The Magic of Generative AI and LLMs



When I started using computers, almost everything about them was magical. The curtains were drawn back as the years passed, and I slowly learned how these things worked. This is not necessarily a bad thing, though. The awe and fascination simply turned into appreciation for how well-designed and crafted these systems were.

The next time I was in awe of these systems was when I understood how a buffer overflow was exploited. I was in awe that systems, so well thought out and so well put together, can be broken, that too with such grace!

Cut to me now pursuing a Ph.D. in systems security. There is an appreciation for the art of exploitation, but systems have grown complex, and the problems I’m working on seem a bit far away from what initially drew me into this field.

Coming to generative AI and LLMs. I was always a skeptic of Machine Learning. From what little I understood, I thought it was just fancy math. Although neural nets always seemed fascinating, I never got a chance to study or work with them properly. When I first saw DALL-E, instantly, the magic returned! Many were skeptical and cynical that the AI was mimicking, that it would put artists out of jobs, questions about OpenAI’s intentions, and that there was nothing of substance here. Although some of these arguments may hold water. I was, simply put, impressed. Then came the barrage, Midjourney, Stable Diffusion, and people creations with them; I saw substance.

To top it off, came ChatGPT. I am no longer a skeptic of ML and AI. We know there is a lot here, as we saw it’s creations and applications quickly break into mainstream culture. Many famous voices in the field of security have expressed concern over how LLMs can be misused. Spam, phishing, and astroturfing are some to name a few. Many are skeptical still, they think of LLMs as just a fancy search engine. Academics are concerned about how this allows students to cheat easily. All while people are concerned, I’m reading and laughing my ass off over a rap battle between Liz Truss and a head of lettuce written by ChatGPT.


I’m more motivated than ever to learn more about AI and its applications, especially in security, and how we can use AI to battle the misuses mentioned above.

While the concerns are valid, I see disruption, much-needed disruption, and the cynicism melts.

At times, the joy that life attacks me with is unbearable and leads to gasping hysterical laughter. I find myself completely out of control and wonder how could life could surprise me again and again and again, so completely. How could a man be a cynic? It is a sin.

Norm Macdonald