi-solid” as another name for mixed solid-liquid, and “all-solid-state” as another name for solid-state. Under that draft’s terminology, “solid-state” means all-solid-state. Everyday industry usage is looser. Was Gu referring to an all-solid-state battery, or the semi-solid cell Jipian reported? The remark alone doesn’t answer that question. The connection to Tesla Optimus is now concrete. In its October 21 Q3 update, Tesla said it had begun building Optimus on its high-volume production lines in Fremont, describing the robots coming off those lines as its first production-quality units. XPeng and Tesla announced major humanoid production milestones within weeks of each other. Both programs now face the same practical test: how much useful work can the complete machine do between charges, safely and reliably? A better battery could turn more of a humanoid’s intelligence into productive hours.

BOOM 🆘 Two House bills want to bring back the federal EV tax credit, and one of them would double it to $15,000 and make it permanent 🔥 1⃣ The bold one comes from Rep. Greg Landsman (D-OH), who introduced the 21st Century Vehicle Choice Act on October 22. It would bring back the credits that ended for vehicles bought after September 30, 2025, and make them twice as big. The $15,000 headline is only the start, and the fine print gets even more interesting: 🔋 The max credit for new EVs doubles from $7,500 to $15,000. Each of the credit’s two halves (critical minerals and battery components) goes from $3,750 to $7,500, so a car still has to pass both sourcing tests to get the full amount. The old price caps stay in place. 🚗 The used EV credit doubles from $4,000 to $8,000 and can cover up to 60% of the sale price instead of 30%. 💰 Income limits for new EVs drop to $225,000 for married couples filing jointly (down from $300,000), $169,000 for heads of household (down from $225,000), and $112,500 for single filers (down from $150,000). 1️⃣ Each taxpayer could claim the new EV credit only once in their lifetime. The way the text reads, anyone who already claimed it in an earlier year may not qualify again. ⚡ The EV charging equipment credit rises from 30% to 50% of the cost. ♾️ All three credits become permanent. 2⃣ The second bill takes a very different route. The Energy Bills Relief Act from Reps. Sean Casten (D-IL) and Mike Levin (D-CA) has more than 160 House Democrats behind it. Instead of building something new, it hits the undo button on the part of last year’s tax law that ended the credits. Here’s everything that undo button would bring back: 🔋 The new EV credit returns at up to $7,500, with the same two $3,750 halves and the same sourcing tests and price caps as before. 🚗 The used EV credit returns at up to $4,000, capped at 30% of the sale price. 💰 Income limits for new EVs go back to $300,000 for married couples filing jointly, $225,000 for heads of household, and $150,000 for single filers. 🚚 The commercial clean vehicle credit for businesses and fleets comes back too. ⚡ The EV charging equipment credit returns at 30% of the cost. 📅 All of these credits run through 2032 instead of becoming permanent. The EV credits are just one piece of that bill. It’s a 419-page energy package that also goes after electricity bills, grid upgrades, and clean energy permitting. ⚠️ So how close is any of this to reality? Not very close yet. Both bills are still sitting in committee, the very first stop for any bill in Congress. To become law, a bill has to pass the House, then the Senate, and then get the president’s signature. Support is also thin where it matters most. No other member of Congress has signed on to Landsman’s $15,000 plan yet. The top House Democrat, Hakeem Jeffries, hasn’t put his name on the Casten-Levin bill either. Then there’s the price tag. Republicans, who control both the House and the Senate right now, ended these credits just last year. Cutting the new EV credit alone was expected to save the government about $78 billion through 2034. Doubling the credit would cost far more. An analyst at the Cato Institute, a libertarian think tank, puts it at roughly $380 billion over ten years. One bill has the bigger number, the other has the bigger crowd, and neither one has the votes to become law yet.

Sources: 1/ 2/