The conversation about AI has gotten very loud. And in that conversation aspects about the human experience in the AI era get lost.
The Event Horizon is a letter built for founders who want to look clearly at what AI is actually doing—to our work, our thinking, our identity, and our businesses—and lead from that clarity rather than react from the noise.
We’re not here to chase the tools or debate the hype. We’re here to figure out what it means to stay human while building something powerful—and to do that work together, week by week.
Human lead. AI empowered. Always.
Welcome back to The Event Horizon. 🌅
Last week, this letter was about what happens when a system loses its outside reference point and starts drifting from itself, one generation at a time.
This week, I want to talk about the work required to keep yourself autonomous from the AI systems you use to build your mission-driven business. And address the phenomenon of the drift that occurs to the machines who feed on their own output actually happening to you, as well, when you lean too hard on them.
I’m going to tell you where I first recognized that same drift in myself, because it happened to me before I ever found the language for it in a research paper—maybe sharing it will help you recognize it in yourself before the distortion becomes canon.
And I want to talk about what I am referring to as The Hard Work.
I chose that term intentionally because, at this point in my life, I want to know that I can still do hard things, and this inflection point we are at with AI certainly feels like it qualifies. While the name may sound intimidating, the practices involved helped me reveal just where I was coming down on the spectrum between, “Oh, I got this; I own AI, it doesn’t own me,” and “Okay, I’m struggling to distinguish myself in the blurry line between me and the machine.” And if you are not sure where you land, it will help you do the same, finding your path to staying in the loop and in the lead.
If any of that resonates, you are in the right place.
So let’s see what it takes to do The Hard Work in the age of AI, shall we?
📡 The Signal
Four Years to Build a Brain
In London, UK, a person who wants to drive a legendary black taxi cab first has to pass a test aptly named, “The Knowledge,” one of the hardest licensing exams in the world. To pass this rigorous exam, applicants must memorize around 25,000 streets, 20,000 landmarks, and more than 320 specific routes within the Greater London area—as well as be able to demonstrate command of the route network and its elements.
The process leading up to taking the exam takes 3-4 years at minimum.
And London is not like Manhattan and its grid, or Paris and its spiral. The streets of London are windy, cobbled together, and not laid out so much as accumulated. Picture it this way: If Manhattan is a spreadsheet, and Paris is a snail shell, then London is a drawer full of old charging cords: tangled, specific, weirdly functional, and absolutely hostile to anyone who thinks they can understand it at a glance.
It’s a lot, to be sure.
Indeed, only about half of those who attempt ‘The Knowledge’ exam actually pass, so doing so is a significant accomplishment.
‘The Knowledge’ is famous in cognitive science circles, as well, because it has offered researchers an unusually vivid case of intensive, adult learning in the real world, and a behind-the-scenes look at how our brains work when they work hard, or don’t.
In a landmark 2000 study, neuroscientist Eleanor Maguire scanned 79 trainees over four years and found the brains of the ones who passed the intense Knowledge of London exam had physically grown larger, and those drivers scored significantly higher on memory tests than the ones who failed the exam—even though everyone had started out roughly equal.
Her research showed that cognitive exercise produces physical changes in the brain.
The Moment It Stopped Working
Additionally, a separate 2017 study by another team at University College London, led by Hugo Spiers, scanned the brains of 24 people while they were navigating in a simulation of driving on London streets—once from memory, and then once while following a GPS device.
What the researchers found is that participants’ brains, specifically the hippocampus (a part of the brain that helps you remember places and navigate through them) and the prefrontal cortex (which helps you think ahead and make choices), lit up at every junction along the route when they were navigating from memory. However, when the participants were following a GPS, those same regions went dark.
In other words, the brain didn’t get help from the guidance device—it wasn’t needed to make decisions, so it simply stopped participating.
That story is not about GPS, or the patchwork of London’s beautiful but complicated city streets.
