How I'm thinking about AI

To catch up new readers and a recap for those who missed previous thoughts

Some people think AI is “too new” to have real thoughts on, but I don’t think that’s the case at all. No matter how good these tools get it’s all still just the automation discussion and we’ve been having that on this blog and the old one for many years.

Anyway, I’ve also been tinkering with these tools since the early days and putting down thoughts for a bit, and something many people like to deny is tradeoffs exist, like they do everywhere else. You can get a car and drive to get places quicker and that’s great, but if you don’t also go to the gym and hike you will end up obese and sickly. The intellectual equivalent is real, too.

Because someone asked, I thought this week I’d put down a list of some of my favorite ideas on this topic in one place (not in a particular order). These should hopefully help anyone from creatives to managers walk a better future path.

The rise of AI nihilism: if we did turn all creative work over to machines, we’ll have more output but considerably less meaning.

How ‘The Time Machine’ explains the future of knowledge work: AI will create a new divide b/w people who let machines do all thinking for them while their creative brain atrophies, and a few holdouts remain sharp.

You are not Rick Rubin: Rick Rubin's influence was part of a totally different world and not people think, and AI can't recreate what made him valuable.

AI is already making people less creative: research suggests leaning heavily on LLMs can flatten the ways people think, write, and solve problems.

Don’t let machines or the crowd decide your world: When algorithms and consensus make all the decisions for us, taste gradually become someone else's, deliberately choosing things for yourself matters.

You can’t prompt your way to meaning: Having AI produce finished creative product may save effort, but removes the struggle that makes the work personally rewarding and kills all the small decisions that matter.

Writing is thinking: If you outsource your writing, you end up losing the muscle for thinking that writing forces you to do.

Resist slop world: The easier it is for people to flood the world with mediocre AI-generated spam, the more valuable genuine effort, taste, and craftsmanship becomes.

People are openly bragging about AI spam operations: There’s a huge difference between a real business and webspam, yet some people are clueless enough not to understand it.

You can’t call people “luddites” for having taste: A person can embrace tech while still believing tools are secondary to skill, judgment, and the desire to make something excellent.

Where does AI actually make sense: There are plenty of boring/repetitive tasks where AI is an obvious win, but replacing human judgment makes far less sense when the value of the work lies in its individuality.

Some friction is good: The work some spend so much time trying to automate can frequently be what force people to develop a distinctive voice.

AI’s narrative problem: The industry has a huge communication problem, and it’s worth solving now before too many turn into doomers.

The post-literate age is here: In a culture increasingly hostile to reading and focused attention, being someone who can still sit with a book and follow an argument all the way through has a massive advantage.

I hope you too are spending time deeply considering this topic, what you think is worth automating from your creative work (and what is worth keeping). As I go into in these stories there definitely can be a productive use. But machine intelligence is not the same as human intelligence, it’s just a poor facsimile. Know where it makes sense for your own work, and where human judgement is necessary. And don’t lose your edge to the clankers.