Dan🪐

⏫

@billmaher The Great mystery of the physics defying & vanishing Orbs!

If you don’t believe aliens are here and observing us, maybe you’re the conspiracy theorist.

@billmaher No, those aren't aliens defying momentum, those are just really big laser pointers.

@tavi_chocochip @elonmusk Accurate video generation is a physics simulator, robot training data generator, universe understander, ect... Visual intelligence might be more powerful than language-based intelligence, just much less computationally efficient.

@elonmusk How so? What’s the goal with image generation, beyond creating content for social media?

Still trying to figure out why this is so important to Elon

SpaceX main page, 2002

Our FU Superchargers (Folding Unit) save 20%+ on cost, improve build quality, and deploy 2X faster. The majority of cost savings are in civil, electrical and logistics cost. It's why we're spending so much time improving our pre-assembled Supercharger units. With FU Superchargers, on-site work is further reduced, including no more DC busbar connection and no Tesla service technician needed for commissioning. We can ship 16 Supercharger posts per truck (previously 12), making the logistics cost of shipping pre-assembled Superchargers now cheaper than traditional construction 🤯. This is just Rev1, we'll have Rev3 by next quarter. Every little improvement matters for the scale needed for 100% EV adoption.

To return Americans to the Moon, NASA is shifting to an iterative, execution-focused approach – just as we did during Apollo.  We are standardizing rocket architecture, embedding NASA expertise across industry, and increasing launch cadence to support sustained lunar operations.   We are sending a demand signal for crewed missions beyond Artemis V, with at least two providers capable of bringing astronauts to the surface every 6 months.  The goal is not just to reach the Moon, but to stay.  America will never give up the Moon again.

NASA is advancing nuclear power and propulsion in space to accomplish President Trump’s national space objectives. With SR-1 Freedom, launching in 2028, we will demonstrate nuclear electric propulsion and deliver SkyFall helicopters to Mars. In collaboration with @Energy, these capabilities are key to future missions to Mars and beyond. A new chapter of deep space exploration begins. 🚀

The right thing to do is obviously increase the price of the car by $13, or at least be required to tell the customer that a $13 optional add-on will improve safety, so at the very least required to disclose ALL safety information they are aware of. Agreements cannot be voluntary if they are not honest. This is a solution to the problem that maintains the principle of voluntarism that Milton would defend. Milton was right that the argument was not based on a principle that applies generally, but he failed to show that his principles can solve the problem too, at least within this clip.

NASA is leading the greatest adventure in human history, and it has only just begun.   On Tuesday, we’ll share our plans for the future of NASA across many of our programs.   We’ll see you tomorrow morning for Ignition.

@elonmusk @pbeisel 4.20 sq miles = 117.1M sq ft

@pbeisel Yeah, 100M sq ft is the right order of magnitude

Terafab First, the Advanced Technology Fab— where new manufacturing technologies are proven out. Think of it as the prototype line. Then Terafab itself: By a wide margin, it will be the largest factory in the world. Let's guess at sizing: Quick math tying it to the 1 TW target + 250W AI5 chips 1 TW = 1,000,000,000,000 watts of AI compute capacity. Each AI5 chip draws ~250W Chips needed annually for 1 TW scale: 10¹² W ÷ 250 W/chip = 4 billion chips/year. To produce 4 billion AI5 chips per year (the volume needed for 1 TW of compute at 250 W/chip), a Terafab-scale facility modeled on current TSMC fabs would require 3,000–5,000+ acres or land and 280–350+ million square feet of total building/floor space. Obviously innovations must be applied. I'd hazard a guess at a 100 million sq ft size factory.

@SawyerMerritt He said "on the order of" 100 million square feet. If it is 117.1 million square feet, That would make it 4.20 square miles.

