I use AI every day.
I build things with it, experiment with it, write code with it, learn about new things with it, and generally think it's one of the most interesting technologies I've ever had access to.
So this is definitely not an "AI is ruining everything" post.
But I do think AI is creating a problem that we probably need to talk about more:
It has become incredibly easy to avoid thinking.
And that cannot be good for us.
We used to have to struggle a little
Say you have to write an essay about something you don't care about and are not at all interested in.
Maybe it's for school. Maybe it's for work. Whatever it's for, somebody wants you to write 1,500 words about a topic you know absolutely nothing about.
20 years ago that kinda sucked.
You had to research it.
You had to figure out what people were talking about.
Read the same pages over and over again til they stuck with you.
Take notes.
Then you had to decide which information was important.
You had to organize your thoughts.
You had to stare at a blank page for twenty minutes because you had no idea where to start.
You'd write something, hate it, rewrite it, realize you misunderstood something, go look it up again, take a break, come back to it, and eventually, little by little you would piece together something decent.
But you had to think, and you had to do the hard thing because no one was going to do it for you.
And somewhere in all of that annoyance, you probably learned something.
Not just about the subject.
You practiced researching.
You practiced deciding whether a source made sense.
You practiced turning information into your own understanding.
You practiced explaining something.
You practiced writing.
You practiced being stuck.
That last one might be the most important.
And honestly you probably were feeling pretty good at the end of it all because you felt like you had accomplished something.
Now the hard part is optional
These days all you need to do is type:
Write me a 1,500-word essay about this topic.
And thirty seconds later, there it is.
Maybe you edit it.
Maybe you fix a few sentences.
Maybe you read through it closely enough to make sure it doesn't say anything stupid.
But you didn't actually have to do most of the work that used to happen between receiving the assignment and having the finished result.
AI did the research.
AI organized the ideas.
AI figured out the structure.
AI found the words.
The end product might be perfectly fine.
But producing the end product was never the only point.
Sometimes the process was teaching you something too.
I see this with kids
This part really worries me.
If a kid has 10 math problems for homework and doesn't understand number 7, there are two very different ways to use AI.
One is:
What's the answer to number 7?
The other is:
Can you help me understand how to solve this without giving me the answer?
Those might look like tiny differences in prompting.
They're not.
One replaces the thinking.
The other supports it.
And when the easiest option is always sitting three seconds away, I think it takes a surprising amount of discipline to choose the harder one.
Adults definitely do it too.
Why spend an hour figuring something out when you can ask an AI?
Why read the documentation?
Why wrestle with the bug?
Why write the first draft?
Why sit there thinking about how you actually feel about something?
Why learn the formula when something else can calculate it for you forever?
Sometimes the correct answer genuinely is: don't waste your time doing this manually.
I automate things specifically because I don't want to do repetitive crap manually.
But not all difficulty is useless friction.
Sometimes difficulty is the exercise.
I think we've started confusing "hard" with "unnecessary"
This is probably the part I think about the most.
We've spent decades making technology remove friction from our lives.
Usually that's awesome.
I don't particularly miss having to drive to the library to do research.
I don't think programmers need to give up autocomplete and start writing machine code to prove we're still using our brains.
There is no virtue in making something difficult solely for the purpose of making it difficult.
But I think we've started treating any mental friction as a problem technology should solve for us, and we're starting to lose the desire to accomplish something on our own.
Struggling to understand something is uncomfortable.
Sitting with a problem is uncomfortable.
Being confused is uncomfortable.
Writing badly before you learn to write well is uncomfortable.
Trying 3 solutions that don't work is uncomfortable.
Honestly, just being wrong is uncomfortable.
But those experiences also build something.
They build judgment.
They build persistence.
They build the ability to look at something unfamiliar and think:
I have no idea how to do this, but I can probably figure it out.
I'd really hate for us to lose that completely.
AI can either make us better or much worse
The weird thing is that AI might be one of the greatest learning tools we've ever invented.
You can ask it to explain something to you 5 different ways until you get it.
You can ask it to test you.
You can argue with it.
You can tell it your answer and ask what mistakes you made.
You can ask it for hints instead of the answer itself.
Learning can be incredibly fast with these tools.
That's the part of AI I love.
But the same tool can also let you skip the entire learning process.
And the difference isn't really the technology.
It's how we decide to use it.
Sometimes I'll ask AI something and realize afterward that I probably should have spent another ten minutes thinking about it myself first.
Not because using AI was cheating.
Not because the machine somehow contaminated the answer.
Just because I needed the exercise.
Maybe we need to protect some of the struggle
I'm definitely not suggesting we stop using AI.
Good luck prying it out of my hands.
I think maybe we have to be more deliberate about choosing what aspects of our thought process we're comfortable delegating.
Let the AI help you understand the chapter, but maybe don't let it read the chapter for you every time.
Let the AI point out why your math solution is wrong, but maybe try to work out the problem yourself first.
Let the AI help with debugging your code, but maybe take some time to try to understand the bug a little bit before throwing it at an AI chatbot.
Let the AI critique your argument, but maybe figure out what you think before asking it what you should think.
Because if we outsource every difficult thing the instant it becomes uncomfortable, eventually we're going to become very good at operating AI tools and very bad at operating without them.
AI should give us more ability to think.
It shouldn't give us an excuse not to.
Top comments (1)
this is exactly why i built a "constitution" into my ai mentor. if it just writes the code for u, u learn nothing. koda is forced to explain the why and ask questions back. ai should be a sparring partner, not a crutch. 🐯