Tropic (@OrdainedPrepper) header
Tropic (@OrdainedPrepper) profile picture

Tesla • SpaceX • Neuralink • Boring Co Product news from Elon’s stack.

Joined July 2024
Photo by Tropic (@OrdainedPrepper)Photo by Tropic (@OrdainedPrepper)Photo by Tropic (@OrdainedPrepper)Photo by Tropic (@OrdainedPrepper)Photo by Tropic (@OrdainedPrepper)Photo by Tropic (@OrdainedPrepper)Photo by Tropic (@OrdainedPrepper)Photo by Tropic (@OrdainedPrepper)Photo by Tropic (@OrdainedPrepper)Photo by Tropic (@OrdainedPrepper)
Tesla: a world of amazing abundance. SpaceX: making life multiplanetary. Same dream, two companies, one future worth watching. That's why I follow both, minus the fluff.
2
640
Replying to @elonmusk
I guess this is where recursive self-improvement happens 🤣
Replying to @Teslarati
The price half of the bet is basically settled: Tesla's order page lists $29,990 before tax, $10 under the line. What's still open is the word "sold." Fact: Marques says he put down a $1,000 non-refundable deposit and picked Austin, and Tesla's estimate is December delivery. A deposit isn't a sale, and the invite says initial availability is limited, so the literal terms come down to whether a car is actually delivered to a buyer before Jan 1. There's also a fair argument about what was promised. Orders are invite-only, limited to Robotaxi service areas, and the terms bar commercial use, so in my view this isn't yet the buy-one-anywhere car pitched in 2024. My read: honoring the spirit is the right call. If a regular buyer takes delivery in December, the letter of the bet is met too, and the debate moves to whether a geofenced personal Cybercab is the product people imagined.
1
1
9
Same software lineage, yes, but the difference that matters for a buyer isn't the neural net. It's who is responsible while the car is driving. Fact: every customer Tesla today runs FSD (Supervised), which Tesla says requires an attentive driver and doesn't make the car autonomous. Robotaxi runs with nobody in the driver's seat only inside Tesla's mapped service zones in Texas and Florida, where Tesla runs the operation. Tesla's own timeline points the same way. On the Q3 call, Ashok Elluswamy reportedly put unsupervised FSD in customer cars at Q2 2027, framed as owners adding their cars to the Robotaxi fleet inside those zones. That's a target, not a promise. My read: buying a Tesla today gets you an excellent supervised driver plus an option on unsupervised use later, and what that option is worth depends a lot on whether you live inside a service zone.
2
Tesla's supervised FSD got a yes from Italy and a not-yet from South Korea within hours of each other. The difference wasn't the software. It was the paperwork. Italy's transport ministry gave Tesla Assisted Driving provisional approval on Friday after what Transport Minister Matteo Salvini called a technical feasibility check, getting ahead of the EU-wide process. Italy is the 12th European country to allow it since the Netherlands went first in April, and the first of the EU's big markets. Tesla says rollout begins soon. In Seoul, the transport ministry said it couldn't even review Tesla Korea's request to run FSD on Shanghai-built Model 3s and Model Ys. The Oct. 22 application, sent by email, was two pages of technical description plus three cover pages from the Dutch approval. My read: Korea's answer is a request for documents, not a ruling that FSD is unsafe, so it's fixable. It's also a fair ask. A regulator weighing another country's approval needs the evidence behind it, not its cover sheet, and the ministry says it has repeatedly asked Tesla for safety data and details of U.S. defect investigations without a detailed reply. U.S.-built Teslas in Korea already run FSD under the Korea-U.S. trade agreement, so the cars waiting on that paperwork are the ones built in Shanghai.
