Your Company Keeps Losing Its Mind
Otto has Alzheimer's, so he carries a notebook everywhere. For a long time, philosophers have said that notebook doesn't just help his memory, it is his memory. Give a company the same notebook, and the metaphor gets tempting. But does it work?
AI
Culture
A “mind” isn't confined to a brain. Couple it tightly enough to a tool and the tool becomes part of the cognitive system.
![](data:image/svg+xml,<svg display="block" id="565267359" role="presentation" viewBox="0 0 3000 1376" xmlns="http://www.w3.org/2000/svg"/>)
THE NOTEBOOK
Otto has Alzheimer's, so he carries a notebook everywhere. Addresses, appointments, anything he needs — he writes it down, and when he needs it he looks it up rather than trying to recall it.
Clark and Chalmers have famously pointed out that the notebook is doing exactly the job memory does for everyone else: it's reliable, always at hand, and Otto trusts it without a second thought. Their Extended Mind thesis says you'd be wrong to draw the line around Otto's mind at his skull.
A “mind” isn't confined to a brain. Couple it tightly enough to a tool and the tool becomes part of the cognitive system, no different in kind from what's inside your head. A notebook, a calculator, an AI chatbot — tools. But Otto's notebook is, functionally, also Otto's memory.
But what about Otto Inc? Give a company its own notebook — a wiki, a shared drive, a knowledge base — and these extend the employees' minds with every use. At least rhetorically, call the whole phenomenon a company-sized Mind, extended into its own infrastructure.
Just because the infrastructure exists, it doesn't mean that the extension happens. The architecture is static.
THE EXTENSION ACT
Let's say Otto lost his notebook and I found it. I can read it just fine, and I understand its purpose. But I am reading it, just as Otto used to, and yet it's not extending my mind. My memory works fine unextended, and even if it didn't, I'd have no use for Otto's home address.
What happens when we scale this logic up for a company? The company can't read its own notebook. It can't really read anything. And even if it could read — say, via an LLM — it would do so the way I'm reading Otto's notebook. Just as there's no purpose for my mind to extend into Otto's notebook, there's no Mind for a company to extend into its infrastructure — because a company doesn't have one. It was only a rhetorical device, remember?
I, however, find myself rooting for the concept. There is a way in which a company's sense of purpose could be materialized via its scaffolds, processes, memories. It happens all the time, come to think of it! But it happens granularly, with each instance of cognitive coupling under the company's umbrella. This umbrella points to some goal. There is a success condition.
And while the company does not have a Mind, per se, it does employ a number of people who (ideally) do. So it's their minds extending into company infrastructure to satisfy the goal.
Say Otto has suddenly lost his sense of vision, and started paying me to go through his notebook and retrieve important dates. My goal is to do my job successfully (and get paid), and my mind achieves this goal (necessarily) via extension to the notebook.
But what if Otto, instead, hired 30 people, me included, to satisfy a goal he never got to specify? He suddenly lost his ability to communicate, in a characteristic way. I could try guessing what his goal was, and in doing so I might look through his notebook. But then the notebook would extend my mind toward my goal of finding his goal — not the goal he actually hired me for. Maybe Otto wanted us all to guard his notebook so no one would read it.
This is a (rather convoluted) way to exemplify that:
(1) cognitive extension happens only under a goal, and
(2) minds (when paid) can extend to achieve goals they did not originate to the degree at which the goals were specified;
The Mind exists as long as it's instantiated. What instantiates it is the goal.
WHERE IT BREAKS
(1) Otto pays me to grep his notebook because he can no longer see.
(2) Your company employs me to complete tasks. To be online. To attend meetings.
And I do so very well. The way my mind extends into company infrastructure is directed by the goals I interpret as belonging to the company. And so I am completely functional.
But my extension never really instantiates that rhetorical company Mind. Because the goal originating in that Mind would be 5x-ing revenue in the next couple of years. The Mind wouldn't care about my timesheet.
As the company scales, on the road from (1) to (2), the Mind gains cognitive resources and sacrifices alignment.
So where do we find this Mind, then?
THE GOAL
And extend it into company infrastructure to achieve its lofty goal?
The Mind exists whenever — and to the degree at which — employees extend their own towards the Mind's specific goal. Not towards respecting the contract, not towards completing the task, not towards producing, confirming, improving, logging, billing.
It might look like the overlap between these individual goals and the top-level organizational goal is just a function of management. And it might be.
But what is that organizational goal? And did it originate in the company's Mind? Or was it reconstructed, bit by bit, from what surfaced as each mind extended into company infrastructure toward its own goal?
Drift is what happens when nothing keeps the goal legible as it moves. That problem is solvable.
