Parā Technica #4: When Desire Becomes Executable
The power to will, the discipline of introspection, and the question beneath agency.
A person opens an AI assistant and asks it to draft a proposal, write a message, compare options, summarize the research, create a prototype, and recommend next steps. Twenty minutes later, they have a better plan than they would have produced on their own. The response is polished. The tradeoffs are thoughtful. The next steps are clear. Then the machine asks a question it cannot answer for them: What outcome are you trying to achieve?
Suddenly, the difficult part is not the work. It is the wanting.
The current obsession with agency is both right and incomplete. ‘Agency’ has become one of the words of the moment. Companies want agentic employees. Founders praise agentic teams. Parents want children who take initiative. Technologists now speak of systems that do not merely answer, but plan, call tools, execute tasks, and move through workflows. The word is useful because it points toward something productive.
However, it is also becoming vague. Sometimes agency means action or initiative; sometimes ambition, independence, ownership, drive, or simply doing more. But not all action is agency, and not all agency is action.
A person can move quickly and still be carried by fear. A person can experience distress, yet remain still and composed. A person can choose and still be choosing from desires they never examined. Two people can perform or refrain from the same action—accept a job, build a product, follow a model’s recommendation—when only one is acting with real agency.
Understanding agency becomes urgent because AI is changing the cost of action. For decades, technology amplified the hand. Now AI increasingly operates upstream of it: thinking, summarizing, advising, comparing, planning, generating, and sometimes even choosing.
The previous article asked when assistance becomes substitution. This article explores one level deeper. If AI can help us do, then what should we ask it to do? And beneath that: What do I really want done?
The Three Powers
Mohak: There’s something interesting I’ve noticed about the way people talk about AI. Their first questions are practical: What model should I use? What workflow should I build? What tasks should I automate? But the conversation often drifts somewhere stranger: What should I spend my time on? What else can I use this technology to accomplish?
If execution keeps getting cheaper, merely doing things is no longer the differentiator. The human task shifts from doing to directing. But directing requires knowing where I am trying to go, what matters, and what I am willing to sacrifice.
We are building machines that do not merely make work easier; they make desire executable. They can write memos, generate plans, schedule meetings, summarize documents, and perhaps take the next steps without much human involvement. That can feel like freedom. But it also exposes the harder question: What step is worth taking?
In work, this shows up when a team uses AI to accelerate a roadmap before asking whether the roadmap deserves to exist. We have all seen teams execute against goals that, in retrospect, were not worth pursuing. AI lowers the cost of execution, but it does not lower the risk of pursuing the wrong thing. In some ways, it raises it.
The same pattern appears in parenting, when we help a child achieve before helping them navigate what kind of person that achievement is forming. So agency cannot only mean drive, speed, or initiative. A person can be moving fast but still be moving in a direction driven by fear, comparison, habit, or unjustified aspiration. The question is not merely whether I am acting. It is: Why am I acting this way, and is this how I want to act?
AAD: Indian thought offers a useful framework for dissecting agency. Agency can be understood through three powers: kriyā śakti, the power to act; icchā śakti, the power to desire or will; and jñāna śakti, the power to know.
Action alone does not demonstrate agency. Habit, pressure, imitation, fear, conditioning, or even mechanical potential can all produce movement. To know whether action is truly agentic, we have to ask what knowledge informs it and what desire moves it. Notice how sentience is built into the formula through these elements.
Although these powers are distinct, they do not remain separate. What we know shapes what we desire. What we desire shapes what we do. And what we repeatedly do reshapes both our desires and knowledge. This integration obliges us to be careful about each of the three: the knowledge we acquire, the desires we cultivate, and the actions we repeat.
Icchā (desire) is especially important here because desire is not merely preference. Preference says, “I like this.” Icchā points to something deeper: “This is the direction in which my being is moving.” That direction may be noble, base, directed, confused, shallow, or sacred. But it is not neutral.
Agency is not merely the power to act. It encompasses knowing, wanting, and doing.
Mohak: This is where the framework becomes concrete for me. If AI makes action easier and can supply the materials we often mistake for knowing—information, summaries, arguments, comparisons, and analysis—then icchā becomes more important.
The machine can help me gather information, compare possibilities, and formulate plans. But it cannot finally tell me what is worth wanting. Greater capability does not clarify desire. It amplifies it. Convenience becomes the default. Attention becomes addiction. A vague and unexamined wish becomes a roadmap.
The machine does not make the desire wiser. It makes the desire more executable.
We praise drive without asking what is driving the drive, ambition without asking where it came from, and self-direction without asking what self is doing the directing. A person can be highly driven and still not be deeply agentic. We’ve all worked with someone who appears exceptionally busy yet seems not to get all that much done. Or at least not anything of particular value. That is why “What do I want?” and perhaps more importantly, “What should I want?” are not soft questions. They are technical, practical, moral, and spiritual all at once.
The Desire Beneath Action
AAD: This is why icchā śakti (the power to desire/will) deserves special attention. Desire is not simply “whatever I feel,” nor is it pure because it feels internal. It can be shaped by habit, environment, past action, ego, comparison, attachment, and fear.
Some desires announce themselves clearly. Others hide behind productivity, staged identities, professionalism, or even independence: “I want to lead” may include wanting recognition; “I want freedom” may mean escape from responsibility and accountability.
