When I spoke at TechLead Conference 2026 powered by connpass, I kept hearing a similar theme at the networking session:

I don’t know what to learn in the AI era.

My own growth cannot keep up with AI’s progress.

How should beginners even build up their skills now?

My talk was titled “Tackling Technical Debt in the AI Era,” but at the networking session, some attendees brought up anxieties about learning and career as a separate topic of conversation. Stepping away from technical debt for a moment, this post is a column reflecting on that theme.

Let me lead with the conclusion. As someone who has been engineering for 15 years, I believe this is the best time ever to be a beginner.

2011: Asking a Senior Engineer Basic MySQL Questions

Rewind to 2011. I was a fledgling engineer, attending Python mini Hack-a-thon (affectionately known as “pyhack”) in Tokyo. It is a study group operated by members of BeProud, Inc., where Python enthusiasts gather to work quietly on their own projects.

At those gatherings, I peppered senior engineers with genuinely basic MySQL questions. I had read books and the official documentation, but there were gaps that only clicked after asking a human, face-to-face, “Am I reading this right?”

Looking back, the patience of those seniors was a formative experience. But honestly, did I feel I could “ask as much as I wanted”? Not really.

• I couldn’t ask face-to-face unless I showed up at a weekend event.
• Posting on a Q&A site meant being visible to strangers, so I was careful with wording.
• I hesitated, thinking “is this too basic to ask?”

There was a hard, physical ceiling on how many questions I could actually ask.

2026: Unlimited Access to AI

In 2026, tools like Claude, ChatGPT, and Gemini let me ask anything the moment I think of it.

• Late at night, early morning, or on the move
• Asking the same thing 10 times, with no dirty looks
• No need to preface with “sorry for the basic question”
• Paste in the full code, schema, or error message, and discuss it in context
• Adjust the explanation level: “explain it more simply” as many times as I want

The questions that 2011-me wanted to ask but never did are countless. If Claude Code and Codex had existed back then, I can only imagine how much faster I would have learned.

People often say “AI weakens your thinking.” I see that as a usage issue, not a tool issue. If you only copy the answer, yes. But if you push back and ask “why is that the answer?” and iterate on that loop, the AI era gives you a privilege earlier generations never had.

The Learning Loop Got Shorter

Here is what a learning loop looked like in 2011:

1. Read a book or the official docs
2. Write code, run it
3. Get stuck
4. Ask at a study group or around the office (costs a venue and a week)
5. Ask again next week, or next month

Here is the 2026 version:

1. Learn the concept from books, docs, or AI
2. Write code, run it
3. Get stuck
4. Ask AI right there (seconds)
5. Dig deeper, or try a different approach

One cycle went from a week to minutes. That means the total volume of learning explodes, and beginners benefit most, because beginners are the ones who get stuck most often.

On the Anxiety About Self-Growth

“My growth cannot keep up with AI’s progress” is a feeling I understand completely. Yesterday’s state of the art becomes obsolete after this morning’s model update. I feel it too, daily.

But comparing your growth to AI’s progress on the same axis is setting yourself up to lose. AI is a tool. You are a learner. Tool performance and learner growth should be measured on different axes.

Instead, try reframing with these questions:

• Compared to a year ago, how much wider is the range of topics I can engage with?
• Investigations that took a week a year ago: do they take a day now?
• Questions I could not even formulate a year ago: can I formulate them now?

As long as AI progress becomes leverage for your learning, growth compounds. Feeling left behind often just means the comparison is pointed the wrong way.

For Those Just Starting Out

Looking back 15 years to my pyhack days, that era had its own kind of warmth, carried by the generosity of the people who answered a beginner’s questions. Those memories stay with me.

And 2026, on its own terms, is a genuinely great time too, in the specific sense that anything you want to know, you can research without limit.

If 2011-me could stand in today, the first thing he would do is throw every stuck-MySQL question at AI. Then he would learn at 3x speed, fail at 3x speed, and hit the next wall at 3x speed. That sounds like a joyful life to me.

To those just starting out, or anyone anxious about learning in the AI era: please, ask AI without holding back. There is no such thing as “too many questions.” The environment that 15-years-ago me would have killed for is sitting right in your hands today.

That’s all from the Gemba, still wanting to keep learning 15 years on from being a fledgling engineer.