Mars Economy @ThrustLab

The new graphics in the Tesla Spring Update look so good. I love the shadowing effect when it goes into park during FSD. So clean. 🔥 Tesla has integrated Unreal Engine into its vehicle software for rendering 3D car models.

Flight-12… the clock is ticking ⏱️

To protect passengers or cargo, the powered rear seats & trunk in Model Y will automatically pop back up if detecting an obstruction while folding.

We're hiring an exceptional designer at SpaceXAI. Work directly with me and the team to build the future of human <> agent interactions and ship beautiful, intuitive experiences across Grok and X. job-boards.greenhouse.io/xai/jobs/50419…

JUST IN: Elon Musk says "we gotta get this SpaceX IPO going pretty soon."

keyboard_arrow_left Previous keyboard_arrow_right Next

Please help make Grok Build great! Much appreciated, Elon.

Morgan @morganlinton

3 months ago

Phew, Grok Build is really thorough, pretty incredible. Relay feature for PasteLocal is done, pushing to Github now. And if you want to know what this adds, here's a bit more about it, this is a feature I really wanted in there, and honestly didn't really expect to have done

218 159 1K 8.0M 456

BREAKING: Cursor launched Composer 2.5, its most powerful coding model yet and confirmed the SpaceX AI partnership is now active. Cursor CEO Michael Truell: "This is the very start of our work with SpaceXAI. Hope to have more improvements out soon." Composer 2.5 benchmarks vs. the competition: - SWE-Bench Multilingual: 79.8% vs GPT-5.5 at 77.8% - CursorBench v3.1 harder tasks: 63.2% vs Composer 2 at 52.2% - Terminal-Bench 2.0: 69.3%, in line with Opus 4.7 SpaceX holds an option to acquire Cursor for $60 billion. This is what the early stage of that partnership looks like.

keyboard_arrow_left Previous keyboard_arrow_right Next

JUST IN: Elon Musk waited too long to sue OpenAI, jury finds on Day 14 of trial The unanimous verdict from the 9-person jury in federal court in Oakland hinged on one issue: timing. The jury found that Elon Musk waited too long to bring his lawsuit against OpenAI, Sam Altman and Greg Brockman. The statute of limitations on his core claims had already run out by the time he sued in August 2024. What the jury actually decided: - Musk's breach of charitable trust claim was barred by the statute of limitations - Musk's unjust enrichment claim was barred by the statute of limitations Judge Yvonne Gonzalez Rogers had instructed the jury that California law gives a plaintiff three years to bring a breach of charitable trust claim once they knew or should have known of the breach, and two years to bring an unjust enrichment claim. The cutoff dates were August 5, 2021 for the breach of charitable trust claim and August 5, 2022 for unjust enrichment. In her closing argument, OpenAI defense counsel Sarah Eddy walked the jury through years of Musk's own emails, tweets and public statements showing he was on notice of OpenAI's for-profit pivot well before those cutoff dates. Eddy pointed to a September 2020 Musk tweet calling OpenAI "captured by Microsoft" and arguing the company had become the opposite of its founding open-source mission. Eddy also argued that Shivon Zilis, Musk's OpenAI board observer and now a Neuralink executive, had kept Musk updated on OpenAI's restructuring discussions for years. Zilis was on OpenAI's 2021 special committee. The jury agreed. Musk knew, or should have known, about the alleged breach years before he filed suit. The merits of Musk's underlying claims were never decided. The case was lost on timing alone. (Image of the Jury is from Google)

keyboard_arrow_left Previous keyboard_arrow_right Next

Muskonomy @muskonomy

3 months ago

JUST IN: All three of Musk's claims dismissed, lead counsel signals appeal on Day 14 of Musk v. OpenAI trial Judge Yvonne Gonzalez Rogers confirmed she would accept the jury's findings as her own and dismissed all three claims accordingly. The breach of charitable trust and

keyboard_arrow_left Previous keyboard_arrow_right Next

13 14 32 17K 5

@muskonomy 😡😡😡

NEWS: SpaceX holders approved a 5-for-1 stock split ahead of the company's expected IPO, bringing the share price from $527 down to ~$105. SpaceX is expected to begin trading on Nasdaq under $SPCX on June 12.

