Nicolas Cole explains how to get better AI outputs by breaking tasks into smaller steps, not by learning AI itself. He calls this "unbundling": identify each micro-task, create an AI skill for it, then chain them together like a supply chain. Writing a book, for example, splits into idea generation, outlining, research, and proofreading.
I feel very strongly that these tools have like unlimited use cases and unlimited power, okay? You pretty much anything you come up with, you can do. And if you can't do it today, there's a high likelihood you're gonna be able to do it tomorrow. So I think the more helpful thing to spend time on, spend time thinking about, spend time practicing, spend time learning. Like, this is why I continue to pay for, I continue to buy courses, I continue to join masterminds, I continue wanting to just need input, right, is because the more valuable thing is being able to look at a task and break the task apart. That's ultimately what we are achieving here. A beginner would go, I want to write a book. Someone intermediate would go, I want to write a book and I want to get it ready to publish on Amazon. An expert unbundles those and goes, I want to write a book, and in order to write a book, I have to go through the following sub-steps.. And then in order to publish it on Amazon, it's not just one, it's actually a series of substeps. And so I am going to take the time to reverse engineer each one of those substeps and create documentation for it and then export all of it as a skill and then talk to Cowork and instruct Cowork on how to do that inside of Notion. And so the takeaway is that it's actually not AI knowledge. This is, I think, what's so hard for people to understand. Yes, there are some new things here where you're like, oh cool, I didn't know AI could do that.
Right.
But the real bottleneck is not actually knowing what AI can do or knowing how it works inside of AI. The real bottleneck is looking at a task and going, wait a second, what are all the substeps of this task? Unbundling it. Because when you can unbundle a task, then you can create individual skills for each step. You think of it like a supply chain. And when you can, when you have more than one step, right? One step is a prompt. Run a prompt, get an output. One step. But when you start seeing things in multiple steps, when you unbundle and break things apart, that's where you can start chaining skills together. That's exactly what I, what I did with this workflow here. It's like, I want you to do all of these things. I don't want you to just do one of them. I want you to do all of them. And I want you to do all of them in a specific way. It's why I have a skill for each one. And the bot— this is like the hardest idea for people to understand is that the bottleneck to doing this has nothing to do with AI. It has everything to do with understanding what actually, what actually goes into that task. Beginners like, I want to write a newsletter. I'm like, I'm going to unbundle. I want to write a newsletter into, all right, first I want to generate 10 high quality ideas. Then I want to pick the best one. Then I want to generate 2 different subject lines so that I can A/B test them. Then I want to outline. The newsletter, but I want to do it in a very specific way. Then I want to go through and do like a, like a research step where maybe we could add in a stat or a study. I want to isolate that. Then when it's done, I want to proofread the whole thing. Then when that's done, I want to ship it over to Notion and I want it to reverse engineer 10 Substack notes for me from the thing that I just wrote. So when someone says, I want to write a newsletter, what I hear are all those substeps, and that's the skill to build., right? That has nothing to do with AI. Does that, does that make sense for everyone? That is, that, that is fundamentally why it is laughable, absolutely laughable, when people are like, AI is destroying everything. No, it's not, because you can't even use the technology if you're not aware of the steps that you have that you're going through in the first place. The steps to understanding AI is not like, how do I have a new, a perfectly nuanced understanding of like back— backend file structure, .md, local saving to hardware file, like lost in the sauce. That's not what matters. What matters is like, yeah, yeah, they're— trust me, they're making these tools for the average human being. They are making these tools for the masses. They're not making these tools for like the top 1% programmers. They're making these tools so that, you know, your grandma over Thanksgiving can look something up and get a quick answer. Okay. So they're making these tools extremely accessible. So it's not really about like, I have to perfectly understand every single thing that's happening under the hood. What it's about is taking a more first principles approach and going, I want to do something. What are the sub-somethings that go into that? And if you can't name the sub-somethings, then you— I could give you a Ferrari, but you're not going to be able to drive it, right? Because AI is just an engine. So I'm harping on this because I really want to emphasize like, this technology is very cool. I encourage everyone to go play with it. I really encourage everyone to connect your Notion and just start chatting with Cowork and even ask it like, hey, could I do this? Hey, can you read this? Hey, does this work? Like, interact with Cowork, connect the two and go have fun. But realize that knowing how to use the technology is only 50% of it. The other 50% is going, what task am I actually trying to do? And can you unbundle and explain that? Task, because if you can't unbundle it into a series of substeps, you're— you can't use AI. And yeah, Erica, exactly. Once you unbundle a task, that is really what a skill is. Like here, each one of these skills I built is a different step in that supply chain, right? First, I have a skill for generating and clarifying ideas. I have skills for if I'm writing a 35,000-word novella or if I'm writing a 60,000 or 80,000-word novel. I also have a skill for how I want it written, skill for my fiction voice, for example. I also have a skill for creating the entire pre-publishing checklist. So it's like, I want to do this thing, I unbundle it into steps, and then you isolate each step and go, how do I build a skill to maximize that step? How do I do that step as amazingly as I possibly can? What information does AI need? What instruction does AI need? Like sometimes what, what I notice happens most often is people try to create skills for too big of a thing. They, they don't unbundle it. They're like, I'm gonna create one skill to try and do this giant thing. And instead it's more helpful to take the big task, break it apart into pieces, and then create skills for each step or each piece.
By the way, if you're brand new to writing online and want a really easy place to start, I recommend reading The Art and Business of Online Writing. I wrote this book after I'd been writing online for over a decade and accumulated over a billion views on my work, monetizing my writing in all sorts of unconventional ways. And I wrote this book to give other writers like you a way to skip all the lessons that I had to learn the hard way and start seeing success from your writing a whole lot faster than I did. So if writing online is something you're thinking about or something you're actively practicing, I recommend picking up a copy of The Art and Business of Online Writing.
What Every Writer Gets Wrong About Their Audience #writing #creator #writingtips