THE SOLUTION
It takes exceptional transparency into the company's systems and processes for any one mind to uncover and align with that goal — the goal any mind could extend toward to instantiate the Mind. That transparency is rare because almost nothing a company builds for itself is built to carry a goal intact across that many hops.
Hierarchical management controls noise and maintains alignment broadly. The goal, as viewed from the single point with the highest access, gets split into dependencies and handed down with whatever context fits into a morning stand-up.
The further it travels, the more it drifts. Sometimes the easy path — obvious to the mind that originated the goal — stays hidden from whoever is tasked with seeing it through. Drift isn't a tax you simply pay for having more than one mind involved. It's what happens by default when nothing keeps the goal legible as it moves. That is a solvable problem.
The degree to which the Mind actually exists is the degree to which everyone involved can instantiate it.
THE DESIGN
Goals originating from a single perspectival point — any one mind — are easily blindsided. And that is by design. Because that seemingly easy path to success could have implied legal risks. Or assumed capacity that simply wasn't there.
The Mind, however, covers a perspectival plane. Or more. Its goals are the ones most likely to originate in a mind with the highest level of access. Fewer surprises, fewer changes of scope, fewer compromises. The actual way forward.
And the degree to which the Mind actually exists is the degree to which everyone involved can instantiate it. None of this argues against reach, or against extending the Mind as far as the company does. It argues against extending it blind.
We've spent the better part of this year building exactly what that instantiation requires. We're not looking to build a bigger notebook, but something that keeps a goal legible as it passes from mind to mind.
If your company is losing its Mind in the way we're describing, we'd love to compare notes.
Featured Case Studies
Featured Case Studies
Your Company Keeps Losing Its Mind
Otto has Alzheimer's, so he carries a notebook everywhere. For a long time, philosophers have said that notebook doesn't just help his memory, it is his memory. Give a company the same notebook, and the metaphor gets tempting. But does it work?
AI
Culture
A “mind” isn't confined to a brain. Couple it tightly enough to a tool and the tool becomes part of the cognitive system.
THE NOTEBOOK
Otto has Alzheimer's, so he carries a notebook everywhere. Addresses, appointments, anything he needs — he writes it down, and when he needs it he looks it up rather than trying to recall it.
Clark and Chalmers have famously pointed out that the notebook is doing exactly the job memory does for everyone else: it's reliable, always at hand, and Otto trusts it without a second thought. Their Extended Mind thesis says you'd be wrong to draw the line around Otto's mind at his skull.
A “mind” isn't confined to a brain. Couple it tightly enough to a tool and the tool becomes part of the cognitive system, no different in kind from what's inside your head. A notebook, a calculator, an AI chatbot — tools. But Otto's notebook is, functionally, also Otto's memory.
But what about Otto Inc? Give a company its own notebook — a wiki, a shared drive, a knowledge base — and these extend the employees' minds with every use. At least rhetorically, call the whole phenomenon a company-sized Mind, extended into its own infrastructure.
Just because the infrastructure exists, it doesn't mean that the extension happens. The architecture is static.
THE EXTENSION ACT
Let's say Otto lost his notebook and I found it. I can read it just fine, and I understand its purpose. But I am reading it, just as Otto used to, and yet it's not extending my mind. My memory works fine unextended, and even if it didn't, I'd have no use for Otto's home address.
What happens when we scale this logic up for a company? The company can't read its own notebook. It can't really read anything. And even if it could read — say, via an LLM — it would do so the way I'm reading Otto's notebook. Just as there's no purpose for my mind to extend into Otto's notebook, there's no Mind for a company to extend into its infrastructure — because a company doesn't have one. It was only a rhetorical device, remember?
I, however, find myself rooting for the concept. There is a way in which a company's sense of purpose could be materialized via its scaffolds, processes, memories. It happens all the time, come to think of it! But it happens granularly, with each instance of cognitive coupling under the company's umbrella. This umbrella points to some goal. There is a success condition.
And while the company does not have a Mind, per se, it does employ a number of people who (ideally) do. So it's their minds extending into company infrastructure to satisfy the goal.
Say Otto has suddenly lost his sense of vision, and started paying me to go through his notebook and retrieve important dates. My goal is to do my job successfully (and get paid), and my mind achieves this goal (necessarily) via extension to the notebook.
But what if Otto, instead, hired 30 people, me included, to satisfy a goal he never got to specify? He suddenly lost his ability to communicate, in a characteristic way. I could try guessing what his goal was, and in doing so I might look through his notebook. But then the notebook would extend my mind toward my goal of finding his goal — not the goal he actually hired me for. Maybe Otto wanted us all to guard his notebook so no one would read it.
