A few months ago I was completely lost learning AI.
I was jumping between:
YouTube videos
random tutorials
different tools
Everything felt disconnected.
I knew “a bit of everything”
but couldn’t build anything real.
Then I changed my approach:
→ focus on small practical projects
→ learn only what I need to build
→ apply immediately
That changed everything.
Now learning feels 10x clearer.
If you’re feeling stuck, maybe the problem isn’t you…
maybe it’s how AI is being taught online.
For sure the approach matters a lot. Having some results that you can show while progressing keep the person motivated. Additionally, you can later on reflect on your progress.
Nevertheless ML/DL can feel like a big puzzle at the very beginning. Unless you study it at University, you have to find a correct path to build a good mathematical foundation and sequentially learn about architectures, optimization etc.
Best of luck to you!
This is how I learn things for almost anything, like painting a ledge, AI or a new programming language. It has worked well for me for 20+ years. This is why I like tutorials that take small steps, like the 1990s O’Reilly books, and Indently Python tutorials on Youtube. You learn one simple concept, and apply it before moving on.
Yeah. When I first started learning Python (and a few other programming languages), I read O’Reilly books too. Even if it’s not the quickest route, the foundational knowledge we gain from books is invaluable. Once little design principles and tidbits of information like that sink in, they help us make quick decisions on the spot.
I didn’t read any books. Learnt everything simply by doing it just kept asking AI how to do it and eventually everything sunk in. Universities are redundant now in my opinion. You don’t need to read books anymore unless electricity is an issue otherwise just ask a model how to do it and away you go. They really are excellent teachers. Now there’s nothing I can’t code. I am so far ahead of the whole industry now it isn’t funny.