Purpose and Discernment
Our computational power is expanding our capacity to assert purpose. Discernment must be cultivated at the same pace.
The story of Homo sapiens unfolds inside a single question: how do we get the world to do what we want?
That question names our strange evolutionary inheritance: the capacity to envision multiple futures and choose actions most likely to bring about the ones we prefer—i.e., the intentional assertion of purpose. It is so native to consciousness that we barely notice it.
Purpose, however, is not a static possession. It has evolved by widening its scope—envisioning more ambitious goals thus creating more possibilities. No other species seems to exercise it anywhere near our scale, especially once purpose is extended through tools, coordination, and inherited practices.
Artificial intelligence accelerates this strange evolutionary inheritance at previously unfathomable speeds and scales. It raises an old question with new force: can our powers of discernment keep pace with the opaque speed of our technologies?
In the midst of political rupture and accelerating artificial intelligence, our current dangers emerge from the maniacal pursuit of purpose—politics arranged around the salvation of particular ‘identities’, corporations focused only on maximizing profits at the expense of all else; so-called Christian institutions aggressively pursuing nationalist agendas.
Will our artificial intelligences expand our capacity to envision more and choose wisely, or will they further narrow our vision?
Discernment and wisdom are part of our evolutionary inheritance of purpose. They can keep purpose from becoming too narrow, too maniacal. They are critical skills that must be rejuvenated if we are to turn rupture into passage, rather than nihilistic collapse.
The question, then, is not only whose purposes prevail, but how each of us retains our capacity for discernment amidst political rupture and accelerating AI.
What is purpose?
Any serious engagement with what it means to be alive reveals at least this much: living things are not inert; they can influence what happens next.1
Small scale: A bacterium moves through its liquid environment by chemotaxis—sensing chemical gradients, often nutrients such as sugars, and biasing its motion toward more favorable conditions.
Leveling up: A moth eludes a bat by introducing unpredictability into its flight, breaking the bat’s ability to predict where the prey will be in the next few milliseconds. This unpredictability is the moth’s purposeful defense. It throws the bat’s predictive model of the future into disarray.
Asserting purpose: Humans have a far more expansive capacity to influence what happens next. We can envision many futures across various time horizons, argue about them, revise them, and arrange causal sequences that bring about the futures we want.
This is the assertion of purpose, and it is foundational to how we influence what happens next.
We can also activate a wide spectrum of influence as needed. We can randomize our behavior, and we can move on instinct. Watch any American football game: a running back makes himself unpredictable to evade the would-be tackler, while the moment plays out inside longer horizons with explicit purposes—game and play clocks; choreographed plays nested in game plans; personnel choices made over weeks and seasons; training regimens built over years.
Our repertoire for asserting purpose is wide, and we can switch modes with remarkable speed.
Our signature move is organizing causal sequences from the future backward—designing causes to bring about the effects we want.
Technology puts effects before causes
I’ve set my first-floor thermostat to raise the temperature from 64 to 68°F before I come downstairs around 5:30 a.m. Connected to the furnace and ductwork, it will do this predictably—so long as everything is working as designed.
From a cause-and-effect perspective, I begin with the effect I want—68°F by 5:30 a.m.—and the thermostat triggers the causal chain to make it happen.
This is ‘backward causality’ in a practical sense, and it is fundamental to human technology: we start with desired effects and organize causes accordingly. Without a thermostat and heating system, the indoor temperature simply follows the outdoor world—the laws of physics running unaided.2
This is what differentiates our purposive behavior from that of bacteria, moths and bats. We don’t simply react within causal chains; we imagine multiple futures and arrange causes and effects to favor some over others.
In this way, we get the world to do more of what we want—purposefully and intentionally.
Purpose, Feedback and Discernment
In 1943, cyberneticists Rosenblueth, Wiener, and Bigelow identified a fundamental feature of purpose-driven behavior: it depends on feedback—signals that indicate whether action is carrying one closer to, or farther from, a goal. Negative feedback matters most because it corrects; positive feedback merely confirms that things are proceeding as expected.
All purposeful behavior may be considered to require negative feed-back.3
A guided missile relies on negative feedback to course-correct toward its target, just as a running back relies on it to evade a linebacker and find the momentary opening in the defense.
While ‘negative feedback’ names something real, it risks reducing a vast range of human capacities to a narrow instrumental description. We begin to resemble mere machines—even if Rosenblueth and company meant to blur the line between organism and mechanism.
In practice, the blur too easily favors mechanism over the multi-dimensional vitality of life in pursuit of purpose.
Isn’t feedback—positive or negative—what spiritual traditions have long called discernment and attention in their more venerable sense: sustained, cultivated presence that cultivates greater introspection and understanding?