This is a (rather convoluted) way to exemplify that:
(1) cognitive extension happens only under a goal, and
(2) minds (when paid) can extend to achieve goals they did not originate to the degree at which the goals were specified;
The Mind exists as long as it's instantiated. What instantiates it is the goal.
WHERE IT BREAKS
(1) Otto pays me to grep his notebook because he can no longer see.
(2) Your company employs me to complete tasks. To be online. To attend meetings.
And I do so very well. The way my mind extends into company infrastructure is directed by the goals I interpret as belonging to the company. And so I am completely functional.
But my extension never really instantiates that rhetorical company Mind. Because the goal originating in that Mind would be 5x-ing revenue in the next couple of years. The Mind wouldn't care about my timesheet.
As the company scales, on the road from (1) to (2), the Mind gains cognitive resources and sacrifices alignment.
So where do we find this Mind, then?
THE GOAL
And extend it into company infrastructure to achieve its lofty goal?
The Mind exists whenever — and to the degree at which — employees extend their own towards the Mind's specific goal. Not towards respecting the contract, not towards completing the task, not towards producing, confirming, improving, logging, billing.
It might look like the overlap between these individual goals and the top-level organizational goal is just a function of management. And it might be.
But what is that organizational goal? And did it originate in the company's Mind? Or was it reconstructed, bit by bit, from what surfaced as each mind extended into company infrastructure toward its own goal?
Drift is what happens when nothing keeps the goal legible as it moves. That problem is solvable.
THE SOLUTION
It takes exceptional transparency into the company's systems and processes for any one mind to uncover and align with that goal — the goal any mind could extend toward to instantiate the Mind. That transparency is rare because almost nothing a company builds for itself is built to carry a goal intact across that many hops.
Hierarchical management controls noise and maintains alignment broadly. The goal, as viewed from the single point with the highest access, gets split into dependencies and handed down with whatever context fits into a morning stand-up.
The further it travels, the more it drifts. Sometimes the easy path — obvious to the mind that originated the goal — stays hidden from whoever is tasked with seeing it through. Drift isn't a tax you simply pay for having more than one mind involved. It's what happens by default when nothing keeps the goal legible as it moves. That is a solvable problem.
The degree to which the Mind actually exists is the degree to which everyone involved can instantiate it.
THE DESIGN
Goals originating from a single perspectival point — any one mind — are easily blindsided. And that is by design. Because that seemingly easy path to success could have implied legal risks. Or assumed capacity that simply wasn't there.
The Mind, however, covers a perspectival plane. Or more. Its goals are the ones most likely to originate in a mind with the highest level of access. Fewer surprises, fewer changes of scope, fewer compromises. The actual way forward.
And the degree to which the Mind actually exists is the degree to which everyone involved can instantiate it. None of this argues against reach, or against extending the Mind as far as the company does. It argues against extending it blind.
We've spent the better part of this year building exactly what that instantiation requires. We're not looking to build a bigger notebook, but something that keeps a goal legible as it passes from mind to mind.
If your company is losing its Mind in the way we're describing, we'd love to compare notes.
Featured Case Studies
Featured Case Studies
Your Company Keeps Losing Its Mind
Otto has Alzheimer's, so he carries a notebook everywhere. For a long time, philosophers have said that notebook doesn't just help his memory, it is his memory. Give a company the same notebook, and the metaphor gets tempting. But does it work?
AI
Culture
A “mind” isn't confined to a brain. Couple it tightly enough to a tool and the tool becomes part of the cognitive system.
THE NOTEBOOK
Otto has Alzheimer's, so he carries a notebook everywhere. Addresses, appointments, anything he needs — he writes it down, and when he needs it he looks it up rather than trying to recall it.
Clark and Chalmers have famously pointed out that the notebook is doing exactly the job memory does for everyone else: it's reliable, always at hand, and Otto trusts it without a second thought. Their Extended Mind thesis says you'd be wrong to draw the line around Otto's mind at his skull.
A “mind” isn't confined to a brain. Couple it tightly enough to a tool and the tool becomes part of the cognitive system, no different in kind from what's inside your head. A notebook, a calculator, an AI chatbot — tools. But Otto's notebook is, functionally, also Otto's memory.
But what about Otto Inc? Give a company its own notebook — a wiki, a shared drive, a knowledge base — and these extend the employees' minds with every use. At least rhetorically, call the whole phenomenon a company-sized Mind, extended into its own infrastructure.
Just because the infrastructure exists, it doesn't mean that the extension happens. The architecture is static.
