When “Human-In-The-Loop” Becomes Ceremony
The oversight designed to keep humans leading and in control of the technology is becoming the place where human judgement steadily acquiesces.
The conversation about AI has gotten very loud. So far, in that conversation aspects about the human experience in the AI era have gotten a bit lost.
The Event Horizon is a Thursday 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 to the inaugural issue of The Event Horizon! 🎉
I started this Rooted & Rising edition because the conversations around AI have gotten very loud, and most of it is noise and not signal to be zeroing in on to stay in the lead in this era. …Tools, benchmarks, capability races, workflow hacks—the noise is real. The Event Horizon aims to raise the level of conversation, and address a different part not being had enough: what all of this actually means for the human underneath the productivity gains.
I want to ask different questions to bring better, more consequential, conversations to Stewards in The🌳Grove: What does it mean for our thinking, our judgment, our sense of agency, our identity, our capacity to stay in the lead?
This first week, we’re starting with addressing the safeguard that was supposed to keep humans in control of the technology—and what’s becoming of it.
📡 The Signal
“Human-In-The-Loop” (HITL) began as a design idea. It was the mechanism in interactive computing and human-computer interactions that selectively included human participation rather than fully removing humans from the process and reverting to automated systems by default.
It’s kind of like autocorrect on your phone.
Your phone is perfectly capable of making suggestions. Sometimes it fixes a typo before you even notice it. It’s lovely and helpful, with tiny digital butler energy.
But other times, your phone is very confident. Aggressively so, even. It will suggest the word it thinks you meant, swap out your spelling, and occasionally turn a normal sentence into something that makes you question the entire future of technology.
That’s where the human comes back in; you’re still in the loop.
You don’t need to personally type every letter with a quill and candlelight. But you do need to glance at the suggestion and decide whether the machine helped… Or was wildly wrong, and only one unfortunate tap away from sending your mother a message you’ll never emotionally recover from.
That’s the basic idea of human-in-the-loop.
The system can assist. It can predict. It can speed things up.
But the human still stays involved where judgment, context, and “please don’t send that” matter.
There’s long been an idea in technology that humans should supervise, correct, and guide automation in situations where judgement and discretion are a vital part of the process.
Autocorrect is one little loop.
The system suggests a word. You accept it, reject it, or mutter something unprintable and fix it yourself. Generally, this is the camp I end up in.
But imagine if accepting that one autocorrect didn’t just change the sentence.
Imagine it also updated your calendar, sent a follow-up email, made a list of your groceries, changed the status of an important project, and notified three people in Slack on your current progress.
Now the human-in-the-loop role is no longer “check the suggestion before it goes out.”
It becomes “understand the cascade before it starts.”
And that is where the modern HITL conversation gets much more complicated.
As agentic AI becomes pervasive in organizational processes, its development has made HITL a volume operation for humans. The paradigm is no longer 1:1—one AI output, one human evaluation. In reality, it’s not one task being automated. It’s one action triggering another, then another, across the system.
It’s a domino effect, but with AI making the moves.
And in that environment, HITL becomes something different from what it was originally designed to be.
In researching the topic this week, I came across State of AI Trust in 2026: Shifting to the Agentic Era, a report on the findings of the McKinsey 2026 AI Trust Maturity Survey [see the original source here ->]. One line stood out, summarizing the HITL dilemma at hand:
“If a human receives 200 requests a day, they will stop looking at the evidence.”
The human in the loop was supposed to be the safeguard. The research is beginning to reveal it to be the very place the loop is most likely to collapse.
🪨 First Principles
Staying in the loop is not the same as leading it
The Sitting Pretty Strategies (SPS)’s position—AI for the mundane and repetitive, human energy for the creative, strategic, evaluative, irreplaceable—was always implicit in what HITL was supposed to be. AI handles the work; human stays in the judgment role. That’s the design.
What this week’s signal reveals is that when oversight becomes the job—and it’s 200 approvals a day—the oversight itself becomes the mundane and repetitive. Human energy ends up in exactly the place SPS believes it shouldn’t: processing volume, not exercising judgment.
This gives way to the idea that there are two versions of “staying in the loop.“
Genuine discernment—evaluating, questioning, intervening. And procedural compliance—approving, moving on, approving again. The first protects the human operating system. The second doesn’t just fail to protect it—it depletes it.
For founders, this means something specific.
While large organizations lose institutional judgment capacity when this happens, something more damaging can happen for the solo founder. For a solopreneur or small team, this kind of oversight means losing something harder to rebuild—the discernment and reasoning that distinguishes their work.