keyboard_arrow_left Previous keyboard_arrow_right Next

🚨BREAKING: Samsung expands Texas chip project to $37 Billion with second fab The site will produce 2nm chips for AI and Tesla’s next gen AI5 and AI6 systems.

keyboard_arrow_left Previous keyboard_arrow_right Next

BREAKING: Elon Musk unveils Tesla–xAI project “Macrohard” aimed at automating software work March 11 – Elon Musk said on Wednesday that Tesla and his artificial intelligence company xAI are developing a joint project called “Macrohard” or “Digital Optimus,” a system designed to control computers and perform complex software tasks automatically. In a post on X, Musk said the project combines Tesla software with xAI’s Grok large language model. He described Grok as the high level “navigator” of the system, while a Tesla built AI agent processes real time computer activity such as screen video, keyboard input and mouse actions. According to Musk, the system analyzes the previous few seconds of computer screen activity and user inputs before deciding what actions to perform. He compared the setup to an advanced navigation system that directs software operations. “Grok is the master conductor or navigator with deep understanding of the world to direct digital Optimus,” Musk wrote, adding that the technology functions like a much more advanced version of turn by turn navigation software. Musk said the system combines two different forms of artificial intelligence reasoning. “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),” he wrote. The project will run primarily on Tesla’s internal AI4 chip, which Musk said costs about $650, while using more expensive server infrastructure powered by Nvidia hardware through xAI’s computing systems. Musk said this approach would allow the system to operate at a lower cost while still handling advanced AI workloads. “This will run very competitively on the super low cost Tesla AI4 paired with relatively frugal use of the much more expensive xAI Nvidia hardware,” Musk wrote, adding that the system could become a fully real time intelligent AI platform. Musk suggested the technology could eventually replicate many functions typically performed by software companies. “In principle, it is capable of emulating the function of entire companies,” he said, explaining that the name Macrohard is a humorous reference to Microsoft. The announcement comes as Musk increasingly connects his companies working on artificial intelligence, robotics and computing infrastructure. Tesla earlier agreed to invest about $2 billion in xAI as part of the startup’s funding round, while SpaceX recently merged with xAI in a deal aimed at combining AI development with satellite networks and space infrastructure. Industry researchers say systems capable of autonomously controlling computers are becoming a major focus in artificial intelligence, with several companies racing to build software agents that can perform tasks such as navigating applications, writing code and managing digital workflows without human input. Source: Elon Musk post on X

Elon Musk @elonmusk

5 months ago

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

8K 10K 77K 47.9M 10K

NEWS: Tesla moves more Cybercabs off site as production start approaches Tesla appears to be ramping up preparations for Cybercab production as more vehicles are being transported away from its facilities ahead of the expected launch. At least 16 additional Cybercabs were seen lined up for transport off site, according to Tesla observer Joe Tegtmeyer. The vehicles appear to be part of Tesla’s growing fleet of testing units that have been spotted across the United States. Joe Tegtmeyer wrote on X: “At least 16 (inventory was changing through the morning) more Cybercabs lined up for transport off site and these are more of the testing versions with the wrap that we have been seeing all over the country, including in Washington DC yesterday.” Images shared alongside the post show rows of Cybercabs staged in a parking area before being loaded for transport. According to Tegtmeyer, the number of Cybercabs already moved off site may be significantly larger than what was seen in the latest shipment. “Hard to know the exact number of Cybercabs that have been transported offsite, but easily well over 100 so far. Check out these fresh views … less than 3 weeks to the stated official start of production!” The vehicles appear to be early test units wrapped in camouflage, which Tesla commonly uses while vehicles are undergoing testing or validation. Tesla’s Cybercab is designed as a purpose built autonomous vehicle intended for the company’s future robotaxi network. The vehicle does not include a traditional steering wheel or pedals and is expected to rely entirely on Tesla’s Full Self Driving technology. Moving large numbers of test vehicles off site suggests Tesla is expanding real world testing across different locations before the start of production. Testing fleets are often distributed across cities and environments so engineers can gather data and refine software performance before vehicles are released to customers. Tesla previously stated that Cybercab production is expected to begin soon, and Tegtmeyer’s observation suggests the timeline may be approaching quickly. If production starts as planned, Cybercab would represent Tesla’s first vehicle designed specifically for autonomous ride hailing rather than personal ownership. That vehicle is expected to play a central role in Tesla’s long term robotaxi strategy. Source: Joe Tegtmeyer on X

