This is one of the saddest things I’ve read. My god they deep captured IEEE. This article is bullshit and I don’t think I’m going to renew my membership.
Engineers often prefer ICs to have interpretable structures, so that if a problem crops up, they can understand how the chip works well enough to debug it.
lol good luck with that
Imagine you go to your favorite image-generation engine and ask it to create a painting of the sky in the style of Picasso, Van Gogh, or Michelangelo. You will get images that capture the essence of their brushstrokes, their use of colors, and their framing. All are pictures of the sky nonetheless, but in different styles.
IEEE drank the koolaid
Yep, I dropped a few years ago.
Oh holy shit, this is such a bad idea. The skill curve on understanding RFIC layouts is so crazy that maybe only the top 1% of electrical engineers should be allowed to get anywhere near it.
I can understand using a convolutional neural network in conjunction with human designers as they suggest at one point, but putting anything with a Transformer architecture near this stuff is going to be genuinely apocalyptic.
This sounds like it’ll be an expensive mistake
Whoever wrote this is a moron. Not only is RF design not a dark art where wizards conjure up designs out of pure intuition, but machine-learning methods have been in use for years & years by engineers. Nothing new. https://www.ansys.com/simulation-topics/what-is-inverse-design
RF design is no dark art, but it is standing on the shoulders of giants. Machine learning has definitely been helpful for iterating and integrating, but using something like GPT Image to generate your circuit layouts is absolutely a misuse of the software.
Jim Williams died for this.