There’s no institutional backup. No extra layer overlapping individual loss of judgement capacity. For the small business founder, the operator and the business are one.
What atrophies in the person disappears from the business.
My belief, as a solo founder, is not only “offload the mundane, keep the irreplaceable.“ It’s also “protect the capacity for the irreplaceable by exercising it—even when the tool could do it faster.“
And I am on this journey with you.
Discovering the amazing things AI can do for my business. Then discovering the consequences of allowing it to do everything it has capacity to do. And arriving now at the part of the journey that was always the work: keeping what’s inherently and creatively me strong and clear—my cognitive sovereignty, discernment, and independent thought—and delegating only what helps me support and maintain my energy in order to be able to scale my impact.
Which brings us inside the loop itself—to what’s currently happening to a founder’s judgment, reasoning, and sense of authorship when oversight becomes ceremony.
🫀 The Human Layer
Are you forming thought, or receiving it?
This isn’t the first time the conversation on cognitive sovereignty has been an important discussion. Throughout the history of technological advancement, there has always been a component of concern on the cognitive sovereignty layer of being human.
Cognitive sovereignty is the idea that people should retain control over their own thinking—their attention, judgement, memory, and decisions—and avoid environments that would steal and steer it manipulatively or coercively. Staying the author of your own reasoning.
There’s a gap between feeling like you’ve reasoned something through and actually being its author—and with the rise of AI, most of us have stopped noticing the difference.
The biggest concern in this area is that AI, recommendation systems, and data-driven platforms can shape what we notice, believe, and choose—before we realize it, or even without us being aware at all.
It’s not just in getting a bad or wrong answer, and not challenging it.
The worse scenario is the loss of the habit of even checking and verifying the output—or questioning and reasoning it out for ourselves, no matter the quality of the output.
It’s very much a “use it or lose it” kind of capacity.
Cognitive sovereignty asks, “Am I still genuinely forming this thought, or is something else doing too much of the thinking for me?“ And I think we all should be asking that question more often, as well.
And I am, more and more. I’ve also begun taking back certain aspects I originally handed over to AI joyfully as I learned what it could do. My goal is to have a better balance as I move forward.
You see, I have simultaneously a hopeful and realistic view of what AI can do for us, as well as to us, in the coming years. And at the same time, I also hold a dichotomy of views on human potential and what it means to be human.
I don’t think that just because you get your information from an AI source that inherently means you have given up your ability to think for yourself. In fact, I would argue that most thought is built by other thought.
All research is, by its very nature, outsourced to external sources. All thinking is shaped by context and the world around us. This is where the crucial skills of interpretation, reflection, discernment, and critical thinking come into play.
So, I don’t think this is about becoming isolated from influence, AI or otherwise. What makes something our own thought is the reflection and the refinement and the time we take to actively think through the information we have taken in, and to then form our own output insight about it.
I also don’t think the answer to maintaining our cognitive sovereignty is to stop using AI. Using the tools is like using any other tools that have changed the landscape before. It starts with a strong mind, and strong character, and a strong intention of boundaries.
And those boundaries begin with critical thinking.
It becomes—history has shown time and again—one of the main skills critical for humans to not only develop but also ritualize and maintain. It is one of the main tools that makes cognitive sovereignty possible. Critical thinking helps you evaluate claims or spot gaps and test assumptions for yourself, while cognitive sovereignty is the broader condition of still being able and in charge of your own attention, your judgments, and your authorship of thought while you do your critical thinking.
Critical thinking is the active skill of questioning and comparing ideas. It’s a practice. It asks, “Is this true?“
Cognitive sovereignty is the right and capacity to keep that questioning genuinely your own. It’s the state of not surrendering the practice to automation or persuasion or habit. It asks, “Am I still the one deciding how to think about this?“
In this article on critical thinking in the age of AI [see the source material here ->], author Sabryna Alsfasser reasoned that valuing speed over cognition and judgment is the root condition of decline in the HITL process. And this moment has elevated critical thinking from a soft skill into something closer to the core infrastructure of the human operating system.
I don’t think this all should overwhelm or dismay us.
Rather, the smart play for conscious founders is to earnestly categorize the autonomy (what you keep as human as possible) and the automation (what you delegate through a human lens). The founder who names this, who protects the internal synthesis layer, who practices deliberate discernment even when the tool could move faster—you’re not being anti-AI.
You’re doing exactly what the AI era actually requires: Operating from a stance of human lead, and AI empowered.