Joe Tegtmeyer 🚀 🤠🛸😎 @JoeTegtmeyer

5 months ago

At least 16 (inventory was changing through the morning) more Cybercabs lined up for transport off site and these are more of the testing versions with the wrap that we have been seeing all over the country, including in Washington DC yesterday. Hard to know the exact number of

keyboard_arrow_left Previous keyboard_arrow_right Next

91 224 2K 392K 80

@elonmusk When we will see first demo?

🚨 BREAKING: FCC accepts SpaceX filing for ‘Orbital Data Centers’ The Federal Communications Commission Space Bureau has formally accepted for filing a new application from SpaceX to launch and operate what the company calls an Orbital Data Center System. This is not approval yet. This is the FCC saying: the filing is real, valid and now open for public comment. Public comments are open until March 6, 2026. What SpaceX is proposing: SpaceX is asking permission to deploy: • Up to 1,000,000 satellites • Operating between 500 km and 2,000 km altitude • In orbital shells up to 50 km thick • Using high-bandwidth optical laser links between satellites • Linking into the existing Starlink laser mesh • Downlinking to authorized Earth stations using 18.3–19.3 GHz and 28.6–29.1 GHz bands These satellites are described as: Solar-powered orbital computing nodes Designed to function as data centers in space The exact framing from the filing: In the application narrative cited by the FCC notice, SpaceX states this system is: “a first step towards becoming a Kardashev II-level civilization one that can harness the Sun’s full power” That language is directly referenced in the FCC public notice. This is why you are seeing the Kardashev quote repeated everywhere. Why this is radically different from Starlink Starlink = communications network. This system = compute network in orbit. SpaceX is proposing to: • Use solar power in space instead of Earth grids • Use vacuum cooling instead of water and air cooling • Avoid building massive data centers on Earth • Route traffic through the petabit Starlink laser mesh • Turn orbit into an AI compute layer They even quantify it in the filing: Launching 1 million tonnes per year of satellites Each tonne generating 100 kW of compute = 100 gigawatts of AI compute added per year For comparison, 100 GW is similar to the power draw of entire countries. What the FCC notice actually means: The FCC document (DA 26-113) says: • The application is accepted • Waivers are requested for several NGSO rules • The system will rely primarily on optical intersatellite links • The Bureau is seeking comments on the proposal and the waivers This is the first official regulatory step for this system. Not a rumor. Not a tweet. A federal filing. Why this matters for AI: SpaceX is directly arguing: • Earth’s grid cannot scale fast enough for AI demand • Orbital solar data centers are cheaper, cleaner, and faster • Starship makes this economically possible •This becomes the backbone for AI for billions of people This connects directly to: • Starship • Starlink lasers • xAI compute needs • SpaceX IPO rumors • The “data centers in space” narrative Musk has been posting about for months Timeline • Jan 30, 2026: SpaceX files the application • Feb 4, 2026: FCC issues Public Notice accepting it • March 6, 2026: Public comment deadline After that, the regulatory review begins. The big picture SpaceX is attempting to turn: Orbit → AI infrastructure Using: Starship lift capacity + Starlink laser mesh + solar power This is the first time any company has formally asked a regulator to approve data centers in space at planetary scale. Source: FCC Public Notice DA 26-113

@muskonomy I don't know why but they are registering this good tech for a long time 🕰️ It's time to integrate starlink with apple

@muskonomy @elonmusk 🔥🔥🔥

@muskonomy Wow incredible, slowly spacex is becoming one of the greatest company in the world 🔥

@muskonomy Big stuff