Vol. III, No. 62
Covering 21 April - 4 May 2025
Monday, May 5, 2025
Late edition · A Relentless publication
The Relentless Dispatch
All the fortnight that mattered, in technology and in the world, read next to what we were building at the time.
KASHMIR

A massacre of tourists pushes India and Pakistan toward the brink again

Gunmen killed 26 people, almost all tourists, at a beauty spot in Kashmir. India blamed Pakistan, suspended the treaty that shares their rivers, and moved troops to the border. Two nuclear-armed neighbours edged, once more, toward war.

Gunmen opened fire on tourists at Pahalgam, a scenic meadow in Indian-administered Kashmir, on April 22, killing 26 people, almost all of them Indian visitors, in the deadliest attack on civilians in the region in years, singling out victims, survivors said, by religion. India blamed Pakistan-based militants and, beyond them, the Pakistani state, and responded with a ferocity that signalled this would not be contained to the usual exchange of condemnations: it suspended the Indus Waters Treaty, the 1960 agreement that governs the rivers on which Pakistan's agriculture depends and that had survived every previous war between the two countries, expelled diplomats, closed the border, and moved forces forward. Pakistan denied involvement, called the treaty suspension an act of war, and readied its own forces. Two nuclear-armed nations, with a history of four wars and a permanent, bleeding quarrel over Kashmir, stood once again at the edge of the abyss, and the world, distracted by tariffs and a papal funeral, was slow to notice how close the edge had come.

The suspension of the water treaty marked something new and dangerous. For 65 years, through wars hot and cold, the Indus Waters Treaty had held, an island of cooperation in an ocean of enmity, because both sides understood that water was too fundamental to weaponise. India's suspension of it, in fury over the massacre, crossed a line that had held for two generations, threatening the water supply of a nation of 240 million and signalling that this crisis would be fought without the old restraints. Whether it escalated to open war or was pulled back from the brink depended on decisions being made in Delhi and Islamabad under the pressure of public rage and the logic of deterrence, the same escalation machinery this paper has watched grind toward catastrophe elsewhere, now operating between two states that hold, between them, some 300 nuclear weapons.

THE WORLD

Voters abroad punish anything that looks like Trump

An unexpected pattern emerged from elections across the democratic world: association with Donald Trump had become electoral poison. In Canada, the Liberal Party, written off for dead months ago, surged to victory on April 28 under Mark Carney, its revival powered almost entirely by a nationalist backlash against Trump's tariffs and his musings about annexing the country. Days later, on May 3, Australia's Labor government under Anthony Albanese won a thumping re-election, as its conservative opposition, tarred by comparisons to the American president, collapsed. The global anti-incumbent wave of 2024 had reversed: in 2025, incumbents who could position themselves against Trump were thriving, and the opposition parties that had borrowed his style were being punished.

IN BRIEF

A conclave gathers; a president's hundred days; an oracle steps back

The cardinals of the Catholic Church gathered in Rome to bury Pope Francis on April 26, a funeral that drew world leaders including a brief, awkward encounter between Trump and Zelensky, and then to begin the conclave that would choose his successor. Trump marked 100 days in office on April 29 with the lowest approval ratings of any president at that milestone in the modern era, as data showed the American economy had shrunk in the first quarter. And Warren Buffett, at 94, announced that he would step down at year's end as chief executive of Berkshire Hathaway, ending the most storied run in the history of investing.

The Column

The line you do not cross, even in war

India suspended a treaty this fortnight that had survived four wars, because water was the one thing too important to weaponise, until it wasn't. The suspension matters beyond the region, because it is the breaking of exactly the kind of restraint that keeps conflicts from becoming catastrophes.

By Meera Anand

After a massacre of tourists in Kashmir, India suspended this fortnight the Indus Waters Treaty, the agreement governing the rivers it shares with Pakistan, a treaty that had held through four wars and sixty years of enmity because both sides had understood, tacitly, that water, the thing on which the lives of hundreds of millions depend, was too fundamental to be turned into a weapon. The suspension is being read as a bilateral escalation, and it is that. But it matters beyond the subcontinent, because it is the breaking of exactly the kind of restraint that separates a dangerous rivalry from a catastrophic one, and the logic of that restraint, and of what happens when it breaks, was mapped by a strategist whose work every nuclear-armed rival should still be made to study.

