Core Insights
These days there is a solution to almost every problem: ask AI.
Your code is not working? Send it to AI.
You have no idea what to design? Ask AI for some ideas.
You do not understand something? Ask AI to explain.
You do not know what to write? Ask AI for an ideas and pick one.
I use these tools too. It would make sense to ignore something that can save hours of work and solve certain problems in seconds.
There is an important difference between using AI to get something done faster and using AI so that you do not have to think about the problem at all.
I think that is where things get interesting.
Take a programming problem. You can send the code to an AI assistant. Ask AI to find the bug. There is nothing with that.
There is another way to approach it. Spend a few minutes looking at the code first. Try to figure out what is wrong. Make a guess. Maybe even write a fix. Then give it to AI. Ask, "I think the problem is here. Am I missing something?”
In both cases you might end up with the answer.
The difference is what happened before you got there.
In the case you actually had to engage with the problem.
That might not seem important when you are fixing one bug.. When you repeat this process every day for months it starts to matter.
If every time we face something we immediately ask AI for the answer we spend less time practicing the part that actually makes us better at solving problems.
This does not mean AI is making us stupid. It is more straightforward than that.
The things we practice tend to get stronger. The things we stop practicing tend to get weaker.
Problem-solving works the way.
The funny thing is that the hardest part of learning is often the part we're most eager to remove.
When we do not understand something, when our code refuses to work or when we have no idea where to start with a design getting an instant answer feels great.
Sometimes those frustrating ten minutes are more valuable than the answer itself.
During those ten minutes we are trying things making guesses, connecting ideas getting things wrong and trying again.
That is learning.
So I do not think the answer is to use AI
I think we need to become more deliberate about when we use AI.
For example when I am working on a design I do not necessarily want my first idea to come from AI. I can start by understanding the problem sketching an ideas and exploring them myself. Once I have something I can ask AI to challenge my decisions.
Now AI is not designing for me.
AI is helping me look at my design from another perspective.
The same idea applies to learning.
If I do not understand a concept I can ask AI to explain it.. I should not stop there. I can close the answer afterward. Try to explain the concept in my own words.
If I cannot explain it I probably have not learned it yet.
I have just seen the answer.
Those two things are very different.
One habit that I think is worth keeping is simply pausing before asking.
Not for every thing. There is no point spending thirty minutes solving a problem that AI can solve correctly in ten seconds.
When I am actually trying to learn something I can give myself a few minutes first.
I can make a guess.
Even if it is wrong.
In fact being wrong can make the correct answer more useful. When I already have an idea, in my head I can compare it with the answer. Understand where my reasoning went wrong.
Another thing I have found useful is not always asking AI to create something from scratch.
Sometimes it is better to create something yourself and then ask AI to review it.
Of:
“Design this page for me.”
Try:
“Here is the design I made. What UX problems do you see?”
of:
“Build this feature for me.”
Try:
“Here is the code I wrote. What could be improved?”
It feels like a second pair of eyes than a person doing the whole job for you.
I think this matters beyond design and programming.
If we start asking AI for every idea, every decision, every paragraph and every small problem we may slowly develop a habit of waiting for an answer before we even form our opinion.
That is the part I would be careful about.
We are not going to beat AI at everything.
AI can process information generate more variations and produce text or code much faster than we can. Trying to compete with AI on those terms does not make sense.
What still matters is knowing what we are trying to build why we are building it. Whether the answer we got actually makes sense.
AI can give you a hundred ideas for a product.
AI cannot decide which problem is worth solving for you.
AI can generate thousands of lines of code.
AI cannot automatically know whether that architecture makes sense for the product you are actually building.
AI can write an article.
Ai cannot give you a reason to care about the subject.
That is where I think our role changes.
The future is probably not about people who use AI the least. It will be about people who know how to use AI
Using AI well does not mean handing over every part of the thinking process.
For me a simple rule makes sense:
Think first.
Try it yourself.
Ask for help when you need it.
Check the answer.
Then make your decision.
It does not need to be followed every time. For tasks there is little value, in doing everything manually just to prove that you can.
When you are learning something developing a skill or making an important decision those extra few minutes of thinking can be worth it.
Because in a world where getting an answer is becoming easier every day the ability to ask a question sit with a difficult problem and form your own opinion before seeing the answer might become more valuable not less.
AI is not necessarily going to take thinking from us.
We just have to decide how much of our thinking we are willing to give AI.