Pointed at Whom

Power funds the eye aimed at you and buries the one that would name who decided. The word for both is efficiency.

There is a famous photograph of Mark Zuckerberg back in 2016 at his desk, and the detail people noticed was not his face. It was the laptop behind him, with a piece of tape over the webcam and another over the microphone jack. The man who built the most powerful machine ever assembled for watching other people did not want to be watched through his own. This is not hypocrisy, exactly. It is a precise understanding of what a lens is for. Aimed at the billions, it is a business. Aimed at him, it is a threat, and threats get covered.

I spent two decades inside these rooms at corporate scale, and the same asymmetry runs through every one of them. A company will instrument the individual contributor down to the keystroke: every call logged, every ticket timed, every rep’s pipeline lit up on a screen the whole floor can see. The one thing it will not build is the instrument pointed at the decision that is actually sinking the ship. The four-year, twenty-million-dollar bet that everyone in the room quietly knows is dead has no dashboard, because a dashboard would produce a receipt, and the receipt would carry the name of whoever made the call. The eye points down, at the people with the least power to deflect it, and never up, at the choice that would cost someone at the top. It was never an efficiency problem. Aiming a camera never is.

A camera is power aimed. Everything depends on which way.

A government can learn the same lesson, and ours has.

Over the past year and a half the federal government has been building new instruments and destroying old ones at the same time, and if you read the headlines one at a time it looks like chaos, or budget-cutting, or ideology. It is none of those. Sort the instruments by which way they point and the pattern resolves into something almost elegant. Every instrument that got funded is pointed at citizens. Every instrument that got pulled was pointed at something that would cost the government to look at. The state is not going blind. It is choosing, with great care, what to see.

In March of 2025 the president signed an executive order directing federal agencies to merge the data they hold on Americans and eliminate information silos, in order, the order says, to “increase government efficiency and save hard-earned taxpayer dollars.” The instrument being used to build it is largely Palantir’s, the analytics firm co-founded by Peter Thiel. Its software has spread into Homeland Security, Health and Human Services, the IRS, where engineers are working on a single searchable layer over the agency’s entire store of tax data, and Immigration and Customs Enforcement, where a thirty-million-dollar contract powers real-time tracking of migrants. The administration has sought access to hundreds of data points on individual Americans: bank account numbers, the size of a person’s student debt, their medical claims, their disability status. Palantir has taken well over a hundred million dollars from the government since the term began, on top of a defense contract worth up to one and a third billion.

Palantir says, and this may be true, that it does not itself collect the data, and denies it is assembling a master file to surveil Americans. Take the denial at face value. It does not touch the point. The point is the direction of the spending. This is the state paying real money to build the capacity to see one particular object in extraordinary detail, down to your bank balance and your disability status, and that object is you.

Now hold that against what the same government did to the instruments pointed the other way.

After the 2022 gun bill, Health and Human Services paid for a study on red flag laws, written by researchers whose work shows the laws save lives, and posted it publicly. After the administration changed, it vanished from the government’s site. It survives on an obscure nonprofit page its own lead author was startled to find still existed, and he was given no reason for its removal. The surgeon general’s advisory naming gun violence a public health crisis came down too. And the CDC division that studies how people get shot, and how to stop it, was hollowed out by layoffs.

Here is the part that matters most, because it is the exact inverse of the Palantir investment. A researcher who spent seventeen years in that division said the data is still there. The government is still counting gunshot injuries. What was removed was the capacity to do anything with the count: the staff to study it, to build a program from it, to help a city keep a wounded person from being shot a second time. Gun injury is the leading killer of American children. About a hundred and twenty people die of it every day. The numbers go on being collected, faithfully, into a drawer no one is left to open.

Set the two side by side. For citizens, the state is building the machinery of consequence: merge the records, act on what they show, track and flag and police. For the thing that kills the most children, the state kept the raw count but destroyed the machinery of consequence. It is not that one fact is seen and the other unseen. It is that the state invested in the power to act on what it sees in exactly one direction, and the direction is down, at you.

Deletion is a purchase too.

If it were only guns you could call it gun politics and half the country would stop at the label. It is not only guns. In May the National Science Foundation announced it was dismantling the Ocean Observatories Initiative, some nine hundred deep-ocean sensors moored off five coastlines, three hundred and eighty-six million dollars to build, a decade of real-time data into a planned thirty-year mission. The instruments were not switched off and left. By early June the Foundation had spent ship-days hauling more than two hundred of them out of the water off Oregon and Washington, which costs money that mere neglect would not. Among the things those sensors watch is the Atlantic current system whose collapse is one of the real tail risks of a warming world, and they came out during a stretch of record ocean heat, exactly when someone who wanted to see the danger coming would most want the gauge. The Foundation’s explanation was that this was not a cancellation but a move to, in its phrase, “a nimbler approach.”

