I am keenly interested, and I have been researching ways of making AI more energy efficient. Here is an interesting idea. The world and laws of physics are analog, converting sound, temperature, colors, for example back-and-forth to digital uses a lot of energy. What if we could skip that and take analog input to run ML diagnostic? I am not fully versed in everything that it represents, but I see the potential of this technology trend, particularly linked to IoT, automation, and operations.
We agree that the future of computing and AI will be analog. Check out this episode of Undecided by Matt Ferrell to learn why using analog is critical for sustainable AI solutions - reducing the enormous amount of energy and resources required to support all of today's AI applications. And learn more about Aspinity's analog near-zero power always-on AI solution at ~ the15:02 time stamp. https://lnkd.in/eZAtJmDj