LLMs and performative productivity - Josh Collinsworth blog
Search post titles<br>Popular posts<br>AI optimism is a class privilege<br>The blissful zen of a good side project<br>Things you forgot (or never knew) because of React<br>Ten tips for better CSS transitions and animations<br>Why you should never use px to set font-size in CSS<br>The self-fulfilling prophecy of React<br>Building accessible toggle buttons (with examples for Svelte, Vue, and React)<br>Categories<br>opinion<br>personal<br>AI<br>satire<br>popular<br>web<br>design<br>advice<br>wordpress<br>gear<br>css<br>a11y<br>react<br>javascript<br>svelte<br>vue<br>php<br>More links<br>All blog posts<br>More about me<br>Home
LLMs and performative productivity<br>Published: June 5, 2026<br>I began to wonder exactly how productive AI1 is making me, personally, when I stepped back a bit from what I’d been using it for.<br>At first, when I began to use agents day-to-day, I was blown away by their capabilities, and by how much they elevated my own. Certainly, the floor was raised. I could do far more, much faster. That much was undeniable.<br>Armed with this newfound power, I accomplished a flurry of tasks I either wasn’t capable of, or didn’t have the time for previously:<br>At work, I could get up and running in new codebases without asking for help, and could contribute to them much more easily<br>I got some projects updated, moved, or refactored, all in record time (including a particularly gnarly Nuxt upgrade I’d been putting off for years, and that would’ve taken me days of work, done in about an hour)<br>I added several new features to a handful of apps here and there that I wouldn’t have otherwise<br>I scaffolded new things and built out greenfield projects in record time<br>I wrote more tests, faster than ever<br>I pushed out a whole bunch of bug fixes<br>That all sounds fantastic, of course. It felt fantastic.<br>But when I got done with all that, I had to wonder: could I really call any of that productive ?<br>At work, I didn’t understand the codebases I was working in, and though I was contributing to them, I gained no real context about them. I was opening PRs, but I couldn’t really defend what was in them, or say whether or not they worked with the system. I was constantly afraid I’d messed something up without realizing it<br>Most of the other updates weren’t really needed per se, and afterwards, the apps themselves weren’t any different; the changes just made me feel good, while making little to no difference on the user side of the software<br>The new features I built were neat, but they weren’t actually being used<br>The greenfield projects were quickly abandoned<br>I didn’t really need the tests that bad, and I wasn’t quite sure whether they were doing the anything worthwhile in the first place<br>I didn’t know what had caused the bugs, or what the fixes had been<br>I had mainly just checked off a bunch of old to-dos, most of which hadn’t gotten done previously because they never mattered that much in the first place.<br>I had mainly just checked off a bunch of old to-dos, most of which hadn’t gotten done previously because they never mattered that much in the first place.
And even where they did matter, I paid a cost for doing more, faster. I added a bunch of abandoned side-projects to the old pile, but unlike before, I didn’t even come away with any new skills or experience.<br>If anything, it seemed like I knew less than before.<br>Maybe the code improved, but I sure didn’t.<br>And that was all in the best-case scenarios, where the agent worked well. Other times, I’d spend so long prompting and re-prompting it would’ve just been faster to do the work myself in the first place—but by that point, of course, I was so deep in the hole it seemed easier to just keep digging.<br>I had to ask: when the dust had settled, was it really all a net gain?<br>If I’m being honest with myself, what I was doing was often more theatrical than productive; lots of show, with not a lot to show for it.<br>If I’m being honest with myself, what I was doing was often more theatrical than productive; lots of show, with not a lot to show for it.
But being honest with myself, it turns out, was actually a lot more difficult than it should have been.<br>Because, despite everything above…I loved it.<br>I loved using my agent like a guilty pleasure. I wanted to keep using it, any chance I got.<br>I could see with my own two eyes that I was trading away something valuable for something petty, like a kid blowing their allowance at a gumball machine.<br>But I still wanted what the machine had. It triggered something in my brain. And I believed maybe if I used this thing enough, all the tiny little meaningless tasks would eventually add up to something important.<br>Sometimes I’d feel a compulsion to fire up Claude Code and have it work on something, even when I had nothing in mind to accomplish.<br>I’m very familiar with that sort of compulsion, and I recognized it as soon as I stepped back: I wanted to play AI, like a video game. If I had no reason or goal in mind, I’d try to make one up. I just wanted more of that feeling, like I was...