index

Why I'm publishing my AI workflows for free

I just started ai.terrydjony.com, a site where I write up the AI workflows I actually use. Content, SEO, site audits, agents running on my own server. In English and Bahasa Indonesia.

All of it is free. Some people asked me why I don’t sell it as a course, so here’s the honest answer.

I learned everything for free

I’m a self-taught programmer. I grew up in a small city in Indonesia with a PC and an internet connection, and I learned by tweaking HTML on W3Schools.

Nobody charged me for that. Everything I know came from people who just put stuff on the internet for free.

So this feels like paying it back.

The gap is getting wider

A post by Andrej Karpathy is what pushed me to actually start.

His napkin math: around 6 billion people have barely touched an LLM. Maybe 1 to 2 billion use a free chatbot now and then, like a better Google. And a tiny slice, mostly people in code and math, are watching agents finish in a day what used to take them weeks.

And that gap is widening, not closing.

Andrej Karpathy's post on X describing the funnel of LLM understanding: about 6 billion people have barely touched LLMs, 1 to 2 billion are casual free-tier users, around 20 million professionals in math and code see agents finish weeks of work, and about 5,000 people have internal access to frontier systems
Andrej Karpathy's post on X describing the funnel of LLM understanding: about 6 billion people have barely touched LLMs, 1 to 2 billion are casual free-tier users, around 20 million professionals in math and code see agents finish weeks of work, and about 5,000 people have internal access to frontier systems

I see this gap everywhere in Indonesia, and honestly it’s bigger here.

I don’t want people here to only meet AI as the thing that took their job. I want them to be the ones using it.

Prompt engineering isn’t really a thing anymore

This is the other reason, and maybe a slightly hot take.

I don’t think you need to learn “prompt engineering” anymore. Not for normal, day-to-day work.

Maybe it still matters if you’re building an AI-first product in production, where every token and edge case counts. Fine.

But for the rest of us? Today’s models are good enough that you can just ask. Even if you know nothing about the topic, you can type your question like you’d ask a friend, and you’ll get a helpful answer. If it’s not what you wanted, you just say so.

There’s no secret prompt template. No magic words.

So paying for a “prompt engineering course” in 2026 feels like paying for a course on how to Google.

Spend the money on the tools instead

Here’s what I’d tell a friend.

If you have money to spend on AI, don’t buy a tutorial. Subscribe to the AI itself. A paid plan of Claude or ChatGPT, or whatever tool fits your work.

The gap between the free tier and a frontier model doing real work for you is huge. That’s where your money actually changes something.

This matters even more in Indonesia right now. Money is tight for a lot of people in this economy. Every rupiah should go to the thing that actually makes you more productive, and that’s the tool, not someone’s PDF about the tool.

What I’m doing

So I just document what I do, step by step, and put it out for free.

Most people I know aren’t lazy about AI, they’re skeptical. They tried a free chatbot once, it was underwhelming, and they moved on. But once you see what a frontier model can do inside your real work, it’s hard to go back.

That’s what I’m trying to show, one tutorial at a time.

If you want to follow along, it’s all at ai.terrydjony.com. And if you build something from it, tag me, I’d love to see it.