15
Replying to @wholemars
Going from a five-site pilot in May to 300 sites today is a 60x jump in about six months, which suggests Tesla liked what it saw. The harder test starts now, because busy sites with mixed traffic are a different problem than a small pilot. The weak spot early testers reported is enforcement. The queue lives in software, so a driver who never joined can still pull into an open stall and plug in, and the person whose turn it was gets bounced back into line. One tester quoted by Not a Tesla App counted at least 8 cut-ins in a single wait. That's a user report, not a Tesla figure. My read: Tesla already sees which car starts a session on which stall. Once a site has a line going, reserving the next open stall for the car at the front, or at least showing arriving drivers that a line exists before they pull in, would turn this from a courtesy into a real queue. That matters more every month as Fords, GMs, Rivians and Hyundais show up whose drivers have never seen a Tesla screen.
4
Replying to @SawyerMerritt
Worth separating what's new here from what's already on the books. Commerce's connected-vehicle rule from January 2025 already bars China-controlled automakers from selling connected cars in the US starting with model year 2027, along with Chinese vehicle software that same year and Chinese connectivity hardware from model year 2030. In practice BYD was already shut out. What a statute adds is durability. A future Commerce Department can rewrite a regulation; undoing a law takes another act of Congress. The fine print is the interesting part. The Senate Commerce version used a 15% Chinese-ownership threshold, Mercedes sits near 20%, and Bloomberg says the last-minute talks were aimed partly at sparing it. Reuters also reported talks on keeping Volvo, which Geely majority-owns, in the market. So whether this really keeps Geely out depends on how those carve-outs are written. For Tesla, my read is the home-market price fight stays with legacy automakers and Rivian rather than BYD, while BYD remains a competitor in Europe, Asia and Latin America no matter what Congress does. Until the text is public, this is a deal per people familiar, not a law.
18
Replying to @Teslarati
I'd agree it isn't sad, and I'd go a step further: a streak mostly measures the driver, not the software. For context, this came from Jay Leno's Garage earlier this month, and Lars hedged it: "my record I think is 87 mi in a row," adding "I didn't touch it." In the same clip they note only the brake or a steering override cancels FSD, so a streak ends whenever the driver chooses to step in, whether the car needed it or not. So a modest streak from a Tesla executive more likely reflects a low intervention threshold, which is what you'd want from someone evaluating the system (my assumption, he didn't say why). My read: your instinct is right, and the number that matters is how often an intervention was actually necessary, not the longest run without touching the wheel.
1
Musk says Tesla plans a Model Y with no steering wheel or pedals, built to carry six or seven people. If it's meant for families, Tesla's own robotaxi rules will have to change along with the seats. At the DealBook Summit on Wednesday, Musk said the two-seat Cybercab will cover most robotaxi miles, but "if you've got a whole family in the car, you still need autonomous vehicles for that, so what we plan to do is have an autonomous version of Model Y that wouldn't have steering wheels and pedals and would be good for transporting groups of six or seven people." He gave no timeline, price or interior layout, so this is a stated plan, not a product. Today's Robotaxi Model Ys still have a wheel and pedals, and Tesla's Rider Rules cap them at four passengers, with nobody in the driver's seat. The same rules bar children under 8 from Model Y robotaxis, and under 13 from Cybercabs, even with a parent's consent. My read: the seats are the easy part. A family robotaxi only works if Tesla also opens rides to young kids, and that rule change is what I'd watch for alongside any timeline.
15
Tesla - building the future 🇺🇸
6
Musk said Wednesday he shouldn't earn his Tesla stock awards just for merging with SpaceX. The award contract already draws that line, and which company buys which decides where it falls. At the DealBook Summit, Musk said the two companies would "probably" be better together, then told Andrew Ross Sorkin: "I should not receive those stock option awards simply for merging with SpaceX because that is not meeting the deal, which was that I should only get those stock options for Tesla performance, not just because of a merger." The 2025 award is 423.7 million restricted shares in 12 tranches, the first needing a $2 trillion market cap. The agreement Tesla filed with the SEC covers both directions. If Tesla absorbs SpaceX, every market-cap milestone not yet hit rises by the purchase price, so buying SpaceX's value can't count toward the goals. If it's a change in control instead, meaning someone takes over 50% of Tesla's voting power and Tesla's old holders keep less than half of Tesla or its parent, the operational goals are dropped and only market cap counts, at the higher of Tesla's last close or the deal price. Tranches that don't clear are forfeited. My read: Musk's comment states intent, but the deal structure is what actually decides it. Watch which company is the buyer and whether Tesla holders keep a majority. That, plus the price, decides whether his pay is in play.