Thomas Schelling, who thought harder than anyone about conflict between adversaries who cannot fully communicate, identified the strange power of what he called focal points and tacit bargains: the unspoken thresholds that rivals come to observe not because any treaty compels them but because both recognise, without saying so, that certain lines are different in kind, and that crossing them removes a restraint whose value depends on its never being tested. Some limits in a conflict are conventional, a matter of degree; others are focal, qualitative, salient, lines that both sides tacitly treat as uncrossable precisely because they are recognisable as lines, and the not-crossing of them is a coordination that holds the whole rivalry within bounds. Water, in the India-Pakistan conflict, had become exactly such a focal point: a thing so fundamental that both sides, through four wars, treated its weaponisation as beyond the pale, and the treaty's real function was less legal than Schellingian, a marker of a line whose power lay in both sides' shared understanding that it would not be crossed. To cross it now is not merely to escalate by a degree; it is to erase a focal point, to remove one of the tacit limits that had kept the rivalry survivable, and Schelling's warning was that focal points, once erased, do not come back, because their power was never in the line itself but in the shared belief that it was a line.

The counterargument is understandable and I will not wave it away: a nation whose citizens have been massacred is entitled to respond, restraint can look like weakness that invites the next atrocity, and there is something perverse in asking the wronged party to honour a limit its enemy has, by sponsoring terror, already dishonoured. This has real moral force, and I do not dismiss the fury behind the suspension. But Schelling's point is not a moral one about who deserves what; it is a strategic one about consequences, and it is this: the focal points that restrain a conflict are not favours the sides do each other but shared infrastructure that protects them both, and breaking one to punish an enemy does not merely hurt the enemy, it dismantles a limit that was keeping the whole rivalry from tipping into the unlimited. Water was too important to weaponise not out of tenderness but out of a shared recognition that a conflict in which water is a weapon is a conflict without a floor, and the tragedy of erasing that focal point is that it makes every future crisis more dangerous, because the tacit understanding that some things are not done has been shown, once, to be revocable.

And that is the loss that reaches beyond this fortnight and this border. Between adversaries who cannot trust each other, who share no sovereign, who communicate mostly through threats, the tacit limits, the focal points, the lines both sides observe without a treaty to enforce them, are very nearly the only thing standing between a dangerous rivalry and a catastrophic one, and they are held in place by nothing more solid than a shared belief that they will be held. Every time one is broken, the belief that sustains the others weakens, because the whole system of restraint runs on the perception that these lines are real, and a line shown to be crossable when the provocation is great enough is a line that will be crossed again. India had its reasons, and they were real. But the treaty it suspended was not merely a bilateral agreement; it was a focal point in Schelling's exact sense, a marker of a limit whose value lay in its never being tested, and the water that was too important to weaponise, until it wasn't, is now a weapon, and the floor it represented, for two nuclear-armed nations, is one line lower than it was.

Field Notes
A Relentless build, told plainly

The system nobody uses, and the unglamorous truth that building it is the easy half

This fortnight's front page is about restraints abandoned under pressure. This is about a quieter failure I have seen destroy more good software than any bug: the excellent system that no one adopts, and why getting people to actually use the thing is harder than building it. No client is named.

By Alosh Paul

I have built systems that were, by every technical measure, excellent, and that failed completely, because no one used them. It is the most common and most demoralising failure in enterprise software, and it taught me the single most important lesson of my career: that building the system is the easy half, and getting human beings to actually adopt it, to change how they work, to trust it, to use it the way it was designed to be used, is the hard half, the half that determines whether all the building was worth anything at all. A brilliant system that people route around, that they quietly keep doing the old way beside, that they enter bad data into because they never understood why it mattered, is worth less than a mediocre system that people genuinely use, because software delivers its value only when it is actually used, and use is a human problem, not a technical one.