Then, in mid-June, the pattern broke, and the break turns out to be the most instructive receipt in this essay. Fishermen who plan seasons by that data objected, forecasters objected, coastal senators of both parties objected, and the Senate passed a bill to save the array by unanimous consent. Within days the Foundation reversed itself: no further removals, the recovered sensors to be serviced and put back in the sea.

Notice the word doing the work in all three cases. The surveillance order was to increase government efficiency and save taxpayer dollars. The gun study cuts were the work of “responsible stewards of American taxpayer dollars,” ending what was wasteful. The buoys are a nimbler approach. One vocabulary—efficiency, stewardship, nimbleness—funds the eye on the citizen and pulls the eye off the ocean and off the body count. It is the same word aimed in opposite directions, and the direction is the entire meaning.

Some of this is defensible, and a serious version of the case is easy to make. Government data really is siloed and duplicative, and merging it catches fraud that costs all of us. Watching genuine threats is a real function of any institution. An aging instrument is a fair thing to retire when a better one takes its place. A company that killed a redundant dashboard and a state that consolidated three broken databases would both just be doing their jobs.

The tell is that the efficiency only ever runs one way. Real reprioritization would sometimes turn the new instruments on the government’s own performance, and sometimes build a sharper eye on the ocean. Left to its own judgment, this never does. The money for sight flows toward watching citizens, and the instruments the state chooses to retire are the ones that watch what it would answer for: the warming it is accelerating, the children it is failing to protect, its own record. The buoys are the exception, and I want to be honest about what the exception cost my argument, because it cost the clean version and bought a truer one. The eye went back in the water. It went back because a Senate that could not be waved off, voting unanimously, compelled it. So the aim is not fate. It is a decision, and a decision can be overruled. What it has never once been, in the record of the past eighteen months, is reconsidered from the inside. That is not the retirement of obsolete instruments. It is the curation of sight by whom it threatens, and it bends exactly where, and only where, someone retains the standing to threaten back.

Calling it efficiency is not a description. It is the tape over the lens.

There is a comforting way to end here, and it is false, so I want to name it and put it down. The comforting read is that the facts don’t care: the ocean warms and the children get shot regardless of what gets deleted, and reality always wins in the end. It is true and it is beside the point, because reality was never the thing that sent anyone a bill. The instrument was.

And this is not a story about one government, which is the easy version and the one that lets the rest of us off the hook. It is the reflex of every room where power sits. The executive tapes the lens that points at him and pays celebrities to sell the ones that point at everyone else. The company lights up the rep on the floor and builds no dashboard for the decision that is actually killing it. The state reads your bank balance and files the count of its dead children in a drawer no one is left to open. The scale changes and the move does not: aim the instruments down, at the people with the least power to answer back, and go dark on whatever would name the person who decided. Every one of them calls it efficiency, and every one of them is telling the truth, because seeing selectively is efficient, for the one who chooses where the eye goes.

So the question is not what your government or your hyperscaler can see. It is the same question in every room you have ever worked in, and the pattern does not change with the size of the room. Of every instrument that gets built and every instrument that gets buried, ask the only thing that matters. Pointed at whom.

✦ ✦ ✦

And now for my closing ritual: Thesis & Texture

Each week, I end with two book recommendations. One that sharpens how we think about technology, systems, and power. Another that holds what systems can’t: the texture of being human.

Because building good systems while living a beautifully full human life requires both.

The Thesis: Automating Inequality by Virginia Eubanks. This essay is about the direction sight gets aimed, and Eubanks is the book about that direction. She reports from inside the welfare offices and child-services algorithms where automated systems watch the poor in punishing detail while the systems meant to serve them are never built. It is the same asymmetry, at the scale of a single family: the eye pointed down, at the people with the least power to answer back, and the machinery of consequence built only to discipline them, never to help.

The Texture: Klara and the Sun by Kazuo Ishiguro. Klara is pure attention. An artificial friend who watches the humans around her with total, patient care and sees precisely what they are too close to see themselves, she is the lens the essay never gets to show: an instrument aimed with love instead of power. It is a quiet, devastating book about being seen fully by something that wants nothing from you, which is the opposite of everything watching us now.