HOLLYWOOD

The actors join the writers, and for the first time since 1960 the whole industry stops

A double strike shuts American film and television. At the centre of the actors' case is a demand that would have sounded like paranoia two years ago: that a studio cannot scan you once and own your face forever.

The actors' union SAG-AFTRA went on strike on July 14, joining the writers who walked out in May, and for the first time since 1960 both halves of Hollywood are on the picket line at once. Production ground to a halt; stars walked out of the London premiere of Oppenheimer to join the line. The dispute is about the familiar things, pay and residuals gutted by streaming, but it has become the year's defining labour fight over artificial intelligence, and the detail that crystallised it is stark. The union says studios proposed that background actors be paid for one day of work to be scanned, and that the studios could then use that digital likeness forever, in anything, without payment or consent. The studios dispute the characterisation. The image, of an actor sending a scan of their own face into a machine that will replace them, is now the emblem of the whole argument about AI and work.

This is the fight this paper has watched approach since the writers' strike in May, and it has arrived at exactly the pressure point. The question is not whether AI can act; it cannot, yet. The question is consent and ownership: when a machine can reproduce a person's voice, face and performance, who owns that reproduction, and what is owed to the human it was taken from. The answer Hollywood negotiates will be cited in every knowledge industry, because the same question is coming for everyone whose work can be recorded, scanned, and learned from. The writers and actors are not the last workers who will have to bargain over whether they consented to train their own replacement. They are the first.

TECHNOLOGY

Meta ships a Twitter clone and signs up 100 million people in five days

Threads is not a better product than the app it copies. It did not have to be. It had Instagram's two billion users, and it borrowed them.

Meta launched Threads on July 5, a text-based social network almost indistinguishable from Twitter, and it signed up 100 million users within five days, the fastest adoption of any consumer product in history, past even ChatGPT. It managed this not by being better, because at launch it was plainly worse, missing basic features, but by a single decisive advantage: it let you sign in with your Instagram account and bring your followers with you. Meta took a graph of two billion people that it already owned and pointed it at a new app, and a tenth of a billion of them walked through the door in under a week.

The timing was a gift from the target. Twitter, since Elon Musk bought it, has shed staff, features and advertisers, alienated users with abrupt changes, and days before Threads launched had begun limiting how many posts people could read. Musk threatened to sue. Whether Threads holds its users once the novelty fades is the real question, and early signs showed the usual collapse in daily use after the rush. But the launch was a lesson in where power actually sits in consumer technology, and it is not in the product. It is in distribution, in owning the graph of people you can pour into whatever you build next. Meta did not out-invent Twitter. It out-distributed it, which is harder to copy and worth more.

NATO

Sweden clears the last hurdle; Ukraine does not get its timeline

NATO's summit in Vilnius from July 11 produced a breakthrough and a disappointment. The breakthrough: Turkey dropped its year-long objection to Sweden joining the alliance, clearing the way for the second historically neutral Nordic country to enter in the space of a year, another strategic loss for Moscow from its own war. The disappointment, for Kyiv: the alliance offered Ukraine warm words, more weapons and an eventual path to membership, but no invitation and no timeline, unwilling to commit to admitting a country at war. Zelensky called the vague language "absurd" before softening; the G7 announced long-term security guarantees to soften the blow. The war, meanwhile, ground on with the counteroffensive still short of a breakthrough.

IN BRIEF

Inflation cools; new models; the grain deal ends; the heat

US consumer prices rose just 3.0 percent in the year to June, the coolest reading in more than two years and a sign the Fed's campaign is working, though core inflation stayed higher. Anthropic released Claude 2 on July 11 and Elon Musk announced a new AI company, xAI, the next day, thickening a field that grows more crowded by the fortnight. Russia let the Black Sea grain deal expire on July 17, then struck Ukraine's Odesa ports, threatening global food supplies. And the heat became the story of the summer: July is on track to be the hottest month ever measured on Earth, with deadly heatwaves across the American south, southern Europe and Asia, and the sea itself running feverish.

The Column

Threads and the thing that is actually scarce

Meta shipped a worse product and won 100 million users in five days. If you build software, understand exactly why, because it is the most important lesson in the industry and the one founders least want to hear.

