My Tryst with AI.

My journey with artificial intelligence began in 2002 when I first encountered Genetic Algorithms by Melanie Mitchell. The idea of evolving algorithms that could adapt and solve problems captivated me, and I knew then that AI was the field I wanted to explore. However, my path was far from smooth, marked by academic and professional struggles but driven by an enduring passion for intelligent systems.

From the outset, my non-CS background made breaking into AI difficult. My bachelor's thesis focused on machine vision, where I attempted to automate the quality inspection of screw thread profiles using a camera and rudimentary image processing algorithms. While it was an early foray into computational problem-solving, it didn’t give me the footing I needed to fully explore AI. When I transitioned into the IT and CS workplace, I found myself sidelined, with my desire to dive into AI research dismissed due to my academic background.

Frustrated by the lack of opportunities, I applied for M.Tech programs in AI but faced rejection. Instead, I pursued a master's in cognitive science, where I chose a Natural Language Processing (NLP)-based project for my thesis. This project felt like a bridge between my cognitive science studies and AI, but it wasn’t enough to fully satisfy my curiosity about AI's broader landscape. After graduating, I moved into data-centered roles, working in the field for about five years starting in 2016. While I gained practical experience in data science and machine learning applications, these roles were often centered around business needs rather than cutting-edge AI research.

After completing my master’s around 2010, I also delved into the world of Artificial General Intelligence (AGI) and transhumanist ideas. These concepts expanded my view of AI, shifting it from problem-solving algorithms to the potential for machines to possess human-like cognitive abilities or even transcend human limits. I became fascinated with computational models of cognition and dynamical models of cognition—both of which dovetail nicely with reinforcement learning models and frameworks. These subjects, rooted in how systems can model complex cognitive processes, felt like the right direction for deeper exploration, though they remained outside the industry focus at the time.

As AI exploded with advancements in deep learning, NLP, and other fields, I watched from the periphery with a mix of awe and trepidation. I felt the field advancing at such a pace that catching up seemed almost impossible. This trepidation, I now realize, stems in part from the fact that the areas that genuinely spark my curiosity—ethical AIs, causality, data validation, and model monitoring—are not always prioritized in industry. These topics don’t get the same attention or investment as headline-grabbing advancements in deep learning or generative models.

Furthermore, the high costs of experimenting with AI models have made it challenging to learn by "playing around." In AI, especially in modern fields like large language models (LLMs), practical experimentation is key, but the computational resources and data access needed can be prohibitively expensive. This barrier has often left me feeling disconnected, watching the field progress without the means to fully participate. Thankfully, at least in the realm of LLMs, the costs are beginning to trend downward. The more accessible these tools become, the more I feel that I might be able to reengage with AI in a meaningful way.

Despite all the obstacles, my passion for AI is still alive. Over the years, I've accumulated a range of experiences in cognitive science, NLP, computational and dynamical models of cognition, and data science, which give me a unique perspective on AI’s evolution. Even though my journey hasn’t taken the direct route I had initially envisioned, these experiences have given me insight into AI from different angles. The industry’s focus may not always align with the questions that fascinate me, but the landscape is evolving, and the democratization of AI tools could be my opportunity to dive back into areas like ethical AI and causality.

Perhaps my tryst with AI is not just about catching up, but about finding my place in this ever-expanding field. The journey continues, as does my curiosity and passion.

P.S: The highlight of this journey is that this article was generated with the help of Chat GPT 4o model, with me feeding it facts and some guiding themes.