How to use AI and still love writing

Writing About Writing

oh man,

who am I gonna piss off with this one?

I've actually been avoiding writing this, for a few reasons I'll get into momentarily. And in the midst of this I've found myself trying to write about all sorts of other things ...

An article on productivity (ironic), an article on free writing, one on making money as a writer through copywriting, a rant on generic advice, a fiction story that's slowly coming together kind of like how oil slowly comes together over millions of years compressing bones in the earth ...

But every time my eyes slide across the title of this article, I get that 'catch' - a little feeling which says YEA THAT’S THE ONE GO WRITE THAT ONE.

Sigh

Fine.

My working title is "how to use AI as a writer."

But one of my major problems with this is I don't see myself as an AI expert. I should probably just ignore that since I've sold a course on using AI and I use it quite effectively every day, AND my usage of it is only getting more effective and evolving further.

Secondly,

My specific area of experience using AI is with copywriting, primarily focused on email (so the tactical direction and example is narrow, but the principled and even strategic application is wide).

THIRD ...

I actually hate AI.

Well.

Hate is a strong word but the dis-appreciation of it's existence is pretty strong.

So strong that a bit over a year ago I made a course on AI and pretty quickly came to the conclusion that I didn't want to be an AI expert and I really just wanted to write all the words myself.

Because that's where I'm alive.

It was actually in that process of getting heavily into AI (a bit over a year ago) which highlighted just how much of my strengths and gifts are in writing itself.

And so, in spite of AI's existence, I went full in with The Guardian Academy and Man Bites Dog just writing everything myself. I wrote the entire launch of The Arena (sales pages and sales letters) with no AI, because I had it all in me and I couldn't conjure up a compelling reason to have AI do that work when, not only could I very well do it all myself, but there was no way I was gonna get AI to do what was in my head anyway.

I've written every article I've published for TGA without AI, and ... for the most part ... the same with MBD.

(Here's one of those 'Arena' sales pages if you're interested — that’s on a google doc — it's a bit of a different style than you're probably used to)

We basically spun The Arena up out of nothing and now it's the backbone of The Guardian Academy - everything is about doing the work - Engaging the Field - where you can actually learn by changing your behavior (and the structure of The Arena is that structure of learning and changing your behavior).

So obviously ...

I didn't need AI.

But,

I still have used it when it's suited me.

In the original course I made, I talked about how the place I primarily use AI has been in my "prescriptive" writing - or rather, writing where I have a process, known inputs and expected outputs. A lot of copywriting projects fit into that category.

There's one idea which really has defined well, why I don't like to use AI except when I do ...

I write because I don't know what I think until I see what I say - Flannery O'Conner

It became clear to me that writing is a process of thinking, processing, understanding and presenting all at the same time to me.

It also became clear that if I don't write regularly, I kind of start to go INSANE.

Like I just can't think straight or do things well at all if I don't consistently write. It's like the ideas get blocked up in my head and then tangled and mashed together if I don't allow them to clear out and progress. Like being stuck in one place in time on a harddrive and constantly writing over that spot.

(For you young'ns reading this, it used to be that a hard drive was kind of like a record player on a metal disc - and a record player was kind of like a cd player on a biiig vinyl disk with a needle that traveled over bumps to make sound - and a cd player is like a record player except they use a laser to read the bumps --- so on the hard drive the information was physically written into the disk with this needle and if the disc didn't spin properly you'd just write over the same spot over and over and f#ck it all up)

That's been #oldmanspeaking about things that don't exist anymore

(Actually I think spinning hard drives are not obsolete yet since they can pack a ridiculous amount of storage in the space for cheap)

ANYWAY

I go insane.

If I can't write.

And if I'm having the AI do all my writing - THEN I'M NOT WRITING

No writing and no writing make Joseph something something

BUUUT

I kept coming back to AI for things here and there because the amount of time it saves and the level of iteration that can be gone through on ideas is absolutely f#cking bonkers.

I can write a years worth of daily emails in a day. Pre-ai that would take me a month of writing every day (and I'd absolutely hate the density and repetition of it).

I can do an entire email strategy buildout in an hour (emails written and all). Pre-ai that would take me 2 weeks.

Absolutely. F#cking. Bonkers.

So I've had this existential crisis around AI for like the entire last year. Even - as I said - having made an entire course on the matter and then still having used it effectively here and there.

Well,

Two things happened recently.

One - I realized a mental framework I've used for a lot of other things perfectly applies to this situation and actually makes it easier for me to write more on what matters (more on this in a moment)

Two - because I was starting to apply AI to client projects again to create efficiencies, I started looking at what "the experts" are doing annd ...

I discovered that the best AI people in the world are working with, talking about, and teaching AI the exact same principled way that I demonstrated in my course.

Nothing's different.

There are no secrets.

And the principles have proven themselves out over all this time (which ... is also absolutely f#cking bonkers in the AI world to have that level of consistency).

Ok back to One ...

That mental framework.

Explore/Exploit

A brief explanation on the base concept ...

