Getting Started with AI: Amplify What You Already Do Well

Today let's talk about how to actually start using AI — to lift your output, or to handle the daily grind.

In this AI era, AI is basically the best teacher you can ask for. Almost any question you put to it gets a solid answer.

So why is the market still flooded with courses making you feel like you'll fall behind if you don't learn the latest tool? Why does that just create more anxiety?

Here's the thing: in the AI era, your fundamental capabilities are what matter most. They're the foundation AI can amplify.

---

The key to using AI isn't the tool — it's what you already do well

Remember: no matter how powerful AI gets, it's still a tool. Without a clear purpose, no tool can take you anywhere.

This AI wave is different from anything before, though — AI has reached "almost" omnipotent territory. The ceiling now is your own imagination and capability.

A few well-known examples:

PJ Accetturo, AI filmmaker. Made a commercial for prediction-market platform Kalshi that aired during the NBA Finals. A commercial at that tier used to mean a full team, six-figure budget, weeks of work. His version: one person, two days, $2,000. 20 million impressions during the broadcast.

The one-person marketing agency — already a common operating model in 2025. A 5-to-10-person team used to deliver a handful of videos per client per month. One person with about $150/month in tools now ships 20 short videos in a single day.

Independent publishing — Rossum Press, run by Ryan Cook alone, using AI-assisted literary translation. The books he puts out are obscure-language works no publisher would touch before, because translation was too expensive and the market too small. Not faster. Going from impossible to possible.

None of these three became AI experts. PJ is still the director who knows how to shoot. Ryan is still the publisher with taste.

The point: AI tooling skill isn't what creates the productivity gap. Your own ability and way of thinking is.

---

So if I don't rush to learn AI tools, am I going to fall behind?

The learning curve for AI tools is low, and the pace of change is brutal. A year ago, courses were selling you magic prompts and magic workflow templates.

Now everyone has pivoted to Skills and agentic workflows. And one-shot generation tools keep multiplying — describe the product you want, and it builds it.

Models get updated. Tools get updated. The iteration speed only accelerates. There's no day when you've "learned it all."

So flip the question: what's "always-on" — what stays constant no matter how the outside world shifts?

I'd say: the deeper-layer skills — logical reasoning, breaking a complex problem into small ones, judging whether a claim is trustworthy, how fast you can grasp something new and put it to use.

These skills mattered before AI. They matter more now — because AI executes for you, but whether to do it and whether it's done right is still your call.

Not passively absorbing third-hand, fourth-hand summaries someone else packaged for you. In this era, curiosity and the act of exploration become invaluable — your own distinct capability.

---

So should I buy those AI courses on social? Should I pay for a subscription?

The AI course ads on social: "Zero to AI agent in 3 weeks," "n8n one-person company bootcamp," "vibe coding monetization class."

80% of what those courses teach is transcribed from YouTube and official documentation. The remaining 20% may not even fit what you actually need.

YouTube and LLMs can answer essentially any question.

One: YouTube

For English speakers, channels like AI Explained and Matt Wolfe are excellent for in-depth explanation, plus the official OpenAI and Anthropic channels for the latest capability demos.

Find first-hand sources. Use AI alongside them to deeply understand AI's nature and how it actually works. That investment pays off across everything else AI-related you'll ever learn.

Two: the LLM itself

This channel is the one most people skip. You don't need another person to teach you how to use AI. Just ask it.

Two prompts worth trying:

• "I come from a ___ background and have never used you before. In terms I can understand, teach me the three things I should know first." (If it's still confusing, add: "Explain it like I'm a 6th grader.")
• "For ___ task, search the web for how people are solving this with AI today. Give me the five most popular approaches, with source links."

Run with those for a week.

Now — courses aren't worth paying for, but the paid subscription is. With one caveat: don't buy annual. Monthly auto-renew is fine. AI changes too fast — don't lock yourself into one service. Spread the risk.

And the strongest, latest models all sit behind subscriptions. Claude Cowork, the buzzy one lately, is paid-only.

For Gemini Nano Banana, the current image-generation king, the paid tier's output quality is dramatically better than free.

$20 a month gets you the world's strongest generalist plus top-10% domain experts you can ask anything and have help you solve real problems. Genuinely a bargain.

---

So practically — how do I take the first step?

Enough mindset and resources. The most concrete answer: there are two things you can start today.

One: learn to ask a good question

In your back-and-forth with AI, "asking a good question" is actually the most critical skill.

Say crawfish ("xiao long xia") is trending and you want to know more. You'd just ask AI: "Tell me what crawfish is," right?

But try this instead: "Why did crawfish go viral? What problem did it solve? What problems remain unsolved? If I were running a similar marketing campaign, what alternatives are there besides crawfish?"

I'd bet that on any AI, the second prompt gets a totally different answer.

Make the question more explicit. Bring your purpose. Or ask AI to think it through with the 5W1H framework or "first principles" — the answer shifts dramatically.

Asking AI to cite sources also reduces hallucinations and makes the response more reliable.

Two: start from what drains you the most

Tools change. Versions change. Any "must-learn" list expires eventually.

Instead of chasing tools, think about which daily or work tasks drain you most.

Spending hours organising client business cards? Try snapping photos and letting AI extract them. Then update AI on conversation progress over time, and let it prep your next meeting too.

Filling out the same form daily? Paste the format to AI and let it fill them out. Email overwhelming you? Let AI organise it — even ship you a daily summary.

Take the first step. Let AI into your work and your life. If you have specific scenarios or tasks where you'd want a deeper AI playbook, let me know — those will shape what I write next.

---

Related reading

If you'd like the company-side view first, the previous post covers that: Three Questions to Answer Before Adopting AI.