I am sitting in the car trying to describe why AI feels like an opening. If you enjoy creating, solving problems and learning, here is another way to act on that curiosity.

In the video I compare it to money: it can amplify what you bring to it. That is a metaphor about habits, not a scientific rule about who will succeed. A tool also has limitations of its own, and a motivated person can still get a wrong answer from it.

Keep the invitation, correct the exaggeration

I say the field has no established experts, degrees or rules. Taken literally, that is wrong. Carnegie Mellon announced an undergraduate AI degree in May 2018.[1] The European Commission records that the EU AI Act entered into force on 1 August 2024.[2] Both predate this video.

The newer opportunity I am reaching for is access to conversational tools and the chance to build useful things with them. That is different from saying the underlying field began yesterday.

The “99%” figures in my speech are not survey findings. Nor can I establish that a window will close in one or two years, or promise that practice for that long makes somebody an expert. Those lines communicate urgency; they should not become statistics in an article.

Show what you can do

The more useful question is what I could make for somebody today. A clearer information page. A working prototype. A comparison checked against its sources. Something with an identifiable user and a result that can be examined.

Make curiosity produce evidence. Try one real task: Choose a result you know how to check. Find the weakness: Compare the output with a reliable reference. Improve and explain: Keep the revision and explain what changed.
A practice loop, not a promise of expertise or employment.

I would keep a small record for each attempt: the task, the first output, the mistake I found and the corrected version. That record tells a different story from a folder full of impressive screenshots. It shows whether I can recognise a failure and improve the work.

Formal study and practical experimentation can support one another. Neither a certificate nor confidence in a chat window removes the need to understand the problem. The depth required depends on what I am doing: experimenting with a personal page is different from deploying a system others depend on.

Start where your judgement is strongest

Choose work you already understand well enough to evaluate. Ask the tool for alternatives, inspect its reasoning where possible and check the actual result. If I cannot tell whether the answer is correct, that is the next thing I need to learn.

My service-business article takes this into a concrete project: faster output still needs checking, customers and realistic costs.

I do not need to prove that everyone else is late. I need to do something useful with the time I have now.

Sources and calculation notes