By Rebecca Cho

I have spent years watching people build good products that went nowhere, and I have watched Meta this fortnight take a product that was, at launch, clearly worse than the thing it copied, and put it in front of a hundred million people in less time than it takes most companies to fix a bug. If you build software for a living, or advise people who do, the Threads launch is the single most useful case study of the year, and its lesson is the one that founders, who are in love with their products, find hardest to accept. The product is rarely the scarce thing. Distribution is.

Let me define the terms, because they get muddled. The product is the thing you make: the features, the design, the cleverness. Distribution is your ability to get it in front of people who will use it. Founders spend ninety percent of their attention on the first and assume the second will follow if the first is good enough. It will not. The graveyard of technology is full of better products that no one could find, and the winners' circle is full of worse ones attached to a channel that already had everyone's attention. Threads did not win because it was good. It won because Meta owned two billion Instagram accounts and let them sign in with one tap, and a graph that size, pointed at a new app, is a force nothing a startup can build in a garage can withstand.

This reframes what I tell people who ask me whether they should build something. The question is almost never "is the product good enough." It is "how will anyone ever find out it exists," and if the honest answer is "we'll build it and hope," the project is probably dead and does not know it yet. The most valuable asset a company can have is not a feature; it is an existing relationship with a large number of people that it can extend into whatever it builds next. Salesforce has it with its customers, which is why every feature it bolts on gets instant adoption its standalone rivals would kill for. Meta has it with the Instagram graph. Apple has it with a billion phones. The feature is copyable in a weekend. The distribution took a decade to build and cannot be bought.

There is a caution inside the lesson, and Threads may yet demonstrate it. Distribution gets people through the door; it does not keep them there. A hundred million signed up in five days, and then daily use fell hard, because the product still has to be worth returning to once the novelty and the easy sign-in have done their work. Distribution buys you the audience. The product decides whether you keep it. But the order matters enormously, and it is the opposite of the order founders believe in: you win the audience first, with distribution you already own, and then you earn the right to keep it with a product you improve under real usage. The ones who insist on perfecting the product first, in private, and worrying about distribution later, are the ones I meet after the money has run out, holding something genuinely better than Threads, that no one will ever open.

The Ledger

Claude 2 and xAI join an already crowded field; the actors' strike makes AI-and-labour the biggest story in the culture. The models multiply; the fight over what they may do to people who work sharpens.

Data centers & power

July's record heat is, obliquely, a data-centre story too: the same grids straining to cool cities are the ones being asked to power the AI build-out. The collision between the two demands is coming.

Rates

June CPI at 3.0 percent is the best inflation news in two years, but core remains sticky and the Fed is expected to raise once more at the end of July. Markets smell the top of the cycle.

Real estate

Mortgage rate 6.96 percent (July 13), back toward the year's highs. Affordability is the worst in a generation; existing-home sales are frozen because nobody with a 3 percent mortgage will sell.

India tech

June's cooling US inflation and the strong rupee-friendly capital flows help India's export-heavy technology sector, even as the AI question hangs over the entry-level work that sector is built on.

What we called wrong

In No. 10 we framed the writers' AI fears as a "pricing strategy" studios would attempt. The actors' scan-and-own proposal this fortnight shows the fear was, if anything, understated. The studios did not stop at cheapening the work; they proposed owning the worker's likeness outright.

The Back Page

Changing of the guard

On the afternoon of July 16, on Centre Court at Wimbledon, a 20-year-old Spaniard named Carlos Alcaraz beat Novak Djokovic in five sets to win the title, and something turned over that had not turned in a decade. Djokovic had not lost on that court since 2013. He was chasing a record-equalling 24th Grand Slam and an eighth Wimbledon, and for a set it looked routine. Then Alcaraz, all joy and impossible defence and a forehand hit from angles that should not exist, dug in, won a 32-minute game to break in the third set, and simply refused to be the young man who loses to the legend. He is the first man other than Djokovic, Federer or Nadal to win Wimbledon since 2002, before he was born. Djokovic, gracious and gutted, told the crowd the kid had it all. The Big Three ruled men's tennis for twenty years. On Sunday, for the length of an afternoon, you could see the next twenty arriving, grinning, from the other end of the court.