Everyone is talking about AI at this point. Every week, there is a new model, new tool, a new “game changer” (if you scroll LinkedIn or X).
If you have tried to keep up, probably you felt exhausted, the same way I did. It is not that AI is too hard to learn, but the noise around make it seem almost impossible.
Well, the good news is that you don’t need to learn everything.
You just need a system.
The Problem with “AI Overload”
AI world is moving faster than any other in the tech space. New models drop before you are even done reading the release notes of the previous model.
And every company all of a sudden claims that they are “AI powered”.
If you try to keep up with everything, chances are high you will burn out. You will spend more time reading threads than actually learning something.
Treating AI Like a Data Problem
Think of learning AI like managing data. You won’t capture everything. You capture what is relevant, structured and useful for your purpose.
Same thing with AI. You do not need to know every model, framework or paper out there. You just need to understand:
What AI is actually good at
How it fits into your domain
How to use it practically
The rest? Noise.
How I Would Approach It
Here is a simple 3 step approach I use, which worked pretty well for me without giving me the feeling I have to work full time on trying to learn all of it.
Learn Concepts, Not Tools
Understand the basics of how AI systems work. You do not need to read research papers, just try to get familiar with the terms like:
Model training vs inference
prompting vs fine tuning
embeddings, vector search, agents
If you understand these, then you can adapt to any new tool that comes out next month.
Learn by Doing (Small Projects)
Pick one use case that is applicable to whatever you do. If you are in data, automate a reporting summary, classify metrics, or built a chatbot for FAQs.
Don’t aim to “learn AI”. Aim to use AI meaningfully where you see impact directly. Every project you work on yourself compounds way more than you watching Youtube videos on the same topic.
Filter Ruthlessly
You do not need 20 to subscribe to 20 newsletters and join 50 Discord servers. Follow a few people or sources who simplify the space for you and ignore the rest.
AI moves fast, but core ideas evolve slowly. Learn fundamentals, stay curious, and let the hype pass you by.
Learning the Basics Stack
Here is what I would recommend if you want to learn AI without feeling overwhelmed:
Google’s “Introduction to AI” or DeepLearning AI short courses.
Youtube Channels like: StatQuest (for conceptual clarity), Two Minute Papers (for staying inspired), Mat Wolfe (for tool overviews).
Books: “Architects of Intelligence” - Martin Ford, or “You Look Like a Thing and I Love You” - Janelle Shane.
DataExec Newsletter (if you are interested to learn about data and AI)
There is plenty of available sources out there for free. So do not pay for anything at this stage.
Build Small Projects
Try Hugging Face Spaces to play with models without coding much.
Use Google Colab for free Python based experimentation.
Pick one idea, maybe “Summarize my daily metrics report using AI”, and just build it.
The Takeaway
The main point I am trying to get across here is that you need to focus (a.k.a lock in) with learning the basics and slowly expand as you get familiar with AI. You won’t need to master every tool or model out there. Try to build one thing which makes your life easier, understand how it works and then rinse and repeat.
I would rather have you move from a passive observer into an actual builder, without feeling overwhelmed and burn out.
In the upcoming newsletters, I will try to get more practical on AI and Data. If there is any topic specifically that you would be interested to learn more, reply to this email or comment and I will make sure to capture that in one of my upcoming newsletters.
While we have you here, if you are in the process of trying to apply for jobs, I wrote a detailed guide on how to use AI to land your next job: How to Use AI to Land Your Next Job.