Today at 8 PM ET, @NormalComputing's @MaxAifer will present our latest work on thermodynamic computing on a live stream with the @InferenceActive! As AI development accelerates, semiconductors have become the crucial bottleneck. Normal Computing is disrupting the fast growing industry poised to reach $300B+ by 2029. Max Aifer will speak about how we develop physics-based computing chips for matrix inversion and the applications we are exploring in training large-scale AI models to transform AI hardware efficiency. Normal’s approach to de-risking novel AI chip architectures depends heavily on AI itself. Even for simpler kinds of physical architectures, like memory, complexity is now at "PhD level". At Normal Computing, we have trained the first AI to genuinely understand formal chip logic to help de-risk chips design, development, and verification. Have any questions about thermodynamic computing and probabilistic ML? Make sure to tune in: youtube.com/live/DAZNlhxOhSg

Linear algebraic primitives are at the core of many modern algorithms in engineering, science, and machine learning. The inefficiencies of digital hardware for AI are widely known – one @ChatGPTapp session requires 150 times more power than all-encompassing brain processing. Accelerating these primitives with novel computing hardware would have tremendous economic impact.  Normal Computing's @MaxAifer will be speaking about our work on thermodynamic computing, how we develop physics-based computing chips for matrix inversion, and the applications we are exploring in training large-scale AI models to transform AI hardware efficiency on the @InferenceActive YouTube channel live stream on November 11th at 8:00 PM Eastern Time! Hit the bell to get notified when the stream is live!  youtube.com/live/DAZNlhxOhSg