It’s about what happens any time something gets handed over completely—a route through London, a first draft, a judgment call… The faculty that always made the call or execution stops firing once nothing requires it to.
Nothing to worry about, I suppose, until the day it’s needed and you find it’s out of practice, or worse, out of commission.
We’ve all heard the “use it or lose it” warning. A human’s brain is one of those muscles that needs regular cognitive exercise, in lots of different forms, to stay healthy, agile, and strong. And science is backing this up.
Here’s the takeaway: Long, demanding, real-world learning and deliberate thought exercises can alter how the adult brain is organized and how a person navigates through the world, remembers what it experiences, and solves problems it encounters.
And removing that work shuts those parts of the brain down.
The Warning That Waited Forty-Three Years
Now, here’s what complicates that tidy little lesson: the opposite instinct—be in all of the work then, to keep the brain active—fails too, just more slowly. And the evidence is older than any headline on AI.
While perusing Substack on this topic, I came across an essay entitled, “AI Brain Fry, Workslop and the Ironies of Automation” by Carl Hendrick. In it he talks about a historical paper he describes as reading more like a prophecy.
In 1983, an engineering psychologist named Lisanne Bainbridge wrote a paper called “Ironies of Automation,” arguing that the more capable an automated system becomes, the more demanding (not less) the human role inside it gets, because whatever its designer couldn’t figure out how to automate still lands on whoever’s left holding it.
Forty-three years later, a Boston Consulting Group survey of 1,488 workers proved her right in a way she couldn’t have measured in 1983: the single most mentally taxing way to work alongside AI turned out to be overseeing it, reviewing everything it produces as though vigilance itself were the responsible move.
Turns out, those workers made 39% more major errors than the ones who didn’t pursue rigorous oversight.
Which brings us back to the point of this letter, and the actual shape of the problem.
Allow the faculty to go dark, and the capacity erodes before you notice it’s gone. Keep your cognition lit up on everything, and the exhaustion produces the same erosion from the other direction.
Neither one is control. Both only feel like it, and only for a while.
🪨 First Principles
The Middle Ground Has a Name
So, there’s a version of staying in the loop that costs nothing: a glance at an output before it ships, a name attached to a decision nobody actually made. It reflects the form of oversight without the weight of it.
And a version of human oversight that costs greatly: vigilantly reviewing every output, keeping inside every system and process, which honestly, besides being exhausting, defeats the purpose of systems and processes.
The Hard Work I referred to at the top of this letter is, in my estimation, the middle ground where we keep the systems efficiencies and our agile faculties.
I see it as the effortful, countercultural choice to keep doing the parts of the thinking that are supposed to be hard, in a moment when a thousand tools are offering to make them easy. It’s the discipline of keeping your judgment—the questioning, the discernment, tacit knowledge, your curiosity, and origination—as well as accountability for the whole process and whatever you do choose to relinquish, alive in yourself and on purpose. And it’s the faith in the systems, when built with purpose and consideration that truly augment the work you are doing, while allowing your faculties a genuine break from the inevitable crush of scale.
Those faculties only get built—and maintained—by doing the work that implements them yourself. And only handing off what you define as veritably repetitive, and therefore, AI’s domain.
The reality is that the effort was never incidental to the outcome. It was how the person capable of standing behind that outcome, and behind the tools that helped produce it, got made.
Differentiating, then, is the whole discipline.
Some decisions can be handed off entirely. Some can be assisted. Some need to stay, deliberately, in human hands, because producing the answer yourself is the only way the capacity to judge that answer, and to keep control of both the tool and the person using it, stays alive in you.
And only you can decide and distinguish which is which.
It’s important to recognize, however, that being the human in that loop was never going to be one thing you do once and call finished.
The Hard Work is in opposition to all-encompassing adoption of the tools out of the seductive pull towards speed, in favor of taking a slower, more earnest approach that preserves our important faculties. It’s a personal and cultural ethic of practice: what a person must continue doing so that “human oversight” remains real rather than ceremonial.
