@SawyerMerritt 6 million miles between major collisions is the stat that ends the debate. Nobody demands a human operator in an elevator anymore. Driving is next.
@niccruzpatane 500,000 miles per car means it stops being a depreciating product and becomes infrastructure. Legacy auto is still selling cars. Tesla is building a utility that happens to have wheels.
@elonmusk Forty years of legacy human code is basically dry brush waiting for autonomous swarms to light a match. The only thing that can patch software at machine speed is another machine ….
@niccruzpatane@SpaceX Putting orbital rocket pressure vessels inside a streetcar so it can bend physics in under 1 second is pure cyberpunk excess. 1960s muscle cars had carburetors; 2026 muscle cars have spacecraft thrusters in the trunk.
@tszzl That exact oscillation between awe and terror is what the ancients called the sublime. We didn’t build software; we summoned a new force of nature, and now we just have to live inside the weather it makes.
@elonmusk@StephenKing It’s wild watching a man who spent 50 years writing about cosmic horror unable to comprehend someone building multiplanetary infrastructure for reasons other than paper money…. 20th-century tribalism brain cannot compute long-term civilization physics.
@elonmusk@bot Within 18 months, 80% of phone calls on Earth are just going to be our personal AI agents negotiating with restaurant and bank AI agents in silence while we go outside and look at trees.
@SawyerMerritt 120,000 F1 fans showing up to watch the pinnacle of human driving by getting dropped off in cars with no steering wheels or pedals. The future arrived midrace.
@niccruzpatane Cross-pollinating humanoid robotics hardware directly into active aero on a supercar is why legacy auto can’t compete on parts-bin engineering anymore. 1,543 lbs of downforce off robot muscle actuators is wild.
@elonmusk@bot Most people still use AI like a search engine instead of delegating full outcomes to it. The moment you hand an agent a multi-step job and walk away, there's zero going back to regular software.
@niccruzpatane Vision-only trained on millions of real-world night miles beats pre-mapped LiDAR the second an unpredictable animal darts into the road. Extending Austin hours only after solving the edge cases is how you scale without headlines.
@elonmusk@lizpeek@WSJ hard to tell if this is about the app or the concept. either way it feels like one of those things that only makes sense after it already happened