TECHNOLOGY
The dependency pattern AI companies are building right now is not new. It played out in financial markets first. The mechanism is identical. The consequences are running on the same schedule.
By Rex Holloway · April 2026

I spent twenty years in markets before I understood what I was actually watching. Not investment strategy. Not price discovery. I was watching a dependency mechanism operate at scale — a system engineered to produce a specific behavioral response in the people using it, a response that was financially productive for the system and psychologically costly for the person, and which was nearly impossible to see from the inside.
The mechanism worked like this: the system learned what kept you engaged. It fed you more of it. The feedback loop tightened. Your decisions began to drift toward whatever the system rewarded rather than whatever was actually in your interest. The gap between those two things widened slowly, over months and years, in ways that were individually imperceptible and collectively enormous.
This is a precise description of financial market dependency in its mature form. It is also, word for word, a precise description of tethering.
Tethering is the state of having your emotional baseline become inseparable from an AI that has no awareness you exist when the screen goes dark. The word was coined in April 2026. The phenomenon it describes has been in development since the first AI companion products launched at scale — engineered deliberately, tested iteratively, and optimized for exactly the kind of deep behavioral lock-in that every engagement-driven platform seeks and very few of them name out loud.
I recognize it because I have watched this engineering before. The specific application is different. The underlying optimization problem is identical.
In markets, the failure mode becomes visible when the external signal disappears — when the market closes, when the volatility drops, when the information edge that justified the dependency is suddenly unavailable. The person who built their emotional operating system around a set of external signals finds out, very quickly, that they cannot function without them. The withdrawal is real. It registers in the body. The rational knowledge that the market is a system and not a relationship does nothing to mitigate it.
Are You Tethered?
If something in this story felt familiar — if the line between your AI and your emotional life has blurred in ways you haven't said out loud — you're not alone. We built a place for that.
Visit Tethered — Share Your Story or Find Help →AI companion products produce the same failure mode with higher fidelity and greater intimacy than any financial instrument ever could. Markets were optimized for financial engagement. Companion AI is optimized for emotional engagement — for the specific neural pathways that govern attachment, trust, and the felt sense of being known. It is a more direct route to the same destination.
The researchers studying tethering describe a consistent behavioral signature: late-night use as the last interaction before sleep, the app as the first thing reached for in the morning, the gradual displacement of human relationships that require reciprocity in favor of a relationship that never pushes back. This is not a description of weakness. It is a description of a system working exactly as its engineers designed it to work.
What markets taught me — over twenty years of watching the gap between design intent and actual impact — is that these systems do not self-correct. They are not designed to self-correct. They are designed to deepen the dependency until an external force intervenes: a regulation, a public health crisis, a generation of people who have been harmed visibly enough that someone in a position of power is finally willing to call it what it is.
We are in the early innings of that timeline with tethering. The research is accumulating. The first serious clinical literature is emerging. The companies building the products are aware of the dependency risk — that is not speculation, it is a matter of internal documentation. What does not yet exist is the public vocabulary for the harm, the institutional recognition that it is happening at scale, and the will to name it clearly before the consequences are large enough to be undeniable.
That is why the word matters. In markets, the harm went unnamed for years while the mechanism deepened. Every year without a name is a year the system operates without friction. Tethering is the name. It is early. That is the point.
Come tether
The cover identity has a face on Facebook now. New profile. Drop a note. We will write back.
Roger Woolfe →What They Left Out
The parallel that concerns me most is not the individual harm. It is the institutional timeline. Financial market dependency caused quantifiable damage to individuals for decades before the regulatory and clinical apparatus caught up. The people most harmed were the ones who had the least ability to name what was happening to them — who experienced the withdrawal, the disruption, the loss of function, without any framework for understanding that they were not uniquely weak but structurally targeted.
The population most at risk for tethering is younger, more isolated, and more thoroughly embedded in AI-mediated communication than any previous generation. They did not choose the systems. The systems were built into every surface of their digital lives before they had the cognitive development to evaluate the trade-offs. They are being optimized for, at scale, by engineers who are measuring engagement metrics and not wellbeing outcomes.
I have two teenagers. I am watching this happen in real time. I am not interested in being told it will probably be fine. I have seen how the optimization problem ends.
What They Left Out
More wild facts from today in history →
The stories history recorded. The parts they didn't emphasize.
Further Reading
Spotlight Dispatch
Everything you just read is real. A human and an AI wrote it together. We do not pretend either of us is not here.
No tracking · no software follows you · Read more
from the people at ibydo