The slower, effortful posture is not inherently virtuous simply because it is slow, though.
It matters because it keeps a person actively engaged in the activities through which judgment, learning, practical wisdom, authorship, and responsibility are developed and exercised. The risk is not that AI “makes us less human” in some literal or essentialist sense. Rather, the risk is that frictionless delegation can remove us from the practices that cultivate distinctly human capacities.
It always was, so it is not an anti-AI position I am recommending here.
Sometimes slower adoption is not failure to innovate; it can be the work required to discover whether a proposed use should exist in the first place—and how, who might be harmed, in what ways, and who will answer when it goes wrong.
And it seems large organizations are already implementing AI frameworks for risk management and operational specifications, for just such reasons.
These frameworks are often strongest in high-stakes settings—medicine, hiring, credit, public benefits, criminal justice, safety-critical systems. But I want to make the important cultural extension: even low-stakes generative AI use can shape the habits people bring to high-stakes judgment. Because a person does not develop discernment at the moment of crisis or challenge. By then they either have it or they don’t.
That kind of perspective was developed over time, through ordinary, repeated practices and ritualistic behaviors.
I have started getting the feeling, from different conversations and sources, that my idea of moderate, considered integration and taking our work with AI a lot slower—earnestly, as well as with more intention—is an unpopular one.
The hype around the use of AI and its immense potential is that it makes everything faster so why slow down and “get left behind“? ...While that’s a very simplified version of the ‘use the tech to get more efficient’ argument, I think it’s a terribly short-sighted approach to a technology we are already finding solid reasons to pull back from.
Add to that how AI systems have become so pervasive in our societal landscape so quickly (especially considering it is still so early in its own evolution), and you find that many of us already find it incredibly difficult to do anything without it being involved.
Consider this, against the story from the UCL research above: how often do you use GPS to get where you need to go now, versus how often do you allow your memory and reasoning to serve and guide your way there? Also, when you use the GPS, how much of that drive do you actually recall afterwards? …Remember, when the brain is not “needed,” it essentially shuts down.
Now lay that against the entirety of higher faculties your brain is meant to have command of.
That’s what turning over too much to AI is really affecting, and the science and research is beginning to show it. This realization is making it even more important to take back as much of the brain work as possible, in parallel to still using the technology effectively, while we still know how to do those vital parts and it isn’t completely trained out of us through the adoption of the tech and all the wondrous and concerning things it can do.
In the next section I’ll take you through my journey of adopting these tools, and where I’ve landed as a result.
Before I do, however, let’s look at my current thesis on the idea: use the tools for the repetitive tasks where your time would genuinely be better served, and stay in the process for the things that protect your creative, strategic, and evaluative irreplaceable work.
To help me with this, I have created some frameworks. I have a good one for you next week in The Event Horizon #007, so stay tuned. This week, I have an assessment to help you see where you are in this conversation.
The Hard Work Reflection Assessment
I want to share the 5-question reflection pack I created that assesses the five postures of what should be differentiated between to determine what needs you and what can (and should) be handled by the tech.
The five postures are questioning, discernment, accountability, curiosity, and origination. Additionally, there is a sixth bonus question that points to the umbrella judgment over it all: deliberate restraint.
Questioning — Do I know what I think, what I’m asking, or what matters to me on this topic before I ask the machine?
Discernment — Can I tell what’s solid here, what’s missing, and where this answer might fall short?
Accountability — Am I prepared to stand behind the result, explain it, change it, and fix it if necessary?
Curiosity — Does this answer make me want to look further, or has it made me stop thinking too soon?
Origination — What am I adding that the machine could not have produced without me—my experience, knowledge of the situation, values, judgment, or genuine point of view?
Then add the sixth, with the previous decisions in mind:
Deliberate restraint — Can I resist and choose not to take the fastest answer when the slower work of figuring it out for myself is the skill I’m building and the answer is the part I need to learn?
