John Miller's Visit & AI's Role in Identity

Recently, our class had the privilege of hearing from John Miller, VP of Project Management at Insight Software and co-host of the podcast Gettin' to Market. With a background spanning supercomputing, enterprise work at Apple (where he was instrumental in convincing skeptics that Macs belonged in the workplace), and years of advocating for emerging technologies, John brought a practitioner's honest perspective on how AI can reshape documentation workflows. His talk, "Documentation and AI: A Practitioner's Honest Take," focused on the practical applications of AI as a tool for amplifying human work, not replacing it. Between technical insights, John casually mentioned his collection of instruments—a banjo (the most expensive one), drums, and bass—revealing the kind of person who brings passion and depth to everything he does.

Alongside John Miller’s talk, I also read Dave Gauer's A Programmer's Loss of Social Identity and Bob Nystrom's The Value of Things. I expiremented with LLM Arena as well to determine how certain LLM models performed.

What I Took Away From John Miller's Visit

Where AI Can Help With Documentation

John outlined several concrete ways AI is already transforming documentation practices. Rather than replacing documentarians, these tools accelerate the parts of the process that are typically slowest:

Core Philosophies & Memorable Quotes

My favorite quotes from his presentation:

This was John's central philosophy. AI tools are helpers in the documentation process, not replacements for human judgment, expertise, or writing craft. This distinction matters because it shapes how we should think about integrating these tools into our workflows. We're not automating away the human; we're automating away the tedious parts so humans can focus on higher-value work.

This became the second central theme. The realistic expectation isn't that AI will produce perfect documentation on the first try. Instead, think of AI as a capable research assistant or junior writer who can get 80% of the way there. Your job as a professional is to refine, verify, and ensure accuracy. In a busy organization, that's a massive time savings.

The Reality Check: We're Not There Yet

One thing John didn't sugarcoat: despite all the hype, he hasn't yet seen AI agents autonomously solving actual customer problems in production environments. That honest assessment was refreshing. It reminds us that while AI is genuinely useful for specific workflows, we're still in the early days. There's no substitute for deep expertise, customer understanding, and careful thought.

LLMs & Identity

Testing Model Quality Using Arena

As part of a recent assignment, my professor introduced us to LLM Arena. We were tasked with using it to keep our assessments of AI models objective and up-to-date. This tool allows me to issue the exact same query to different models side-by-side and evaluate which response is superior.

This was an interesting challenge, as I gave the models several different types of prompts. The most interesting one was when I asked them to generate a customized workout routine based on my goals, time frame, and physical stats. The results were quite different. I ended up with GPT-5.2 and Ember as the two responding models. GPT was highly detailed and explained the reasoning behind each selected exercise. Ember, on the other hand, provided a very minimalistic and simple response. Surprisingly, both models also gave me a diet plan to follow, explaining how a calorie deficit works, which foods I should avoid, and how to plan my meals.

Identity and Your Future Self

How did you see your future self four years ago?

Four years ago, I was in my second semester of college, and the future seemed surprisingly bleak and hopeful at the same time. I struggled heavily with imposter syndrome during my first semester, especially in my computer science classes. I was also dealing with the personal battles that the majority of first-year students face, like feeling homesick, which led me to consider transferring to a college closer to my family. Deep down, I think I knew I wanted to stick with CS, but I wasn't sure if I was truly cut out for it, which is where the imposter syndrome stemmed from.

How do you see your future self now?

Although the current age of AI and the job market can be a bit intimidating, I know myself well enough to trust that I will persevere and adapt. I see my future self in a position where she is genuinely happy, one that satisfies her desire to be challenged and continuously pushed out of her comfort zone.

How has AI changed how you invest your time and energy?

AI has vastly changed how I learn, study, and research. My workflow for all three tasks has shifted drastically. I now spend far less time on the initial steps, like gathering materials or setting up a step-by-step plan. Organizing tasks that used to take me at least 30 minutes has been reduced to 10 minutes maximum. Because of this, my energy is entirely focused on the main task at hand.

How will you live your life so that AI can't make you feel obsolete?

I'm not completely sure how to answer this question, mostly because I haven't genuinely felt that fear before. By focusing my career on the machine learning side of things, understanding exactly how LLMs and generative AI models work gives me a strong sense of security. I also make a conscious effort to keep up with new technologies and developments in the field. Maybe it sounds arrogant, but I don't feel threatened that AI will eliminate all jobs or render my degree useless.

Nogramming Progress This Week

I have came up with the list of people to interview, and have started contacting them to set times to do the interviews. I plan on finishing that up this week, and start planning questions for the interviews.

~Shree