Earlier this morning I didn't feel like sleeping, so I was scrolling through X and seeing a lot of AI content.
I keep seeing AI-workflow content that sells you a template. Like a 20-agent orchestration, some complex pipeline, a "system" you're supposed to copy. It looks impressive on paper. But when I actually think it through for my own situation, most of it doesn't fit.
I'm basically just doing plan-driven development, with a few skills I built, like having Codex run adversarial code review to actually track the output. There's guardrails and a lightweight harness I try to keep improving for each project.
That's kind of it. I don't see the need to suddenly run a really complicated orchestration system with 20 agents, and I don't find that kind of content on the internet to be that useful.
Some of the X posts tried to hook their audience by asking: "Instead of watching TV or doing nothing this evening, go learn about this new thing in AI, and then you'll have a new skill." They were imploring their audience to use their time to learn, and that learning is supposed to result in some kind of gain in the future.
That gave me a flashback to a tempting dilemma I feel every day. Instead of spending two hours looking at how to improve my AI system, I might as well just be using my usual AI workflow to finish the tasks that I told myself I'd do today. There's a kind of odd future calculation that comes from that.
What if improving my AI workflow today for two hours instead of doing what I'm supposed to do somehow makes me faster at completing stuff later? There's not one right way to approach it at each moments. But decisions can compound.
If, in key phases, I'm not hitting the milestones that I've set for myself because I keep deciding to do something that feels productive, like "improving my workflow," then what that compounds into is nothing.
If I focus 90% of my time on just getting the tasks done using the workflows and tools that I have that work, that compounds into something, into the actual shipping of products. How delightful! I can also spend some time along the way to figure out how to improve my shipping speed with AI.
A team might genuinely benefit from something more complex. But I'm not a team. I think if you want to be more productive, it's based on your own situation, your goals, what you're working on, the outcomes you're looking for — and honestly just the way you think. There's no one-size-fits-all template for this stuff. It's more helpful to pick smaller things you can add and grow your own harness over time.
The thing I keep coming back to is "pain". If something is actually painful, I'll solve it with an improvement to my workflow. If something's already solved with no friction, there's no real motivation to keep using or improving it. Then that same thing that seemed easy and convenient eventually stagnates and I stop using it.
I became a lot more invested in making autonomous team workflows and tools work because of my friction, not because they sounded cool. I feel the pain of having to kickstart every decision myself, and not knowing which bugs in my Linear project I should actually fix next. I'd tried a random "AI team" repo before that — code quality was pretty bad, and it just wasn't autonomous enough.
I think there are two ways to make AI content. You can teach it grounded in your own actual expertise, and make it genuinely useful. Or you can frame your content so it gives people FOMO and they click. I've been noticing posts and orchestrations that look nice on paper but, when I actually trace them through, don't make much sense for me.
Honestly, I (and who knows, you?) might be better off just asking Claude — or whatever AI — based on what I have locally and what my git history looks like, to suggest some things grounded in my real pains. Not the imaginary problems someone's trying to sell a solution to.
So if I could choose between fumbling around and just growing and pruning my own harness over time versus being hand gifted an incredibly complex, cool, supposedly very efficient system from day one, I would choose to fumble around. I think that productivity FOMO, in and of itself, drains time and energy.
It's like life. We can feel FOMO about anything, and spend a lot of time comparing ourselves to other people. Or you can set your own goals and steadily do your best on them at your own pace. That could get you where you want to be.