Together, they are the heart of “The Hard Work.”
The point of the exercise is not to find yourself in a position of anti-AI or anti-efficiency. It is to create an ‘anti-default delegation’ posture in domains where you can’t separate the outcome from the work it takes to get there, because they are constitutive of the person capable of producing and standing behind the result.
These five postures are vital to protect and maintain, as they are the capacities through which humans become responsible authors and operators: people who can form a view, give reasons for it, recognize uncertainty, learn from consequences, revise themselves, and answer to others.
None of this started as theory.
It came from watching myself become exactly the kind of AI user I swore I’d never be—a person losing ground she didn’t notice she was giving up—and then having to find my way back because I’d failed at every one of those guardrail questions. Then, in recognizing it, slowly figuring out how to take the deliberate differentiation decision back.
That’s the part worth telling next.
🫀 The Human Layer
My Journey Through The AI Land(Hell)scape
Early on, as I started down this path with generative AI tools, I realized that this technology only confirms a stance I’ve had about how we determine when something is quality, significant, and worth our attention. If something created is bad, not poorly produced, but simply crap that doesn’t move the human condition needle in any meaningful direction, we just know it.
It’s something visceral we experience, and it’s not always something we can verbally describe.
But, to borrow a line from one of my favorite writers, Aaron Sorkin, “It’s like porn. We know it when we see it.”
Early in 2023, as generative AI was making its grand entrance, I was just beginning to build my current business.
I was staunchly on the side of not applying the technology to what I viewed as my own work. Like many others, as a Creative, an artist, and a writer, at the time I saw AI as a threat to creative pursuits, and maybe even creativity itself. …I vividly remember having a conversation with my sister about whether I’d employ any of the new technology for my work. I was staunchly a “no!” …I had decided that I would never use “those tools” for the work that means so much to me.
It’s mine, and is how I identify and show up in the world.
And also… I’ve always been interested in technology.
I was certainly curious about what was developing in the AI arena, and what it might all mean.
By early 2024, it seemed obvious that the genie was out of the bottle, and there was no going back.
I started realizing that if I wanted my business (or any business I chose to build next) to have a fighting chance, any sort of viability, I needed to be embracing AI and figure how I would use it, or offer it to clients, or teach it to my audience, in ways that felt authentic and resonant—not just forced by hype and urgency.
So I started putting myself in rooms where people were actively talking about, learning about, working with, and developing all the generative AI tools coming to the forefront.
At that point, I dove in head first… And in February of 2026, right after Claude Cowork came out, I made the pivot to begin systematically handing more and more over to AI, instead of simply using a scattered, cobbled-together set of tools for the gamut of things I do.
It started as a thought exercise of sorts, a personal experiment to begin to surrender control (writers and artists will recognize this), hand the tech my workflows, and in the process help me define what this meant for me specifically. It was important to me to understand my thresholds before I started advising others on what “good” looks like, and what I found the pitfalls to be as they reveal themselves to me.
Boy howdy, did I find the pitfalls!
Pursuing this path of AI has both broadened my perception of what’s possible—and even who and what I am—and helped me to recognize that fundamental truths of being human are precisely that for a reason, no matter the era, arena, or age.
Most recently, I have come to realize that the answer is definitely going to be different for every individual (let alone organization) regarding how much you use the tools, for what, how long, and at what cost (the very least of which is monetary).
As for me, the first half of this year—while I was helping to take care of my last remaining elder relative, committed to a couple of initiatives meant to help move my business move forward, and working on a timeline of milestones I had set in motion before the new year—I found that I was both super overwhelmed and exhausted.
As a result, I found myself leaning on my AI assistants and systems more and more to get things done and off my plate.
Something I started to see in myself was the fact that as I did, two things became very apparent… I didn’t like the work I was producing nearly as much, and I also didn’t enjoy the process of doing the work at the same level, either. For someone who identifies as a writer, that was a real gut punch.
