Physical AI will make subversive discoveries that exceed everyone's expectations - space-time integrated computing, rather than the current three-dimensional spatial computing plus discrete time steps
Let me explain again why I said "disruptive" rather than "substantial"：the current "embodied artificial intelligence" still uses 19th-century numerical methods (the Adams-Bashforth integration method from 1883 and the Runge method from 1895) to represent time frames + three-dimensional space calculations to approximate four-dimensional spacetime (relativistic covariance has proven that spacetime is an integrated whole, i.e., four-dimensional spacetime). I will release more specific code later - you might wonder, don't the "scientists" at those big companies know about this? The answer is that they do know, and I will also release the reasons later, which will definitely surprise you!
While 'substantive' would mean major progress within the current framework, I’m predicting a shift that subverts the current foundational assumptions of robotics.
Right now, we treat time as a secondary sequence—an 'add-on' to 3D space. Moving to a unified spacetime architecture isn't just a big improvement; it fundamentally undermines the discrete-frame logic that almost all current CV and RL models are built upon. It’s 'subversive' because it requires us to unlearn the way we’ve been processing motion for the last decade.
Allegedly AI recently discovered vulnerabilities in React Server that were/are being exploited on unpatched systems, so that's subversive, and we might expect a lot more of it before it gets better.
The problem lies here: the current "embodied artificial intelligence" still uses 19th-century numerical methods (the Adams-Bashforth integration method from 1883 and the Runge method from 1895) to represent time frames + three-dimensional space calculations to approximate four-dimensional spacetime (relativistic covariance has proven that spacetime is an integrated whole, i.e., four-dimensional spacetime). I will release more specific code later - you might wonder, don't the "scientists" at those big companies know about this? The answer is that they do know, and I will also release the reasons later, which will definitely surprise you!