Explore/Exploit is one of several bi-modal risk mitigation models I use in decision making. The way this one works hinges around what is known and unknown, what is certain and uncertain, and time.

If you're trying to create a specific outcome, there are - generally speaking - two categories of “doing.” Known and certain ways to create an outcome, and unknown or uncertain ways to create an outcome. Typically the known and certain ways are ultimately likely limited, while the unknown or uncertain ways have the potential to have a much greater result.

One of the most common examples is a gambling one.

Let's say you show up at a casino in Vegas and you start hitting the slot machines. After a couple machines you find one that's regularly outputting positively - it's not huge jackpots worth, but you're getting a few more coins out than you put in every few minutes.

Now

The question is ... do you keep using that slot machine, or do you go find another one?

THE ANSWER ...

IS,

How much time do you have?

If you're just killing 20 minutes while waiting on a friend before you go out to dinner, then you probably want to just keep exploiting the positive outcome you know is working. To try exploring to find a better result with your limited time is just gambling.

But if you've got hours, or days of your time available to quite possibly find a better outcome, well then it probably makes sense to keep exploring.

Hopefully that makes sense.

To step it up a level, the idea of the barbell is to exist on one side or the other, not in between. You can't explore AND exploit simultaneously (that would be akin to marrying someone while trying to date on the side - ends in disaster).

So,

Applying Explore/Exploit to AI is a tool for deciding WHEN to create with AI and when to do the writing myself.

This actually came to me as I was finishing up my article 'Seek The Expansive in Your Potential'

I had figured out at Gray Wolf Summit that my Exponential Potential lies in the writing itself. Every time I have leaned into writing and written more and focused more of my work around the writing - great shit happens. Which is a defining characteristic of discovering your Exponential Potential - things happen and stuff expands in a way that you can't predict beforehand because your entire reality shifts when you find that potential in you which is exponential.

Read more here:

Seek The Expansive in Your Potential

In seeking the ways to lean into my exponential potential, I realized my old framework of looking at AI through the "process" lens was pretty close ...

But what fits better for me now, is Explore / Exploit.

That is,

Whenever we are looking to accomplish something to create a specific outcome, as much as possible we want to Exploit.

This is because the exploitable is known. There is some level of certainty in Exploit (it's not 100%, because time can always change unseen variables). But we look back on what we have done which has created an outcome in the past and repeat that process.

So

With writing ...

When is it the case that I am, most optimally, looking back on what I have done to repeat something to create a known outcome?

Obviously copywriting falls squarely in that category. But anything else where I'm replicating a known process to create a predictable outcome also falls into this category.

Organization. Planning. Observation. Processing. Etc.

When I allow myself to use AI for all that is exploitable, I get to spend a lot more of my creative writing energy and time on that which is explorable.

Which is all of my best writing.

(One of the tricks in going through this process involves a need for present self awareness, so that when it appears I’m exploiting, I don’t accidentally eliminate what could be explored, thus limiting the overall process - but that nuance I think is too much for this article)

Everything that I write which has turned out as my best work has always been an exploration - perhaps yet looking at a known concept or idea but ALWAYS reconsidering it through the lens of my lived experience and growth, which is creating an unknown. It requires me to write about it so that I can know what I really think when I see what I say and then take those words forward in the learning process that writing itself often is.

As I've embraced this perspective on using AI, I've found it to be wildly impactful on a micro level as well in the processes itself.

For example, let's say I'm writing an article for Man Bites Dog where I'm writing about an idea talked about on a workshop (and then using that ultimately as an opportunity to CTA people into the workshop bundle afterward). Much of that content is known because it's already been discussed. I can bring the transcript into the AI, pull out a structured plan and even have it write the entire thing ...

BUT

If I go straight 100% AI it comes out kind of missing something. I still want to instill my own understanding and ideas.

So,

I can write, very quickly and efficiently, the parts of the article that matter for expanding upon interpretation and understanding which only comes from a human lived perspective ... and then the AI can incorporate those in a much more effective structure and flow for the final execution.

Yes I can hear the writers out there in great dissatisfaction that I'd not take the entire writing process into my hands. Sacrilege!

But on Man Bites Dog I'm trying to clearly communicate useful information which can get people effective outcomes, while on the business side leading people to join the paid tier and walk themselves into The Arena.

That's a specific outcome I'm going for, and we have data of people doing this - so we can look in the rearview mirror, see how it was done before, and keep doing it.

EXPLOIT!!!

On the other hand,

This article here on Growing Trees has no such specific outcome. I'm not repeating a process. I'm actively exploring as I write this and as I publish this.

Therefore,

None of the words of this article have been touched by AI.

Why would I do that?

What would that serve me?

If my goal were to churn out article after article then maybe I'd end up turning there - but my goal here is to cultivate the garden of ME as a writer, and share with you my personal insight and understanding.

AI can't do that for me.

And even if it could, I still need to do this for my own damn SANITY.

(if you're interested in AI related stuff, comment below, I've got something new in the works)