AI is a fascinating area of technology that's been developing at an incredible pace - faster than many of us expected. While those working in tech have known about it for years, it wasn't really until ChatGPT came online that things truly took off. The topic elicits a lot of hate and contentious views, but amidst the hype and criticism, I believe there's something genuinely valuable in this technology.
I've been having a lot of fun with AI-assisted coding. It isn't pure "vibe coding," but rather thoughtful AI-assisted development. Learning syntax seems redundant when I can outsource that to a machine that's superior at this particular task.
The Critical Thinking Problem
I believe that those who hate AI often don't adequately understand the technology. A perfect example: in June 2025, a study was making the rounds that supposedly proves "AI = Everyone will get dumber." Almost all comments I saw amounted to "omg so true 100%." The irony is palpable - people complaining that "AI will make you dumber" while displaying zero critical thought when criticizing AI.
What such research and comments don't account for is that some people have to think. Some of us (call us "compulsive analyzers," "informational omnivores," or "autists" in the colloquial sense) experience discomfort when we don't mull ideas. For us, the temptation to off-load everything is weaker, and when we do off-load, we often reinvest the freed bandwidth in higher-order reflection.
While there's truth in the article, there's crucial nuance buried deep in the report on page 139 (out of 206):
"Brain-only writers who later added ChatGPT actually showed enhanced posterior–prefrontal coupling (Session 4)."
In other words: Learning first, automation second = cognitive dividend.
This insight has important implications for educational policy: teach kids to work without AI first to learn foundational skills, but don't completely remove AI tools. I haven't seen any educational professional advocate that we should only rely on AI, so concerns about future generations becoming cognitively dependent may be overblown.
AI isn't brain rot; it's a mirror. If you skip the heavy lifting, it happily keeps the bar low. Sweat through the first reps yourself, then let AI boost - not replace - your cognition.
My Guiding Principles
My philosophy on using AI can be broken down into a few key principles:
1. Augment, Rather than Replace: AI should be a partner that handles the tedious work (like syntax and boilerplate), freeing up human cognition for higher-level creative and strategic thinking.
2. Critical Engagement is Non-Negotiable: The value of AI is directly proportional to the critical thought you apply to its output. It's a tool for enhancement, not a replacement for thinking. If you skip the heavy lifting, it happily keeps the bar low.
3. Build in the Open: The best way to understand and demystify AI is to build with it and share the results. This is why my projects are open source.
My Workflow
I use AI as an active pair programmer. My primary tool is Cursor, which allows me to have a tight feedback loop with an AI that has full context of my codebase.
I Delegate: Repetitive tasks, writing boilerplate code, generating unit tests, and explaining unfamiliar libraries or codebases.
I Command: High-level architecture, creative problem-solving, and the final decision on whether to accept, reject, or modify any code suggestion. Creating documents that contain project requirements and design thinking are essential.
I Recognize its Limits: AI is not a strategic thinker. It can't tell you what to build, but it's an unparalleled accelerator for how to build it once you have a clear vision.
My AI-Powered Projects
To put my philosophy into practice, I've created several applications with AI assistance that also use AI to perform central tasks.
NovelWriter
NovelWriter is a comprehensive Python application designed to assist authors in writing novels and short stories across multiple genres by leveraging Large Language Models (LLMs). It provides a GUI-based interface built with Tkinter for managing novel parameters, generating universe lore, outlining story structure, planning scenes, and writing chapter prose. The code is fully open source on GitHub.
The original version was created as part of NaNoGenMo 2024 (completed). The first version generated a 52,000-word novel called "Echoes of Terra Nova."
LLM Creative Writing Tester
A tool for testing and analyzing creative writing capabilities of various Large Language Models. This tool allows you to test multiple LLMs with the same creative writing prompt to analyze and compare their responses.
Features advanced similarity analysis between responses:
- Text-based similarity (exact matches)
- Semantic similarity (meaning-based comparison)
- Named entity detection and comparison
- Name component analysis (detects when name parts like surnames appear across different texts)
- Text structure analysis (paragraphs, sentences, and word metrics)
Find it on GitHub.
DungeonGPT
DungeonGPT is an interactive, AI-powered Dungeon Master tool designed to guide players through customized adventures in tabletop role-playing games. The application features dynamic character creation, party selection, game settings customization, and conversational gameplay, all powered by GPT-4.
The code is fully open sourced on GitHub and written in Python (I also rewrote this app in JavaScript but haven't published that version yet).
What's Next?
My journey with AI is ongoing. I'm currently exploring the potential of LLMs for creative writing, gaming, price analysis, and more. My goal is to continue pushing the boundaries of what's possible when human creativity is augmented by machine intelligence.
The key insight I've gained is that AI's value isn't in replacing human intelligence, but in amplifying it. When used thoughtfully, AI becomes a powerful tool for turning ideas into reality faster than ever before. The challenge isn't technical - it's maintaining the critical thinking and creative vision that makes the output worthwhile.
As I continue to develop AI-powered applications and explore this rapidly evolving field, I'll be sharing insights, lessons learned, and practical approaches here on the blog. The future of human-AI collaboration is being written now, and I'm excited to be part of that story.