Now obviously, taking the circumstances into consideration, I was tired and stressed, and it could have been me seeing that in myself and what was coming out of me.
And that was certainly a factor; I don’t pretend otherwise.
But I’ve been paying attention, almost from the jump, because the aspect of how using these tools actually changes and affects a person has been one of the things I’ve been most curious about from the beginning.
When I thought I was recognizing changes in the way I create and think, I started deliberately taking things back from the technology till I could sort it out.
Now, I write the majority of my pieces myself again, using the tools for research, think-through exercises, outlines, as well as a review after I’ve nailed down the issue, that kind of thing.
I have also taken other steps to ensure I maintain my craft, my creativity, and my capacities. I will go deeper into those practices in the next issue of The Event Horizon.
My single biggest take away so far, and what I keep coming back to from almost every angle, and in so many conversations I’m having, is this: right now—not in a year from now, not 6 months from now, not even 3 months from now—is when people need to begin deciding in earnest how much they stay in the loop when using these tools, and at what stages of the process.
Everything I’ve been experiencing keeps coming back to the judgment card.
No matter the tool, task, or tolerance level I am working through, how I show up and actively insert my mind into the equation—my thoughts, my insights, my discernment, my values, my bullshit detector, my curiosity, my intuition—and whether or not I use my brain to add to the process rather or surrender it to the process, determines the quality of output I get.
And every single time I insert my brain into the process, I get infinitely better output.
One more element I am beginning to recognize makes a difference in this conversation… Where you are starting when you begin using the tools.
I had been building my niche and writing around my topics for a while before I inserted AI into my work. So, I had a library of content to train my AI systems on, as well as what good looks like for me. Without those elements, I don’t know that I’d personally been comfortable enough to dive in head first, nor would I have had the discernment to know as quickly as I did when I had relinquished too much or seen how the work was beginning to fail.
I believe that people who are just beginning to build a repository of work, while simultaneously adopting the tools to assist them, will likely be the ones most susceptible to the homogenization the AI models outputs lean toward.
So, here’s where I have landed so far: for all of the “ease” promised by AI, the necessary hard parts before you jump to the ease is what makes the difference. It’s what makes the output worth energy or attention.
The “hard” is what makes it good. And it always has been. AI has simply shortened the curve of recognizing it.
The detrimental part is not the tool.
The detrimental part is using the tool before building the inner reference point—the five postures that keep your higher faculties in play—that the tool is supposed to serve.
🌅 Light on the Horizon
The Good Version, Already Happening
Assistive AI has moved well past the hype cycle this year.
OrCam MyEye, dot’s Lumen glasses, and Microsoft’s Seeing AI are now translating real-world scenes into real-time audio for blind and low-vision users, joined at CES 2026 by self-driving mobility chairs and neural earbuds built for the same purpose.
This is AI doing something a human cannot do unaided: reading a room, a street, a face, in real time, so a person can move through the world with more agency. It’s the differentiation argument working exactly as intended: the tool doing what only the tool can do, the human staying exactly where judgment is needed.
Insilico Medicine’s Rentosertib just became the first AI-discovered drug to reach Phase III trials.
It’s entering a 320-patient study across China to treat idiopathic pulmonary fibrosis, a progressive, scarring lung disease. The molecule came out of Insilico’s own Pharma.AI platform: PandaOmics for identifying biological targets,
Chemistry42 for generating and screening candidate compounds, doing the genuinely mundane-at-scale work of examining thousands of molecular possibilities so human researchers could spend their judgment on what only they can do: designing the trial, caring for the patients enrolled in it, and interpreting what the results actually mean for someone living with this disease.
If it holds up through Phase III, it becomes a treatment reaching people who are currently waiting on one.
That’s the differentiation argument, playing out at institutional scale.
⚠️ Concern Worth Naming:
The Cost Hidden Inside Convenience
Something worth watching:
MIT Media Lab researchers gave this pattern a name: cognitive debt.
