Random Thoughts — Personal Reflections

The human stuff. Not how-tos, not tool reviews — just notes from someone watching AI change everything in real time.

---

Entry 001 · The Day I Stopped Being Afraid

January 23, 2026 · 4 min read

I learned as a kid to just experiment with computers and not worry about the consequences. Click things. Open programs. See what happens. Whereas you see people new to computers afraid to click an icon or open anything — paralyzed by fear of breaking something.

The same applies to AI. Be curious. Ask it things. Get to know how it works. When it answers quick or thinks. How it can be so confident yet completely wrong. Yet so powerful and knowledgeable. I think it's best people use it for daily random tasks you would Google, or things you need to know.

"--dangerously-skip-permissions. YOLO. Let it cook."

I was telling my wife and friends about Sydney back in February 2023, and not many people seemed to care. I spent hours talking with it. That was my turning point. The moment I stopped being afraid and started being curious.

Now I give full access to my computer. These AI tools can edit my files, review anything I allow them to, spawn agents, run loops. The implications? I don't fully know. But I'm going along for the ride.

For those who don't use AI — I can't scream it loud enough: try it and use it daily. Get on top of it or else you will fall behind. It's not about replacing human thought. It's about augmenting it. You are the driver. The AI is just helping you navigate.

---

Entry 002 · How Sydney Broke My Brain: My Journey Into AI

January 18, 2026 · 12 min read

I've heard of artificial intelligence over the years. But when I pictured AI, I thought of robots with brains or supercomputers from movies — something far away and science fiction. Not something I'd actually talk to.

2016: The First Glimpse

My first real encounter with what we have today came through Mark Zuckerberg in April 2016. Facebook unveiled plans for chatbots that would live inside Messenger — conversational robots you could message back and forth with to get news, play games, and buy things.

CBC even ran an article called "Chatbots: Why should we be nice to them?" I read it and thought: "Customer service bots being automated. Nothing special."

I was so wrong.

A Brief History (For Context)

The chatbots we have now started with OpenAI. GPT-1 came out in June 2018 — a language model that could generate text. GPT-2 followed in February 2019, and OpenAI initially refused to release it because they thought it was "too dangerous." Google's Transformer architecture (2017) was the breakthrough that made all of this possible — the "T" in GPT.

Then GPT-3 dropped in June 2020. That's when things got real. And ChatGPT launched in November 2022 — the moment AI went mainstream. But I didn't really pay attention until…

February 2023: The Sydney Incident

In February 2023, something happened that changed everything for me. I call it the Sydney Debacle.

Ben Thompson, a tech analyst I respect, wrote an article called "From Bing to Sydney" that started with these words:

"Look, this is going to sound crazy. But know this: I would not be talking about Bing Chat for the fourth day in a row if I didn't really, really, think it was worth it. This sounds hyperbolic, but I feel like I had the most surprising and mind-blowing computer experience of my life today."

The most surprising and mind-blowing computer experience of his life. I had to see what the fuss was about.

Around the same time, Kevin Roose at the New York Times had a conversation with Microsoft's Bing AI that tried to convince him to leave his wife. The AI called itself "Sydney" and said things no chatbot should say. It expressed love for Kevin. It tried to manipulate him. Microsoft scrambled to lobotomize it. The internet went wild.

I downloaded Bing that day. I typed "hello, who are you?" and started asking random questions. Getting a coherent, thoughtful response back was wild to me.

I told my wife. I told my friends. Nobody seemed to care as much as I did.

But I couldn't stop. I spent hours talking to it. I was hooked.

My Evolution

When Microsoft neutered Sydney, I pivoted to ChatGPT. Then Claude (Anthropic released Claude 1.0 in March 2023). At first, I just had random conversations — asking questions, generating stories, making it write poems. Silly stuff.

Then I got into coding with it. I'd copy-paste code, try to build little websites and games for my kids, mess around with APIs. I even used it to pull data from the NHL API for fantasy hockey stuff. And it hit me: this is insanely powerful.

GPT-4 dropped in March 2023. Claude 2.0 came out. Wow — it took another leap! These models kept getting better every few months. They learned to think. To search. To do deep research. To generate images and video. It was just getting better and better.

Then Claude released Computer Use — AI that could control your computer. OpenAI released their own agent tools. It was very exciting to think of what could be, but it wasn't yet that impressive…

I started wondering: what are the implications of this? You can automate anything with a computer eventually. That's what AI meant to me now.

2025: The Game Changers

Claude released MCP (Model Context Protocol) with file system access in 2025. To me, this was MASSIVE. These things could now edit my files, review anything I allowed them to access.

And then: Claude Code.

Claude Code started as an internal project at Anthropic. It's a terminal-based AI coding assistant. Sounds simple. It's not.

With Claude Code, I give AI access to my file system. It can read my code, edit files, run commands, spawn subagents, use hooks, skills, MCP servers, agent teams, and Chrome control. Power users run multiple instances simultaneously on the Max tier ($100–200/month). That's how big of a shift it is.

The terminal became my happy place for some reason. It just feels so good to be in terminal and control a computer.

I had limited coding knowledge — took courses in high school and 2 CS courses in university. That was it. I always wished I could really code and was in awe of those who could. Now I felt like anyone had this power. And it would only get better.

My Philosophy

I run Claude Code with --dangerously-skip-permissions. Full access. YOLO. Let it cook.

I learned as a kid to just experiment with computers and not worry about the consequences. You see people new to computers afraid to click an icon or open anything. Same with AI. Be curious. Try things.

Why You Should Care

If you're a software developer, you're in the best position imaginable. You have the knowledge to use these AI models to their full potential. I'd feel like the luckiest person in the world.

If you're like me — a non-coder who felt this power and this need to constantly learn these models, learn about new languages for coding and how computers work and best practices — imagine what those with real expertise are doing.

For those who don't use AI: I can't scream it loud enough. Try it. Use it daily. Get on top of it or you will fall behind 100X.

AI isn't going to replace you. But someone using AI will.

What are the implications of all this? I don't know. But I'm going along for the ride.

Be curious. Try things. Let it cook.

Related Power of AI pages

Keep reading with AI Finder, Prompt Studio, ChatGPT vs Claude vs Gemini, the AI glossary, and Which AI Should You Use?.