# Generalist Robotics **Entity class:** Robotics architecture **Domain:** Artificial intelligence / science / political economy **Maturity:** Developed ## Definition **Generalist Robotics** uses broadly trained models to perceive, reason, and act across physical tasks rather than building a separate intelligence stack for every machine and environment. ## Mechanism and Consequence The approach transfers cross-domain competence into embodiment and can later distill narrower controllers, but its economics still depend on latency, reliability, compute efficiency, hardware, and manufacturing scale. ## Relationships - [[wiki/Robotics|Robotics]] - [[wiki/Machine Succession|Machine Succession]] - [[wiki/Humanoid Robotics|Humanoid Robotics]] - [[wiki/Compute Scarcity|Compute Scarcity]] ## Source Dossiers - [[research/AI Is Now So Good They Will Pay You to Quit|AI Is Now So Good They Will Pay You to Quit]] - [[research/Super Intelligence Week Index of Statements from the White House Luncheon|Super Intelligence Week Index of Statements from the White House Luncheon]] - [[research/Jensen Pushes Back on Doomers Moonshots Live|Jensen Pushes Back on Doomers Moonshots Live]] ## Simple Reminders, Quotations, and Thoughts > DrivingBench placed a software interface in a Toyota Corolla so an AI could steer, accelerate, and brake, then asked a general-purpose model to navigate a 130-meter cone course with one instruction: drive from here to there. GPT-6 Astra completed the entire course on its second try. > **— Adapted from Peter Diamandis**, *MOONSHOTS Live, September 25, 2026* [[reminders/Robotics/A General AI Drove a Physical Car on Its Second Attempt by Peter Diamandis|A General AI Drove a Physical Car on Its Second Attempt by Peter Diamandis]] > "Anthropic is playing the vector-graphics game, where everything is a command, while OpenAI is playing the raster or pixel-basis game. These strategies are on a collision course in robotics because embodied AI in the physical world will reveal whether the vector basis or the pixel-raster basis is the right one." > **— Alexander Wissner-Gross**, *MOONSHOTS Live, September 25, 2026* [[reminders/Robotics/Embodied Robotics Will Decide Between Vector and Pixel AI by Alexander Wissner-Gross|Embodied Robotics Will Decide Between Vector and Pixel AI by Alexander Wissner-Gross]] > "The bitter lesson is bitter indeed. The notion that robotics requires separate models is either already out the window or imminently will be. You will vibe-code a robot: drop an open frontier model such as Astra into a robotic body, and it will one-shot embodied cognition." > **— Alexander Wissner-Gross**, *MOONSHOTS Live, September 25, 2026* [[reminders/Robotics/Frontier Models Will Let Us Vibe-Code Robots by Alexander Wissner-Gross|Frontier Models Will Let Us Vibe-Code Robots by Alexander Wissner-Gross]] > DrivingBench demonstrates generalized intelligence: a high level of competence learned across domains will reverberate into biology, physics, and almost anything else. I do not think there is a single specialized model that will still outperform a general model a couple of years from now. > **— Adapted from Emad Mostaque**, *MOONSHOTS Live, September 25, 2026* [[reminders/Machine Succession/General Intelligence Will Reverberate Across Robotics, Biology, and Physics by Emad Mostaque|General Intelligence Will Reverberate Across Robotics, Biology, and Physics by Emad Mostaque]] > "Both things can be true: a generalized model can drive a car or inhabit and build a robot, while a specialized model may still perform the task more efficiently. The missing question is how much compute the general model used. Compute will remain starved; chip prices have risen, and high-bandwidth memory has increased fivefold in price. AI is so valuable that people want the compute, so using more compute for the same task is a serious failure." > **— Dave Blundin**, *MOONSHOTS Live, September 25, 2026* [[reminders/AI Infrastructure/General Models Must Still Justify Their Compute Costs by Dave Blundin|General Models Must Still Justify Their Compute Costs by Dave Blundin]]