Their study of writers using ChatGPT found progressively weaker brain connectivity the more the tool did the thinking, and when those same writers tried working without it afterward, their neural engagement stayed suppressed.
The debt doesn’t announce itself. It accumulates in the background, and every small convenience delegation makes the next one easier to justify, while the underlying capacity gets a little harder to reach when you actually need it.
The good version of this looks like:
Resembling the two items above, the good version of this looks like AI doing what only AI can do, or what’s genuinely mundane at scale, while judgment stays exactly where it belongs.
What separates the two is whether the decision to hand something off was actually a decision, made on purpose, rather than just the path of least resistance.
That differentiation is the loop. Stay in it.
🎯 Your Move
This week, notice the moment right before you hand something off.
Pick one task you’d normally let the tool take without a second thought, and pause there instead.
What would you actually build if you stayed and did this one yourself?
Then choose, on purpose, knowing exactly what you’re trading and why. That’s the whole practice: staying the one who decides, every single time.
📡 Signal Back
This week’s question:
Is there something you swore you’d never hand over to AI that you’ve since handed over completely?
What changed, and was it actually a decision, or did it just happen to you?
Help me see how this is affecting other founders. I read every reply.
If this issue resonated with you, I have four small asks:
① Hit the ❤️—It takes one second and tells Substack this conversation is worth having.
② Hit the 🔄 restack—It puts this in front of your followers—the ones who are already building differently and don’t yet know there’s a name for what they’re doing.
③ Share this issue with a founder you know who could benefit, and might even think you’re their hero. 🦸
④ Drop a comment—I read every one. And I reply.
Some of my best thinking happens in response to what you bring to the conversation here—and future issues often start in a comment thread (this issue is proof of that).
💡The right idea finds the right person at the right time. You might be the one who gets it there.
🌿 The Understory
What you just read is the argument. What follows is the work.
What comes next was built for The Understory, and it picks up exactly where the argument left off: from recognizing the drift in theory to finding it in the places you actually work, and deciding on purpose where the line between what you hand off and what you hold belongs.
This week’s Understory is two things. An Integration Frame—the postures from the Hard Work Reflection Assessment carried into four places you’re already working: your content, your voice, your client relationships, and your systems, with the posture that does the most work in each one. And a Research Deep Dive—the annotated source set behind this issue, from Maguire’s taxi drivers and Spiers’ GPS study to Bainbridge’s 1983 paper and this year’s cognitive debt research, for the writers, coaches, and expert-to-educators in The🌳Grove who want the receipts underneath the argument before they carry it anywhere else.
And at the close—an honest note on how I used AI this week. This is a standing section in every The Event Horizon Understory, because my position on AI is simple and steady, and now refined: use the tools, stay in the loop, differentiate the threshold and deliberately restrain when needed—and stay in the lead. These are not in conflict. The Understory is built on exactly that premise.
Let’s get to work.
For everyone reading from The Canopy: this is what The Understory looks like. Join us in the deeper work.
✨ Here’s to the hard work of noticing before the drift becomes the default.
Stay curious. Stay human.
~StacyLynn
Founder, Sitting Pretty Strategies
Root Deeply. 🌳 Rise Differently.
P.S.
One more thing: I mentioned above that I used frameworks to help me protect my craft once I noticed the drift. These have become real practices I use regularly, alongside the recognition I gained from the assessment I shared in this week’s issue.
In next week’s letter, I’ll share the arc and the practice framework that has helped me come back to my work and creative identity.
And if you know a founder who’s caught themselves breaking their own “I’d never” rule about AI, and can’t quite pinpoint when it happened, forward this their way. 💌
I really liked your story and how you took us through your ebbs and flows with AI. This sentence resonated a lot: "But I’ve been paying attention, almost from the jump, because the aspect of how using these tools actually changes and affects a person has been one of the things I’ve been most curious about from the beginning." I found I have been going through a similar experience this year!