The mistake I made early, and watch others make constantly, is to treat adoption as an afterthought, something you handle with a training session at the end, once the real work of building is done. It is exactly backwards. The people who will have to use the system are not obstacles to be managed after the fact; they are the entire point, and if you have not understood their actual work, their real incentives, the reasons they do things the way they currently do, then you have not gathered requirements, you have gathered wishes, and the system you build will collide with reality the day it ships. People do not resist new systems because they are stupid or stubborn. They resist because the new system asks them to change, and change is costly to them personally, and if the system does not visibly make their own working life better, they will, entirely rationally, keep doing what they know, whatever the org chart says. Adoption is not won by mandate. It is won by building something that actually serves the people who have to use it, and by bringing them along from the start so that it is, in a real sense, theirs.

This is why I have come to believe that the human side of building software, the change management, the training, the patient work of understanding and winning over the people who will use the thing, is not soft, secondary, or separable from the engineering. It is the engineering, the part that determines whether the technical work delivers any value at all. And it connects, strangely, to this fortnight's darker front page: the deepest truth of both is that systems, technical or social, only work when the people inside them actually honour them, use them, hold them. You can build the finest system in the world, or agree the finest treaty, and if the humans do not adopt it, do not use it, do not hold to it when it costs them something, it is worth nothing. The building is the easy half. Getting people to genuinely use what you have built, to make it theirs, to hold to it under pressure, is the hard half, and it is the half that everything actually depends on, and it is the half that everyone, always, wants to skip.

The Ledger
AI
A fortnight dominated by the subcontinental crisis and the papal transition. The models advance in the background; the deeper story, of whether human systems can hold their restraints under pressure, plays out this fortnight on a nuclear border rather than a server.
Data centers & power
Quiet against the geopolitics. The build-out and its power demands continue; the trade war's disruption of chip supplies remains the near-term threat.
Rates
The US economy shrank in the first quarter, a warning sign, as the tariff whiplash bites; the Fed holds, watching a stagflationary fog thicken. Buffett's exit removes a steadying voice from the markets at an unsteady time.
Real estate
US 30-year mortgage 6.76 percent (May 1). Range-bound amid the uncertainty; the American housing market strangely placid while a nuclear crisis builds an ocean away.
India tech
The India-Pakistan crisis is the fortnight's gravest story for the country, overshadowing all else; a full war would be catastrophic for the economy and the technology sector alike, and the world's dependence on Indian software makes the stability of the subcontinent a global concern, however little the world is currently paying attention.
What we called wrong
Nothing to retract. The column on the water treaty extends this paper's oldest theme about escalation: the restraints that prevent catastrophe are precisely the ones that must hold when fury most demands their abandonment, and a line crossed in rage removes a floor that calmer hours will wish had been kept.
The Back Page

The oracle steps down

Warren Buffett announced this fortnight that he will step down at the end of the year, at 94, ending the most remarkable run in the history of investing: 60 years turning a failing textile company called Berkshire Hathaway into one of the largest and most admired enterprises on Earth, and himself into a figure the world came to call, only half in jest, the Oracle of Omaha. This paper noted, a year ago, the poignancy of his first annual meeting without his partner Charlie Munger, the empty chair beside him. Now the man himself prepares to leave the stage, on his own terms, at a time of his choosing, having spent decades preparing his successor so that the enterprise he built does not depend on him, the very lesson this paper keeps preaching about succession and single points of failure. He is handing over not in crisis but in calm, not carried out but walking, having made himself, at last, replaceable, which is the hardest and most selfless thing a founder can do. He leaves behind not just a fortune, most of which he is giving away, but a way of thinking, patient, rational, honest, contrarian, humane, that he taught the world for free, in folksy annual letters that generations read like scripture. He was never the flashiest investor, and that was the point; he was the one who lasted, who compounded, who understood that time and temperament beat brilliance and speed, and who proved it over a span longer than most careers by simply refusing to do the stupid thing, year after year, for sixty years. He is 94, and he is leaving while he is still sharp enough to choose the moment, which is its own final lesson. The oracle is stepping down. The wisdom, thankfully, he already gave away.