Holy moly, Tesla/SpaceX’s Terafab will be ~100 million square feet, which is 10X larger than the main Giga Texas building, which itself is ~1 mile long. No wonder the Terafab won’t fit on the existing Giga Texas campus 🤯

@aaronburnett Space will probably be cheaper than terrestrial in 2028

Macrohard or Digital Optimus is a joint xAI-Tesla project, coming as part of Tesla’s investment agreement with xAI. Grok is the master conductor/navigator with deep understanding of the world to direct digital Optimus, which is processing and actioning the past 5 secs of real-time computer screen video and keyboard/mouse actions. Grok is like a much more advanced and sophisticated version of turn-by-turn navigation software. You can think of it as Digital Optimus AI being System 1 (instinctive part of the mind) and Grok being System 2. (thinking part of the mind). This will run very competitively on the super low cost Tesla AI4 ($650) paired with relatively frugal use of the much more expensive xAI Nvidia hardware. And it will be the only real-time smart AI system. This is a big deal. In principle, it is capable of emulating the function of entire companies. That is why the program is called MACROHARD, a funny reference to Microsoft. No other company can yet do this.

@peterwildeford xAI will catch up this year and then exceed them all by such a long distance in 3 years that you will need the James Webb telescope to see who is in second place

@InTeslaITrust @elonmusk @rohanpaul_ai It said over the next 10 years, So that would include spending based on revenue between now and 2036. Elon thinks in orders of magnitude and on exponential scales so he's likely being serious.

@elonmusk @rohanpaul_ai Not sure if you are being sarcastic or serious 🧐 if 10T or you are not even trying then I guess Tesla is merging with SpaceX, right? We also need to get to the 10T of AI spending!

Google is building a nearly untouchable advantage by controlling every layer of the AI stack, from the chips up to the power grid. This Forbes piece shows Google has a massive plan to spend $1.9T on data centers and hardware over the next 10 years. They are ramping up its capital spending from $90B in 2025 to a staggering $185B this year alone. A major part of this budget goes toward TPUs. These chips have become a hot commodity, with other AI companies now renting them out through Google's cloud services as a solid alternative to Nvidia hardware. To handle this growth, Google is shifting toward modular data center designs, which are like standardized blueprints that can be built quickly all over the world. They are also signing big deals with utility companies to secure the massive amounts of power these facilities need to run 24/7. Most importantly, Google makes enough profit to fund these projects using its own cash reserves rather than relying on outside investors. --- forbes. com/sites/richardnieva/2026/03/02/googles-data-center-buildout-could-top-1-trillion/

First images of the Tesla Cybercab body-in-white - 60% fewer parts than the Model Y ✨ What you are looking at is possibly the simplest vehicle structure Tesla has ever designed: two Gigacastings, stamped panels, and minimal joining. The rear underbody casting has no integrated wheel arches - flat lateral walls only. The die design doesn’t require lateral slide cores. Tesla stripped the casting design to its absolute minimum. A robotaxi has no steering column, no driver controls. That freedom allowed Tesla to start designing the casting from zero, introducing new solutions, some of which counterintuitive to the progress of Gigacasting. The unboxed manufacturing process - where the front section, rear section, and side panels are assembled in parallel and joined only in final assembly - is built around these castings. Read the full breakdown on this here: industryarsenal.com/p/tesla-s-most… And for the complete picture, every vehicle that uses Gigacasting and every Gigacasting machine installed worldwide, visit 📊 The Gigacasting Database industryarsenal.com/database

Hermeus is developing a hypersonic aircraft the way SpaceX is developing Starship: by iterating quickly and focusing on what matters. Looking forward to their next milestones!

American Dynamism 🇺🇸

A space jellyfish was seen far and wide across the U.S. east coast this morning as a Falcon 9 rocket lifted off from Florida at 5:52am ET and climbed into the rising sun 🚀🪼❤️ Check out the new tool to forecast these - started off with a great one! jellyfish.johnkrausphotos.com

@juppjupp00 @NASAAdmin This announcement is clearly about fixing the culture that causes delays. He literally said that the skills needed for progression atrophies if launches are not frequent, He is literally tightening the schedule.

@NASAAdmin Delay, delay, delay, and more delay. Let’s tune in to 2030 for a new delay! The Moon will be bored and leave before NASA is ready.

President Trump gave the world the Artemis Program, and NASA and our partners have the plan to deliver. We will standardize architecture where possible, add missions and accelerate flight rate, execute in an evolutionary way, and safely return American astronauts to the Moon, this time to stay. This is the NASA that once changed the world. This is the NASA that will do it again.