Some Random Thoughts on How LLMs Transformed My Leaning Process

I have some thoughts about how LLMs have transformed my life, but I can’t quite articulate them.

This Friday, I presented the paper “Language Models Use Trigonometry to Do Additions” at the NLP reading group at Michigan (click here for my notes). Now, I’m not familiar with LLM interpretability, but I found it fascinating how, from a vast and general corpus of text, an understanding of the linear and periodic nature of natural numbers emerges in these models.

So I uploaded the paper to ChatGPT, asking questions and verifying my understanding as I went along. It took me just a weekend to get through it – something that would have been far slower before widely accessible LLMs. It’s not that Google Scholar or Wikipedia didn’t exist before, but the conversational nature of ChatGPT fundamentally changed how I interact with knowledge. When I presented the paper, I demonstrated examples using the Cursor AI IDE, which I now use daily (and yes, it significantly speeds up my coding).

Later that day, I read an article “We Live Like Royalty and Don’t Know It” from my timeline, which discusses how modern systems that undergird our lives – food distribution networks, water supplies, electricity grids, and public health systems — provide a level of comfort and security that past generations couldn’t have imagined. When I encounter unfamiliar words in that article (English isn’t my native language), I used an LLM-based etymology tool (that I also found on my X timeline) to analyze them.

It made me reflect on my own experience with using LLMs to optimize daily tasks. It feels like LLMs don’t just boost productivity; they augment my thinking. I really believe that LLMs will change human cognitive process: streamline learning, accelerate problem-solving, and reshape how we interact with information. What a time to be alive.