← All writing

How I'm Thinking About AI Right Now

· 3 min read

I think generative AI is going to change every part of society. But the more I sit with it, the more I find myself pulled in two directions, so I wanted to get down where I’ve landed, for now.

Where I’m coming from

When I first started using Gen AI, I was all for it. I understood the risk people talk about, that if you lean on it too much you can lose the ability to think critically. But I come at this from a specific place.

My mom had to pay for an English tutor because I was that bad at English. Not every family can do that.

For someone who can’t afford a tutor, AI opens up a huge amount of possibility. You can effectively build your own tutor for next to nothing. And yes, that takes something away from the person who used to do that job, and I don’t want to pretend it doesn’t. But the access it creates is real, and it actually matters.

It depends on how you use it

The argument I hear most is “AI is going to kill critical thinking.” I don’t think that’s necessarily wrong, but there are caveats, and I think it misses most of the picture.

An analogy I like to use is Excel. Excel took away the need to write out calculations by hand, and it didn’t make people dumber. It moved them up to higher-level work.

AI is the same, just bigger. If your mindset is “I don’t want to do the thinking, just make the decision for me,” these tools will hurt you. If your mindset is “I want to actually understand this, teach me,” and you use it to work things through in your own head, they’re invaluable. Used well, it sharpens your thinking and keeps you on track with your ideas and plans.

The caveat to my own argument is that which end you land on isn’t purely a choice. Most of these tools are built to hand you the answer, and when you’re tired or on a deadline, the answer is the easy path. Using AI to think better takes discipline, which means the people who benefit most are often the ones who already had good learning habits. That’s worth being honest about.

The part I’m torn on

Part of me is genuinely excited that I can build tools custom to myself. Modular software has been a concept we’ve talked about for years, and AI is making it real.

The other part of me knows there are companies employing a lot of people whose products can now be replicated by one person with AI. Now, there are caveats. Will something I build myself have the security that whole teams spend their careers on? Absolutely not, at least not today. But the pressure on those jobs is real.

I don’t think the honest position is pure optimism or pure doom. Usually the work doesn’t disappear, it gets repriced. Tutoring gets cheaper and reaches more people, while individual tutors earn less. Both of those things are true at the same time.

Where I’m leaning

I hate that it takes away from jobs. But I really like what it does for people like me who can’t afford all of those things right now, and for creators who don’t have the budgets that millionaires and billionaires have. The playing field is levelling out. We can all compete.

The catch is that if everyone can compete, you can’t just do what the well-funded players did in the past and expect to end up where they did. When everyone has the same tool, the tool stops being the edge. The edge moves to what AI can’t hand you: real domain knowledge, judgment, taste, trust and relationships. You need to think above and beyond.

So overall, AI doesn’t make expertise less necessary. It makes people with real expertise more dangerous.