As Pierre Hadot observed in his work on ancient spiritual exercises:
Attention to oneself … gives rise to a whole series of techniques of introspection. It engenders an extraordinary finesse in the examination of conscience and spiritual discernment.4
I prefer these words—and the practices that accompany them—because they open our computational intelligence to a fuller register: introspection, intellectual discipline, wisdom, and other dimensions often flattened by contemporary AI talk. We can call these capacities computational or cybernetic—which is accurate—without reducing lived intelligence to the step-by-step execution of a Turing machine.5
Some may want to hold out these experiences—discernment, attention, and wisdom— as a hard boundary—a difference in kind—between humanity and our computational technologies. Others may want to discount them as fallible human emotions—noise gumming up the workings of a homogeneous and instrumental rationality. Neither response is correct. We can understand attention, discernment, and wisdom as forms of computation—not in the narrow, formal sense of instrumentalism or relentless rule execution, but as embodied, feedback-guided processes of imagination, selection, and revision.6
A deeper dive into the evolution of purpose—our ability to envision possible futures and arrange causes to bring about our preferred ones—should lead us to a rejuvenation of more venerable aspects of intelligence: discernment, attention, and wisdom. Together, they name the capacity to pause, reflect, and sharpen how we navigate a complex world. They keep the expanding reach of purpose from becoming the maniacal and intolerant grab of an enforced ‘destiny’—of oneself, one’s nation, or one’s species.
These capacities are fundamental features of computational intelligence. They are the means by which our capacity for purpose has expanded, precisely because they allow us to hold open multiple futures before proceeding, even if our assertion of purpose can all too easily slide into maniacal single-mindedness.
Attention, wisdom, and discernment are not ancillary to purpose; they are essential to it. Calling them by other names risks sterilizing a tradition of spiritual exercise at precisely the moment it is most needed.
I’ve experimented with ChatGPT by asking it to interpret this essay, which helped me to edit and clarify during the writing process. See the results here, especially the sections where I asked it to summarize (in bullet points) ‘What I am saying’ and ‘What I’m NOT saying.’
This essay is part of my Rejuvenation Series.
I use the words ‘purpose’ and ‘influence’ somewhat interchangeably but with a difference in emphasis. ‘Purpose’ emphasizes the bundle of inherited capacities a living organism uses to influence causation, while ‘influence’ emphasizes the effects of purpose on a given causal chain. The moth can purposefully randomize its movements by activating neurons in a randomized way while its influence must be understood by watching what happens to the bat’s movements in response.
By ‘backward causality’ I do not mean that physical laws literally run backward in time. Rather, this names a well-established feature of purposive systems in philosophy, engineering, and cybernetics: present actions are organized and selected with reference to a future condition or goal. Human technologies routinely work this way: engineers design systems backward from specifications, arranging forward-running causal chains to achieve chosen effects. For a concise description of ‘the self-fulfilling paradox of backward causality’, see the section ‘Cause by Effect’ in Blaise Agüera y Arcas, What Is Intelligence? (MIT Press 2025), pages 125-7.
‘Behavior, Purpose, and Teleology’ p. 19. See Agüera y Arcas What Is Intelligence? pages 163-74 for a discussion of ‘purposiveness’ and ‘negative feedback’. He is drawing from Rosenblueth, Weiner, and Bigelow.
Pierre Hadot, Philosophy as a Way of Life (Blackwell, 1995) p. 136. Hadot provides an excellent synopsis of attention and discernment in ancient philosophy. For the Stoic and Epicurean traditions, see pages 84-5; for the Christian tradition, see pages 130-6. For a Jesuit’s take on discernment and attention in prayer, see James Martin, S.J. Learning to Pray, ‘How Do I Know It’s God: Discerning God’s Voice’, which is valuable reading even for those who don’t profess Christianity.
By computation I mean not only formal symbol manipulation, but any process that predicts and influences possible futures by processing feedback.
The word attention has famously re-entered our technical vocabulary through contemporary AI, but with the mechanistic and engineering connotations that distinguish it from the more venerable practices of attention described by Pierre Hadot. See the influential 2017 paper on Transformer architectures ‘Attention Is All You Need.’
See also the final chapter of Agüera y Arcas’s What Is Intelligence? for the dangers of defining intelligence as a narrow vision of computing, such as ‘optimizing a value or utility function’ or the ‘obeying of rules’. He, too, sees ‘attention’ as a computational power that need not be drained of more spiritual experiences: ‘Attention is a real enough sort of computation, useful for predictive modeling . . . It is reasonable, then, to define a being, a soul, or a “who” as that which can pay attention—and can model that attention’ (What Is Intelligence?, 284).