I was an AI skeptic. Not the fun kind who gets to be smug later, but the boring kind who just didn’t pay attention. I’d tried ChatGPT when it launched, asked it a few questions, watched it confidently make things up, and moved on. “Neat trick,” I thought, in the same tone I’d use for a card trick at a party I didn’t want to be at.
I’ll admit there was a brief window where I was genuinely delighted by AI, and it was when the image generators were absolutely terrible. Remember that? When you could type “the pope at a rave” and get a grid of nine nightmarish fever dreams that looked like they were rendered by a microwave? I loved that. It felt like the right vibe for the technology: weird, funny, obviously broken, good for exactly one thing and that thing was sending cursed images to group chats. That was an AI I could respect.
Then they got good and I lost interest.
This is the part where I’m supposed to tell you what changed. I don’t have a good answer. Somewhere between “this is a toy” and “I am going to give Anthropic my entire bank account if they let me,” something shifted, and I missed it.
Over the holidays, I started actually playing with Claude. Not asking it trivia questions or making it write poems about my dog. Playing with it. I connected MCP servers. I learned what “skills” were and then learned how to make them. I built systems that talked to my analytics, my help desk, my project management tools. At some point my wife asked what I was doing and I said “working” which was technically true but felt like a lie.
The closest comparison I have is playing RTS games at a LAN party in 1999. That same energy.
Staying up way too late, hyper-focused on something that doesn’t matter but also kind of matters more than anything, surrounded by the vague sense that you’re figuring out something important even if you can’t articulate what. Except now I’m 41 and the LAN party is just me alone in my office and the RTS game is “making my help desk talk to my project management software.” Same guy, honestly.
Here’s the thing about being a late adopter: you don’t get the glory of being early, but you also don’t have to pretend the first version was good.
I skipped the part where these tools were unreliable and landed directly in the part where they’re unsettlingly competent. It’s like missing someone’s awkward teenage years and meeting them as a functioning adult. Less charming, more useful.
I’ve spent the better part of a decade building elaborate productivity systems. Obsidian vaults with careful linking. Notion databases with rollups and relations. Systems for systems. I told myself this was leverage, that the upfront investment would pay dividends. And maybe it did. But there’s something darkly funny about watching an AI do in thirty seconds what my beautifully architected second brain was supposedly optimized for. All those careful tags. All those templates.
The experience has left me with a question I can’t quite answer: am I way behind, or is this just how it is for everyone right now?
Both options are uncomfortable. If I’m behind, that means there are people out there who figured this out months ago and have been quietly compounding advantages while I was still squinting at ChatGPT like it was a magic 8-ball. If everyone’s in the same boat, that means we’re all fumbling through the most significant technological shift of our careers with roughly the same level of sophistication as a kid who just discovered the family computer has games on it.
I think it’s the second one. I talk to other founders and developers, and there’s this shared energy that’s hard to describe. It’s not quite excitement and it’s not quite panic. It’s more like the feeling of realizing the test is open-book but nobody told you and you’re not sure if your notes are even relevant anymore.
I don’t know what I’m doing. I want to be clear about that. I’m building things, automating parts of my business, connecting systems in ways that feel vaguely illegal even though they’re not. But I’m not operating from some coherent framework or grand strategy. I’m just poking at things and seeing what happens.
The good news is that after enough conversations and enough reading and enough late nights connecting MCP servers to things that probably didn’t need to be connected, I’ve finally reached the point where I’m confident nobody else knows what they’re doing either. The people who sound like they have it figured out are just talking louder or more often.
The rest of us are all standing in the same fog, squinting at shapes and hoping we’re pointed in roughly the right direction. That’s weirdly comforting. I don’t think I’m actually behind. I’m exactly where everyone else is: confused and making it up as I go.
I’ve caught up to the uncertainty, I think.
Leave a Reply