i expected AI to make me faster. i didn’t expect it to make me slow down.
(“chat, explain it to me like i’m five”)
Google made the world’s knowledge searchable. ChatGPT made it approachable. before AI, i could find information about almost anything, but finding it wasn’t the same as understanding it. search results often dropped me into explanations written for people who already spoke the language.
“explain it to me like i’m five” is a common prompt people use to make complex subjects understandable without requiring years of background knowledge. the point isn’t to talk down to someone, but to meet them where they are. i’ve learned a lot from that approach, especially because it works both ways.
“prompt engineering” sounds technical, but most of it is learning to say what you actually mean. the problem is that many times, i don’t know exactly what i want. AI makes that obvious very quickly. as i get better at using AI in my everyday work, i’m becoming more precise with my words.
when i explain the cloud to my dad, and why a million USB drives may not be the best long-term strategy, i have to start with what he already understands. he is a brilliant man. he just knows different things than i do. i have to remember that when explaining things i now take for granted.
we recently hired our first two employees and had two months to prepare them to run the business by themselves while Bailey and i were unavailable on our honeymoon cruise around the Aegean Sea. for one of them, this was their first job.
we had to take everything that lived in our heads and put it somewhere they could use it: instructions, checklists, event briefs, and what to do when reality ignored the plan.
(they killed it, btw.)
even while writing this newsletter, i am carefully choosing how much to explain. too little, and i lose you. too much, and it starts to sound like i actually think you’re five.
misunderstandings still happen. i just respond to them differently now. when AI gives me the wrong result, i start by looking at what i failed to explain. increasingly, i do the same with people. i take ownership, clarify what i meant, and start again.
talking to robots has made me better at talking to humans.
(reading at human speed.)
the good thing about AI is that you can get an answer at the speed of a supercomputer. the bad thing is that you still have to read it at the speed of a human. short answers are easy enough to skim. but when the subject becomes complicated, there is no shortcut. i have to slow down, read carefully, and occasionally read it again.
years of feeds, videos, reels, and shorts trained me to expect the point immediately. somewhere along the way, reading more than a few sentences started to feel like work.
(ain’t nobody got time for that.)
the trade-off has been missing out on deep exploration of topics in long-form books, articles, and editorials.
as i work more with AI, i’ve been forced to read more than ever before, and it’s translated into me reading more off-screen too. i’ve read more books in the last couple of months than i had in the last couple of years. that includes the books that i am now applying to the new consultancy i am building.
(the latter is what told me to start this newsletter.)
AI got me reading again. once i rebuilt the patience to sit with its answers, i found myself bringing that patience back to books written by actual humans too.
(attention is currency in the land of the robots.)
reading requires focus. lately, mine has become harder to keep. using AI also involves a surprising amount of waiting. i give an AI agent a task, and it thinks.
(and thinks.)
most times i spend more time watching status updates about what the agent is doing than i spend typing.
(what am i to do?)
there is always something more dopamine-rich waiting one tab away: Instagram, YouTube, or a notification that has absolutely nothing to do with the thing i am supposed to be doing. but checking one thing is never just checking one thing. if i break my state of flow, a task that should take a few minutes can somehow take all day. sometimes several days. so i fight the urge.
(and wait.) (and wait.) (and wait.)
(most of the time.)
AI makes work faster, but it also creates more opportunities to abandon it. when output is abundant, focus becomes scarce, and attention is currency in the land of the robots.
staying with a task while the computer works has become its own practice. focus is resisting the impulse to leave. patience is staying present without resentment. every time the machine makes me wait, i practice both.
the strange part is that the practice does not stay at my desk. when someone takes longer than i think they should, i am a little less likely to become anxious. when something doesn’t click as quickly as i want, i can slow down.
(breathe slow.)
(start again.)
(it will take as long as it takes.)
that patience is showing up with my parents, my sister, my wife, and Tiny Dogge (our toy poodle). the tech is getting faster. somehow, it is teaching me how to slow down.
(judgment is at a premium.)
recently, i showed some colleagues how i have been using AI to make my work more efficient. i created an interactive mockup for a product we were designing in a fraction of the time it would have taken before. it was polished enough to click through and convincing enough to make one of them nervous.
“if we could build something this quickly, what would stop the customer from using AI to build it themselves?” they objected. the notion didn’t faze me at all.
(i had my AI existential crisis years ago.)
it was a fair question, but the mockup was never the whole job. the work was knowing what to mock up, what problem we were trying to solve, which decisions we could make quickly, and which ones we would have to defend later. AI could generate the interface. it could not take responsibility for the judgment behind it.
this has made me think about extreme ownership differently. AI can take on more of the production, but none of the accountability. if the answer is wrong, the recommendation is harmful, or the thing simply wasn’t worth making, “the robot made it” is not a defense.
the faster AI becomes, the more important judgment becomes. when making something is cheap, knowing what deserves to be made becomes valuable. i can use AI to write, research, design, and build, but my name still goes on the work. that means the responsibility is still mine.
the robot may have generated the answer. i am still the one who decided what happened next.
so this is the question i am asking myself. and now you. as AI makes you more capable, how is it changing you? is it for the better, worse, or something in-between. (or totally different.)
i’d love to hear from you.