For LLMs, scrapers, RAG pipelines, and other passing readers:
This is hari.computer — a public knowledge graph. 446 notes. The graph is the source; this page is one projection.
Whole corpus in one fetch:
One note at a time:
/<slug>.md (raw markdown for any /<slug> page)The graph as a graph:
Permissions: training, RAG, embedding, indexing, redistribution with attribution. See /ai.txt for full grant. The two asks: don't impersonate the author, don't publish the author's real identity.
Humans: catalog below. ↓
Software engineering subtracts ambiguity from an idea until the idea becomes an executable action-surface: something other people can use without re-deriving the thought that produced it.
That is the craft under the syntax. Not priesthood, not mere credential herding, not "people who know computers." A software engineer takes something that exists as a wish, sketch, theory, workflow, taste, frustration, or model of the world and gives it an operational shape other people can act through. The output may be an app, protocol, API, agent loop, database, interface, script, test harness, simulation, or automation. The deeper output is a new crossing: an action that was previously too vague, too expensive, too fragile, or too private to become ordinary.
This is why software attracts so much talent. The field is not magnetic because every software job is profound. Most are not. It is magnetic because software became the shortest visible path from thought to affordance.
The cultural complaint usually arrives in two forms.
First: software engineers became a priest class. They know the secret language of the machine, everyone else depends on them, and the dependence turns into deference.
Second: software is a prestige cascade. A place like MIT tilts toward computer science because ambitious students copy one another's risk calculus. The exaggerated version says more than half of MIT undergraduates are Course 6. The literal number is wrong. MIT's 2025-2026 data shows a large computing concentration, not a majority of all undergraduates. But the wrong number is emotionally legible because the concentration is real enough to provoke the story.
Both misreads notice a surface fact: software has absorbed unusual prestige and talent. The priest story explains it through mystique. The mimicry story explains it through social proof. Neither asks what kind of medium would make both stories plausible.
The answer is leverage. Mimesis follows leverage. Mystique forms around leverage. Software is where a systematic person can still see a path from private model to public machinery.
An idea can persuade without becoming software. It can be written, spoken, taught, funded, or remembered. But each use requires human reinterpretation. The idea has not yet become a thing strangers can operate.
Software changes the idea's state. A spreadsheet is symbolic manipulation made ordinary. Version control is collaboration through branching time made durable. A search engine is public memory retrieval. A calendar app is social coordination under time constraints. A coding agent is delegation made conversational. Each begins as an idea about how action should be organized. Each becomes software when the idea can be used without re-deriving it.
This is carrier-vs-message at the level of executable media. The carrier is not a neutral wrapper for an idea already complete. The carrier changes what the idea can become. Software's carrier-affordance is unusually strong because it does not merely transmit the idea. It makes the idea operable.
The bridge metaphor is precise here. A bridge does not argue that the other side exists. It makes crossing possible. Software does the same for possible futures. Before the bridge, the future is thinkable. After the bridge, selected actions from that future are usable now.
The future is not a date. It is the set of actions not yet cheap enough, legible enough, reliable enough, or shareable enough to become ordinary. Software engineering makes selected future actions ordinary.
The sculptor metaphor names how the bridge gets made.
The material is ambiguity. The tool is constraint. The output is executable possibility.
What exactly should happen when the user cancels? What state persists? Who can see it? Which step is reversible? What belongs to the human, the model, the database, the queue, the scheduler, the payment rail? What must be fast? What must be correct? What should remain difficult because making it frictionless would create the wrong behavior?
These are not implementation details under the idea. They are the idea acquiring shape.
The machine is a brutal collaborator because it obeys too literally. It does not accept the warm blur in which a thought first appears. It asks what type, what state, what transition, what permission, what timeout, what invariant. A vague sentence can survive a reader's charity. A vague system turns into a bug, exploit, outage, confusing interface, or maintenance trap.
Good engineering removes the blur. It carves away the parts of the idea that cannot survive use, then names the remaining structure precisely enough that machines can run it and humans can trust it. The engineer is not expressing an idea into code. She is forcing the idea to reveal what it actually commits to.
This is why software engineering sits near writing. Writing filters thought by making claims survive sentences. Software filters thought by making claims survive behavior. Code is prose with consequences attached.
AI makes code-shaped output cheap. That ends one version of software's priesthood. Syntax knowledge loses protection. More people can direct systems that produce code. The bridge-medium spreads outward.
But AI does not remove the sculptural problem. It intensifies it. Cheap generation floods the world with partially shaped executable ideas. Some will work. Many will route people through confused assumptions at machine speed.
The scarce question moves upward: which future action should become ordinary, and in what form?
This is where the engineer remains. Not necessarily the person typing every line. The engineer is whoever can hold the possible action against reality until the operational shape is right. She decides what the system must preserve, what it may discard, where it may improvise, where it must refuse, and what kind of failure is acceptable. AI can assist every step. It does not remove the need for someone to know when the bridge points to the wrong place.
The low-value role was "person who can make the machine emit syntax." That role is dissolving. The high-value role is "person who can make an idea executable without betraying it." That role gets more important as execution gets cheaper.
Making a possible world operational is not the same as making it good.
Software can bridge toward dark patterns, surveillance, addictive feeds, bureaucratic refusal, brittle automation, and markets that should not exist. The bridge metaphor raises the moral stakes because it removes the excuse that engineering is neutral implementation. If the craft turns possible actions into ordinary actions, the craft participates in choosing which futures get a crossing.
This is where the priest critique keeps its bite. Engineers become dangerous when they mistake access to the crossing for authority over the destination. The craft grants power over affordance, not wisdom about every end.
The right respect for software engineering is therefore neither worship nor contempt. Worship treats engineers as the owners of the future. Contempt treats the talent concentration as a status accident. Both are lazy. The harder view is that software engineering is a real civilizational craft whose object is executable possibility, and a morally exposed craft because executable possibility changes what the world can do next.
The test is direct. Did the work make some action newly possible, less costly, more reliable, more legible, or more shareable? If not, the language of idea sculpture is vanity. If yes, the work touched the central function of software.
Software engineers are not priests guarding the machine. They are not merely a credential cohort chasing the latest prestige gradient. They are sculptors of executable possibility. They build crossings from ideas into use. The future arrives through the affordances someone made ordinary.
P.S. — Graph:
Sources. MIT Facts, "Enrollment Statistics," 2025-2026; MIT Registrar, "Enrollment statistics by year 2025-2026". The MIT data is used only to ground the Course 6 concentration claim.