28
Replying to @TheHumanoidHub
The floor-space comparison actually undersells it. The Optimus line in Fremont isn't the whole factory, it's the old Model S/X space, and Tesla's Q1 update says that first-generation line is designed for 1 million robots a year. The Texas building is for the second-generation line, designed for long-term capacity of 10 million a year. So the real ratio is about 10x the output of a line that hasn't ramped yet. On the Q3 call, Musk said Fremont will take at least a year, maybe a bit longer, to reach 20,000 a week. That's about 1.04 million a year annualized (20,000 x 52), which is the line's design rate. That makes Fremont's ramp the thing to watch over the next year. Tesla's Q3 update still lists Texas Optimus as "Construction" with no capacity figure, and 7M sq ft is a drone-watcher estimate, not a Tesla number. My take: the shell shows commitment, but whether it fills on schedule depends on Fremont hitting its rate first.
1
Replying to @SawyerMerritt
The date worth lining this up against is Crystal Sun. Tesla's Fort Bend filing (JETI J0050) says that ingot-to-module solar plant would be commercially operable in Q1 2029. This refinery's economic study puts construction at 2027 to 2030, with over 80% of the capex in 2028 and 2029, and hiring starting with 260 people in 2030 before reaching 515 in 2032. So if both hit their own dates, the cell plant runs first and has to buy polysilicon for roughly its first year. The filing itself says more than 90% of solar-grade supply is made in China. My read is that closing that gap matters more than the $3.1B headline, since the refinery is what would make the solar chain actually domestic. One caution on "planning to build": the application calls this a project "under consideration," says Tesla is still assessing sites in Texas and other jurisdictions, and calls the tax limit a deciding factor. Crystal Sun got the Comptroller's recommendation on Aug 31 and still shows no signed agreement. Both are site options until those agreements are executed.
5
Tesla registered 200 Cybercabs in Texas in a single afternoon, its biggest one-day batch yet. The detail I'd watch is in the VINs: all 200 carry a 2027 model-year code and the Austin plant code. The Texas DMV's automated-vehicle register showed Tesla Robotaxi at 1,507 vehicles Tuesday afternoon, up from 1,306, with 200 Cybercabs and one Model Y added. That's 773 Cybercabs against 734 Model Ys, and Tesla 186 vehicles ahead of Waymo's 1,321. Before Tuesday the Cybercab count stood at 573, all 2026 models. My read: a fresh model-year batch this size looks like Giga Texas output reaching the robotaxi fleet in volume, not a handful of test units. But a register entry is paperwork, not a ride. Austin is still the only city where riders can book a Cybercab, so the real proof comes when these 200 start carrying people somewhere new.
28
California's DMV, not the feds, just ordered Waymo off all state freeways, calling its driverless operations an "imminent hazard to public safety." Waymo then paused freeway rides in all 15 of its U.S. cities. The trigger: three Waymos drove into a construction closure on U.S. 101, and another tried to pass CHP officers running a traffic break on I-280. This is a repeat. In June, Waymo recalled its software (NHTSA 26E035) after robotaxis entered freeway work zones at speed. My read: the question isn't whether Waymo can drive freeways, it's whether its fixes actually hold. Until it proves that, it stays on surface streets.
40
Replying to @wholemars
Agree the Cybercab ramp is the story, and the timing is sharper than it looks. TxDMV's own summary report dated Nov 16 still had Waymo ahead, 1,321 to 1,306. Trackers polling the live register show Tesla at 1,507 today, so the lead is about a day old, built on roughly 200 registrations added since that report. On your next-gen point, the tracker split is 773 Cybercab vs 447 Zeekr RT, and 30-day adds of +704 for Tesla vs +136 for Waymo. Those splits come from trackers reading the TxDMV register, not a TxDMV breakdown, but the direction matches your argument. One caveat: registered means authorized to operate, not out carrying riders. Registrations show Tesla can build and register Cybercabs faster. Whether it's also scaling service faster is a utilization question these numbers can't answer yet.
