Some thoughts that have been living rent-free recently in my head about AI... We should be more precise when we talk about it. We’ve had "AI" for decades. Google Translate, Alexa and Siri, computer vision in video games, OCR in mail sorting, protein folding models all fall under artificial intelligence, but there’s no mass panic over these. The current progress is in generative AI, and that's where most of the public concern is. I’m guilty of the shorthand myself, saying "AI" when I mean "generative AI". People are afraid of generative AI, but the real problem is concentrated power and corporate greed. Someone wanting to sow disinformation could do so without generative AI models, and deep-pocket organizations of state could finance extreme actors if they wanted to. The tools evolved, but the risks aren’t new. The bigger problem that few are talking about is corporations who are aligning the technology for their own goals; they increasingly dicta...
We’re told that it will take powerful quantum computers to break RSA encryption, so for now the world is safe. But I wondered, in an era of increasingly sophisticated models, might AI pose a threat? These systems excel at finding patterns in data that humans miss, and if there were any subtle weaknesses in key generation, I would think AI could detect them. Yes, theory says it’s all but impossible to break RSA because it relies on the computational hardness of factoring large prime numbers. But theory and practice don’t always align, and sometimes the most interesting discoveries come from testing our assumptions. So I set up an experiment to test whether a transformer model could learn to reverse-engineer SSH private keys from their corresponding public keys. Experiment and Results I trained a T5-small transformer model (60 million parameters) on a dataset of 50,000 SSH key pairs, split into 70% training, 15% validation, and 15% test. Given a public key as input, the model was as...