🌅 Light on the Horizon
When humans stay the author of the impulse
The Creativity Paradox
Researchers studying AI’s effects on creative output found something worth pausing on: AI enhanced human creativity—but only when the human initiated the idea first.
When participants let AI generate the opening concept and worked from there, their output was measurably less original than when they brought their own seed idea and used AI to develop it further.
The tool did more when the human started. It did less when the human deferred.
For founders, this means the initiation layer isn’t the part to hand over—even when the tool can handle it. The creativity lives in the start. The augmentation lives in what follows. Human-generated → AI-developed → human-refined is a different cognitive loop than AI-generated → human-approved. The first keeps authorship intact. The second is the ceremony problem in miniature.
[Read the research →]
The Diagnostic Horizon
On a different front: AI is compressing the rare disease diagnostic odyssey in a way that shows exactly what human-led, AI-amplified looks like when it’s working as intended.
Families living with rare diseases often wait seven years or more for an accurate diagnosis. AI systems parsing genetic markers, symptom clusters, and research literature are shortening that timeline dramatically—surfacing pattern connections that human specialists, regardless of expertise, couldn’t hold in working memory simultaneously.
Physicians remain the diagnostic authority. AI extends the reach of that authority.
Nobody removed the human from the loop. They gave the human a tool that made the loop worth being in.
That’s the version we’re building toward.
[Read more →]
⚠️ Concern Worth Naming
Something worth watching:
You may feel like you’ve reasoned something through independently—and still not have been the author of the conclusion. This is the subtler risk. When AI wraps its output in language that sounds finished, the illusion of independent thought is easy to miss.
What looks like clarity can be compression. What looks like insight can simply be plausibility. The reasoning feels like yours because it reads like yours.
The good version of this looks like:
A founder who pauses before accepting any AI-synthesized conclusion and asks: Did I think this through, or did I receive it? The habit of that question—practiced regularly, even when the output looks right—is what keeps cognitive sovereignty from becoming a concept you agree with but never embody.
That questioning is the loop. Stay in it.
📡 Signal Back
This week’s question:
Where are you in this? Are there places in your work where you’ve noticed your oversight becoming a formality—where you’ve stopped genuinely looking, even when you’re technically still in the loop?
Reply and help me see the whole picture.
One honest line is enough. I read every reply.
🎯 Your Move
Where in your work have you assigned yourself to “oversee” something you’ve actually stopped genuinely looking at?
Volume has a way of making deep looking feel impossible. Name one place where your human judgment has become a procedural signature.
Then: this week, review that one thing as if it were the first time.
Ask a question about it. Disagree with it, even briefly. Play devil’s advocate, or the steelman. …The act of genuinely evaluating is the practice—and the practice is what keeps the capacity alive.
🌀 The Ecosystem Pulse
Before you go—one honest question: Did this issue nourish something?
Reply with a single letter and let me know how this one feels for you:
A—Yes, I feel more aligned
B—I’m still sitting with it
C—This one didn’t connect
Every signal strengthens the ecosystem. Thank you for being part of it.
🧭 Constellation Compass
🌌 Elsewhere in the SPS ecosystem this week:
🗺️ In this issue of Elegant Email Ecosystems, I dove deep into what AI surfaces about your business before any human decides to trust you, the four signals it scans for, and how to build for the evaluation you didn’t know was happening. [Read E3 #027 →]
📍 In this issue of Pretty Strategic, I made the structural case for coherence as an AI-era indexing requirement—and why the founders who built foundation-first are the only ones a recommendation engine can surface with confidence. [Read PS #039 →]
🔮 Inbox Alchemy is the lab side of SPS—where ideas like these get turned into practice through a sequential, evergreen email experience—the hands-on companion to everything SPS publishes, one issue at a time. If you’re ready to move from insight to implementation, [Join us in the lab →]
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.
💡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 goes where this issue has been pointing all along: into the practice of staying genuinely in the loop.
This week’s Understory is two things. A Practice Protocol—one concrete, daily practice for founders who want to keep their judgment engaged rather than let it drift into a procedural signature. And a Research Deep Dive: the fully annotated source set behind this issue’s argument, including the Wharton “cognitive surrender” study, the HBR Brain Fry findings, and three additional Tier 1 and Tier 2 sources—built for the educators, coaches, and expert-to-educators who’ll carry these ideas into their own work.
And at the close—a brief, honest note on how I used AI to write this issue. This is a standing section in every The Event Horizon Understory, because my position on AI is simple and steady: use the tools, stay in the loop—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.
