I wrote recently about the acceleration being real. About sitting down to tackle challenges that used to need full cross-functional teams, and delivering them myself in a fraction of the time. That hasn't slowed down. Every few weeks there's another multiplier, and each one lands on a base that's already grown. The compounding is relentless.
This isn't a spectator sport anymore. If you're waiting for your company to figure out AI for you, or for things to settle down before you engage, you're making a decision by not making one. And it's getting expensive.
I've spent 25 years building operating models, leading transformations, managing teams locally, regionally, and globally. I know what serious professional capability looks like when it emerges, and I know what it looks like when people miss the window. This is the window.
The opportunity is right now. And it belongs to the people who decide to take it.
A Thousand Hours
So how long does it take? Someone asked me this recently. To really get AI. Not dabbling good. Actually good.
My answer starts with a thousand hours.
I know that sounds like a lot. You've got a job. Maybe a family. You're already stretched. I'm not asking you to quit your life. I'm asking you to carve out an hour at a time.
The capability I described in my last article didn't arrive overnight. It came from building things. Breaking things. Sitting with problems that felt impossible until I figured out how to make the models work for me. There's no shortcut. You can't read your way to AI fluency. You can't watch enough YouTube videos. You have to do the work. Think of it like learning a language. Nobody becomes fluent from an occasional lesson.
Make It Real
Taking this seriously means treating it like the professional development it is. Not something you dabble with when you remember. Something you schedule.
Maybe that's an hour before your workday starts. Maybe it's replacing one evening TV show with tackling a real problem using AI. The slot doesn't matter. What matters is that it's regular, it's protected, and you're bringing real challenges to it.
Get Set Up
You also need to get set up properly. Start with Anthropic's Claude. Something meaningful shifted with the launch of Claude Opus 4.5, and it's where I'd point anyone beginning their serious AI journey. Get a subscription. The free tier is fine for experimenting, but if you're putting in real hours, you need the full capability. If you're on a Mac, Claude's new Cowork feature lets you do things with local files and workflows that previously required technical skills most people don't have. This is a big deal. And if you're technical, jump straight into Claude Code. It's not just for coding. You can build, automate, and orchestrate things that would have taken a team a week.
Then broaden your exposure. Claude, GPT, Gemini, Grok, Llama. They're not interchangeable. Each has a different personality, different strengths, different ways of approaching problems. You need to understand these differences firsthand. Not from reading comparisons. From working with them. I'll write more about model capabilities and personalities in a future article.
You can subscribe to each individually, or use an aggregator like Ozzai (which I built for exactly this reason) to access multiple models through one interface. The point isn't which solution you choose. The point is you're not limiting yourself to a single model's perspective. And whatever you use, make sure your data stays private.
From Tool to Collaborator
Now, here's where the real shift happens.
Most people still treat AI like a search engine or a writing assistant. Ask a question, get an answer. Draft an email, move on. That's using the language app to look up a word. It's not fluency. It's not even close.
Fluency happens when you stop thinking of AI as a tool and start treating it as a collaborator. That means bringing it hard problems. Real problems. Things you're actually trying to figure out in your work or your life. Not "draft me an email" but "I'm trying to restructure how my team handles project intake and I need to think through the constraints and options."
Start by telling Claude what you're trying to achieve. The boundaries. The constraints. What success looks like. Then work through it together. Iterate. Push back. Let the conversation go somewhere. That's how you build fluency. Through immersion and real use.
The Wave
Your company may or may not be investing in AI capability. If they are, great. Take everything they offer and supplement it with your own hours. If they're not, that's not a reason to wait. This is your career, your capability, your future. A year ago, AI was relevant to 36% of jobs. Today it's 49%. That's not a trend. That's a wave, and it's not waiting for you to get comfortable.
Start Now
The thousand hours will happen eventually. You can put them in deliberately, starting now, building genuine fluency while the window is wide open. Or you can drift into them slowly, always a step behind, always translating in your head instead of thinking natively. The multipliers keep coming when you run at this. The sooner you build the base, the more they compound in your favour.
I've done my thousand hours and kept going. I've pivoted my entire career to be AI-centric. I still feel like I can't keep up, but the opportunities are endless.
Make your move, human. Start this week.
Damien Healy is the founder of Qanara, an Australian AI consultancy helping businesses accelerate from strategy to impact. He writes about AI-native workflows, frontier AI capabilities, and practical transformation.