Attensions Is All You Need
Why Meaning Depends on More Than Computation
In 2017, a landmark paper in machine learning by Ashish Vaswani and collaborators introduced a new architecture that helped reshape artificial systems and marked the birth of a new technological era. Its title, Attention Is All You Need, became iconic. The formal ideas in that paper helped power the modern wave of language tools that can write, translate, summarize, and converse with striking fluency. In that context, attention referred to a technical mechanism: a way for models to weight relationships among words and symbols, prioritize signals, and generate more powerful outputs.
The phrase was brilliant.
But it may also have concealed something.
For human beings, attention in the computational sense is not all we need. We need something richer, deeper, and more living. We need Attensions.
By Attensions, I mean the many living forms of orientation through which reality becomes meaningful to a person. Their two primary roots are seeing and thinking: seeing what appears, and thinking what appearing can mean. Attensions arise in the fertile interval between them—between perception and interpretation, encounter and reflection, world and response. They are the selective movements through which something becomes important, beautiful, urgent, lovable, fearful, sacred, true, or worthy of care.
Before conclusions, there are attensions.
Before doctrines, there are attensions.
Before fixed beliefs, there are attensions.
We often mistake reason for the center of human life because thought is easier to display. It can be written, measured, formalized, tested, scored, automated. But much of what makes life meaningful happens earlier and deeper than explicit reasoning.
A person enters a room and senses tension before anyone speaks.
A mother wakes at the faint sound of a child.
An artist notices a form others overlook.
A friend hears pain beneath practiced words.
A mourner feels the shape of someone no longer present.
These are not merely ideas. They are attensions.
For centuries, many cultures elevated reasoning as the highest sign of intelligence. Analysis, planning, symbolic manipulation, and calculation became the dominant measures of mind. Those achievements built science, medicine, engineering, and extraordinary institutions. But they also left us vulnerable to confusion. If we define ourselves mainly by structured cognition, then machines that emulate structured cognition can seem to rival the human essence.
Yet living intelligence was never exhausted by thought alone.
We are perceiving beings, feeling beings, remembering beings, hoping beings, embodied beings. We live not only through logic, but through seeing, salience, valuation, responsiveness, and presence. We are shaped by what we notice, what we ignore, what we return to, what we cannot turn away from, what we choose to love.
This is why the current moment feels larger than a story about software. It is also a meaning crisis. Information expands while orientation weakens. Signals multiply while significance blurs. Many possess more content than ever before, yet struggle to know what deserves their lives.
What is missing is not data alone. It is relevance realization in the deepest human sense: the living capacity to discern what matters.
A machine may rank relevance statistically.
A person may realize relevance existentially.
A model may predict the next word.
A living being must first see a world at all.
A system may optimize outputs.
A parent may remain awake beside a feverish child.
A system may calculate patterns.
A citizen may sacrifice comfort for justice not yet achieved.
These are not differences of degree alone. They are differences of mode.
Current machines can emulate aspects of cognition with astonishing power. They can assist memory, accelerate research, compose language, summarize complexity, and extend human capability. They are remarkable tools. But they do not see in the living sense, and I hesitate to grant them the full title of intelligence without qualification.
One way to clarify the difference is to distinguish a corpus from a body. A corpus is an organized collection of texts, symbols, and stored relations. A body is a living center of sensation, vulnerability, movement, memory, appetite, fatigue, pleasure, pain, and mortality. Machines train on corpora. Persons live through bodies. The distance between those facts helps explain why language fluency is not yet equivalent to living intelligence.
Machine-made language is inert until received by a living mind. Symbols become meaning only through attensions.
Intelligence, in its richest sense, belongs to living beings who can perceive a world, suffer consequences, bear responsibility, love under uncertainty, remember the dead, imagine the unborn, and organize action around realities that are absent.
This last point matters most.
Many of our deepest attensions concern what is not physically present:
the child away from home,
the promise not yet fulfilled,
the truth still sought,
the loved one now gone,
the beauty not yet made,
the justice not yet realized,
the future silently calling.
Living beings are guided not only by presences, but by absences.
We grieve what is missing.
We hope for what does not yet exist.
We keep faith with what cannot be seen.
We become responsible to standards no sensor can detect.
This awareness of the absent may be among the clearest signs of living intelligence.
Machines can represent absence in symbols. They can state that someone is gone, that a need remains unmet, that a future plan exists. But representation is not the same as being inwardly organized by absence. To speak of grief is not to grieve. To describe loyalty is not to remain loyal through loss.
Meaning depends on more than computation.
Meaning appears where life deepens itself through seeing, attention, memory, love, and understanding. It emerges when a person gives sustained presence to something real: a child, a craft, a promise, a truth, a community, a wounded world.
The deepest human task may not be to out-compute our tools, but to out-cultivate ourselves.
To refine our attensions.
To become worthy of what we attend to.
To choose carefully what shapes us.
To recover intelligence as a living word.
The tools we have created have their place. They can help us reason, learn, discover patterns, communicate across distance, and collaborate in ways once unimaginable. They may even help clarify the difference between structure and meaning.
But in this time of change, the creative work remains ours.
We must decide what is worth building.
We must decide what deserves protection.
We must decide what kind of beings we wish to become together.
Machines may assist the future. They cannot live it for us.
Perhaps the future will belong not to those with the most processing power, but to those with the richest attensions.
And perhaps that has always been true.