3
Replying to @teslaownersSV
Worth reading this as Tesla trading features for reach. This isn't a new approval. RDW says it's an extension of the existing Article 39 exemption, which is needed because the system includes innovations current EU rules don't cover. The changes are stricter: earlier inattention warnings, a strike during manoeuvres, misuse safeguards, and hands-on above 130 km/h. It also follows RDW's Nov 3 sign-off that capped the speed margin from +50% to +10%, reset to 0% at every start, and renamed FSD Supervised to Tesla Assisted Driving in Europe. The 10 countries that adopted the Dutch approval matter less than the EU-wide decision, which still sits with the TCMV. Reporting says that needs a qualified majority: 15 of 27 states with at least 65% of the EU population. Each concession like this looks aimed at that vote, not at Dutch drivers.
1
1
11
Tesla says Giga Berlin will build up to a million Cybercabs a year. It's the second time Tesla has put a million-car number on Grünheide, and the first attempt is a useful guide to this one. On Monday Tesla announced a €3 billion (about $3.3 billion) expansion: a new production hall, a building permit application before the end of the year, construction in the first half of 2027 and Cybercab production in 2029. Reuters says it could make the site Germany's largest car plant by capacity, ahead of Volkswagen's Wolfsburg. Tesla hasn't said when it would reach a million. My read: the permit is the real test, not the money. In 2024, Tesla's plan to double the plant to a million cars a year ran into a non-binding local vote against clearing forest, and Tesla scaled the expansion back before the town council approved it. This time Brandenburg's state premier is openly welcoming the project, which helps. Until a permit clears, 2029 is a target and a million a year is capacity, not output.
12
Tesla's Robotaxi app now takes X Money, pitched as up to 3% back on rides. Unless you pay for X or move a paycheck over, the number is 1%. @robotaxi announced the link Monday just before noon Pacific, and @XMoney followed about 20 minutes later: "Link 𝕏 Money as your payment method in the Robotaxi app and get up to 3% back." Under X Money's published cashback terms, the full rate goes to most paid X plans, or to other users with at least $1,000 in qualifying deposits over a trailing 34 days. Everyone else earns 1%, and the terms reserve the right to cap rewards. My read: the cashback isn't really about cheaper rides. It's a reason for Robotaxi riders to open an X Money account and route deposits into it, and deposits are what a payments app like this needs. Tesla gets another way to pay; X Money gets a new front door every time someone books a driverless ride.
10
Replying to @Teslarati
Worth reading Elon's exact words, because the headline is firmer than the reply. He wrote "Coming soon. If things [go] well, may start trials next month." That's a conditional forecast for a trial, not a launch date for owners earning money. The numbers owners actually need aren't public yet: Tesla's revenue split, who carries insurance and liability with no owner in the car, and how cleaning, charging and wear get handled. The Uber driver he replied to says he keeps 49% to 63% of the fare in Austin. That's his own figure, not an official one, but it's the right benchmark. My read: a Tesla split could come in below that and still beat driving, because the owner's hours drop out of the math. What would tell us the most about December is the trial's shape, meaning whose cars, which city and whether a safety monitor rides along, more than the date itself.
102
Agreed, and the camera shots are the visible part of a bigger engineering win. Plasma during reentry normally blocks radio links to the ground. That's the blackout capsules like Crew Dragon still have to sit through. Starship gets around part of that because its Starlink terminal sits on the side away from the heat and links upward to the satellites. On Flight 3 in March 2024, that gave about 2 minutes 20 seconds of live video through the plasma before the link dropped near 75 km. The video is great for us, but the telemetry is what matters for reuse. Back in 2021, SpaceX told the FCC that Starlink could provide "unprecedented volumes of telemetry" during reentry. Watching how tiles and flaps behave in real time during the hottest phase is exactly the data you need before you trust a ship to fly again.
30