So we cannot understand agency by looking only at action. We have to look beneath it to desire. And we cannot understand desire merely by naming whatever appears first in the mind. The first answer may be the loudest, not the truest.
This is where the principle of antardṛṣṭi (un-tur-DRISH-tee) becomes important. It is often translated as introspection or self-reflection, but that can make it sound too thin. Antardṛṣṭi is not merely thinking about myself, nor is it just turning thoughts over in the mind; it is looking inward and “seeing” what is there.
When applied to desire, it asks, “What do I truly want?” And then, more sharply: is that desire worthy? Such reflection does not merely discover desire, as if every inner impulse becomes sacred once named, but discovers what that desire is making of us, and whether that is who we mean to become.
Antardṛṣṭi is the discipline of examining desire long and hard enough to see not only what it is becoming but also what we are becoming.
Mohak: This is the deeper question I keep hearing underneath practical AI questions. I’ve heard versions of it from executives, engineers, students, parents, and teachers. The wording changes, but the question underneath rarely does: What will still matter? What should I become? What kind of life am I trying to build?
If I do not know that, then every tool becomes a source of confusion. A powerful tool doesn’t solve the problem of direction; it just makes it harder to ignore.
I imagine myself in an incredibly fast car, capable of covering enormous distance in no time. But if my destination isn’t clear, the speed is worthless. I can drive for hours and arrive nowhere worthwhile. The speed of the car was never the question.
There is an old image that captures this: You can fill a page with zero after zero, hundreds of them, but without a “one” in front, the whole thing amounts to nothing. The zeroes are the effort; the one is the purpose that gives them value. Without it, more effort just means more zeroes.
“Hundreds of zeroes, patiently drawn,
with not one numeral to stand before them —
what sum can such blackened paper yield?”
- Niṣkulānanda Svāmī
And so, I keep coming back to a very simple question: What have I come to do, and what am I doing?
That question is about alignment with a purpose. Am I living in a way that matches what I claim to value? Am I delegating something that I should be forming in myself? Am I moving because I chose the direction, or because the world around me supplied one?
AAD: The answer is rarely obvious. A person may ask, “What do I want? What is my purpose?” and honestly answer, “I do not know.” That is not a failure but the beginning of clarity, often developed by repeatedly returning to the question. This positions antardṛṣṭi not as a one-time exercise but as a habit, until it becomes second nature.
Beyond Self-Reflection
Mohak: Sometimes people hear this and think it means endless doubt. But the point is not to lose conviction. The point is to make sure conviction can grow. A popular saying in the tech community these days is, “I might be wrong, but I’m not confused.” It sounds wise as it preserves the ability to act under uncertainty. But it’s also incomplete. I have watched leaders confuse the absence of confusion for the presence of clarity. They are not stuck. They are not paralyzed. They are simply locked in early, on a thesis they have not yet given themselves permission to revise. The quote is meant to defend agency, but it can quietly defend something else: the refusal to be moved by what one learns.
The need is not merely conviction. It is a conviction that can grow.
AAD: This kind of growth asks instead of assuming, while humility allows the answer to cost us something. It loosens the assumption that my desire is trustworthy simply because it is mine. Otherwise, the self becomes the defendant, the jury, and the judge; not a good pairing. The desires under review and the standards of review come from the same source. Reflection produces conviction, but not necessarily clarity—the two are easy to confuse when both feel internal.
For reflection to really work, antardṛṣṭi has to be grounded in something beyond the self—something knowledgeable and steady. This builds on the role of the teacher we discussed in our previous article. Sailors have long used dead reckoning: navigating from a last known position using speed, heading, and elapsed time. The method feels rigorous from the inside; each calculation follows from the last. But small errors compound, and without a fixed external reference such as Polaris, a ship can drift confidently in the wrong direction. Introspection without an external anchor is dead reckoning of the soul.
The Gītā describes the alternative as being “established in the eternal” (nityasattvastha). It advises connecting our desires to a stable foundation—a spiritual ideal that Hindu philosophy calls Akṣara, or “the imperishable”—that which does not change, and therefore can serve as a reference for everything that does. It is what makes correction possible at all. Without something outside the drift, there is no way to know one is drifting.
Mohak: That changes the way I think about agency. The world often treats agency as self-assertion: I decide, I act, I build, I move. But that assumes we already know what is worthy of our action. If the self is full of unexamined desire, then assertion is putting the cart before the horse. Real agency demands the opposite movement first: self-examination to ensure that what we are asserting is worth it.
This is why agency and reflection must go together. If we act without examining the inner movement, desire remains in command. If we reflect without disciplining action, the same desires keep strengthening. Repeated action feeds desire; repeated desire colors knowledge; colored knowledge shapes the next action. Repeatedly asking, “What have I come to do, and what am I doing?” examines and aligns all three powers at once: what I know, what I want, and what I am actually doing. Most importantly, it grounds that alignment in a stable, purposeful foundation.
Knowing what is worth wanting matters greatly in a world where AI makes access to knowledge and action plentiful.
The risk is no longer that we won’t get what we want; the risk is getting it and realizing we didn’t want it at all.
Questions to Carry
“What have I come to do, and what am I doing?”
When the distance between wanting and achieving narrows, what becomes worth creating?
Does something need to be useful for it to be well-desired?
What remains stable enough to guide desire when feelings, fears, and ambitions keep changing?