Coming up as a developer, a big chunk of every single day was set aside for creating, for making. I got to experience first-hand the joy and frustration of mashing things together and seeing if they worked. For me, that's how I learned so that's how I developed a sense of proficiency as well as a sense of place within whatever I was creating. "React and d3 don't play well together, so have d3 do the math and react control the DOM" <-- hard-earned, blood sweat and tears attempting to make dataviz work is how I got that model burned into my head. I'm sure there's others rattling around up there, but they escape me and run the risk of rabbit-holing this first attempt at an essay, so we'll leave it be.
The point is, every day was focused around making a thing and seeing it appear directly as a result of your thinking and your logic. When it didn't work, you had clear documentation (sometimes) about what you needed to do differently. That cycle (try, fail, adjust, fail less, repeat) was a huge part of learning and a deeply valuable experience.
There is a deep satisfaction that comes from looking at something out in the world and knowing the relationship between what exists and the effort that went into it being there. Making things, and having it be pure output, is a cathartic experience that the internet has opened up and enabled for billions of us.
Even in the 'before times', I consumed a lot of content. Podcasts are a huge source of inspiration for ideas and concepts. Essays are illuminating, inspiring and educational. Tweets are hilarious and eye-opening blips of community building. All of these input streams helped to inform and inspire what to make, and they balanced the output I created. In theory, if you ingest enoguh high-quality thoughts, process them well, your output may be decent.
Additionally, it's healthy to balance this stuff. You can't just be a giant vent of content with no consideration of what to bring in. Well, you can, but that's how you create garbage.
I've found that when I feel most balanced is when I'm on that middle path of challenging myself with what I bring in, and then sharing the output with the larger community.
But that balance has been off for the past (checks notes) 18 months. I was an early-ish adopter of AI-first development, and it started innocently enough: accepting tabs in vs code, debugging my files and error messages in a chat window with claude, accepting edits in cursor, dangerously allowing in claude code, ditching the IDE for Amp in the TUI, working with background agents like Charlie, and now even thinking about concepts like dark software factories. All of these progressions along the AI productivity curve have been massive tailwinds to productivity and iteration speed. I'll look back at this time in 5 years misty-eyed, knowing it's been a magical phase of evolution and that I got to ride the wave.
But it's come at a cost for me. When I'm in create mode, all I do now is read.
On a good day, the split in create mode between reading and writing was probably 90/10 - 90% writing code, 10% reading error messages and docs. Now, that's flipped - 10% promptcraft, 90% reading. Reading responses, still reading error messages (although a lot less), reading ai generated PR review descriptions, reading agent reviews of PRs, reading code. It's a net-positive for my productivity - I'm able to crank out a week-sized feature in an afternoon of focused reading, but it's an entirely different experience from what it was 24 months ago. I've let go of my attempt to balance out the rich inputs in my life with the hard output of direct creation.
So now, weekends are for writing. No AI "prompt me somethig to write about". No "review this essay, is it coherent?" (as if it isn't obvious). Weekends are for getting the brain used to direct ownership and creation of thoughts again. Weekends are for jumping back into create mode. Weekends are for rediscovering the balance.
Weekends are for writing.