THE EXTENSION ACT
Let's say Otto lost his notebook and I found it. I can read it just fine, and I understand its purpose. But I am reading it, just as Otto used to, and yet it's not extending my mind. My memory works fine unextended, and even if it didn't, I'd have no use for Otto's home address.
What happens when we scale this logic up for a company? The company can't read its own notebook. It can't really read anything. And even if it could read — say, via an LLM — it would do so the way I'm reading Otto's notebook. Just as there's no purpose for my mind to extend into Otto's notebook, there's no Mind for a company to extend into its infrastructure — because a company doesn't have one. It was only a rhetorical device, remember?
I, however, find myself rooting for the concept. There is a way in which a company's sense of purpose could be materialized via its scaffolds, processes, memories. It happens all the time, come to think of it! But it happens granularly, with each instance of cognitive coupling under the company's umbrella. This umbrella points to some goal. There is a success condition.
And while the company does not have a Mind, per se, it does employ a number of people who (ideally) do. So it's their minds extending into company infrastructure to satisfy the goal.
Say Otto has suddenly lost his sense of vision, and started paying me to go through his notebook and retrieve important dates. My goal is to do my job successfully (and get paid), and my mind achieves this goal (necessarily) via extension to the notebook.
But what if Otto, instead, hired 30 people, me included, to satisfy a goal he never got to specify? He suddenly lost his ability to communicate, in a characteristic way. I could try guessing what his goal was, and in doing so I might look through his notebook. But then the notebook would extend my mind toward my goal of finding his goal — not the goal he actually hired me for. Maybe Otto wanted us all to guard his notebook so no one would read it.
This is a (rather convoluted) way to exemplify that:
(1) cognitive extension happens only under a goal, and
(2) minds (when paid) can extend to achieve goals they did not originate to the degree at which the goals were specified;
The Mind exists as long as it's instantiated. What instantiates it is the goal.
WHERE IT BREAKS
(1) Otto pays me to grep his notebook because he can no longer see.
(2) Your company employs me to complete tasks. To be online. To attend meetings.
And I do so very well. The way my mind extends into company infrastructure is directed by the goals I interpret as belonging to the company. And so I am completely functional.
But my extension never really instantiates that rhetorical company Mind. Because the goal originating in that Mind would be 5x-ing revenue in the next couple of years. The Mind wouldn't care about my timesheet.
As the company scales, on the road from (1) to (2), the Mind gains cognitive resources and sacrifices alignment.
So where do we find this Mind, then?
THE GOAL
And extend it into company infrastructure to achieve its lofty goal?
The Mind exists whenever — and to the degree at which — employees extend their own towards the Mind's specific goal. Not towards respecting the contract, not towards completing the task, not towards producing, confirming, improving, logging, billing.
It might look like the overlap between these individual goals and the top-level organizational goal is just a function of management. And it might be.
But what is that organizational goal? And did it originate in the company's Mind? Or was it reconstructed, bit by bit, from what surfaced as each mind extended into company infrastructure toward its own goal?
Drift is what happens when nothing keeps the goal legible as it moves. That problem is solvable.
THE SOLUTION
It takes exceptional transparency into the company's systems and processes for any one mind to uncover and align with that goal — the goal any mind could extend toward to instantiate the Mind. That transparency is rare because almost nothing a company builds for itself is built to carry a goal intact across that many hops.
Hierarchical management controls noise and maintains alignment broadly. The goal, as viewed from the single point with the highest access, gets split into dependencies and handed down with whatever context fits into a morning stand-up.
The further it travels, the more it drifts. Sometimes the easy path — obvious to the mind that originated the goal — stays hidden from whoever is tasked with seeing it through. Drift isn't a tax you simply pay for having more than one mind involved. It's what happens by default when nothing keeps the goal legible as it moves. That is a solvable problem.
The degree to which the Mind actually exists is the degree to which everyone involved can instantiate it.
THE DESIGN
Goals originating from a single perspectival point — any one mind — are easily blindsided. And that is by design. Because that seemingly easy path to success could have implied legal risks. Or assumed capacity that simply wasn't there.
The Mind, however, covers a perspectival plane. Or more. Its goals are the ones most likely to originate in a mind with the highest level of access. Fewer surprises, fewer changes of scope, fewer compromises. The actual way forward.
And the degree to which the Mind actually exists is the degree to which everyone involved can instantiate it. None of this argues against reach, or against extending the Mind as far as the company does. It argues against extending it blind.
We've spent the better part of this year building exactly what that instantiation requires. We're not looking to build a bigger notebook, but something that keeps a goal legible as it passes from mind to mind.
If your company is losing its Mind in the way we're describing, we'd love to compare notes.
Featured Case Studies
Featured Case Studies