Who's watching the baby?

Hi, it's Dror Globerman here with Skip Intro, on innovation, technology, and you. This week:

1. The AI project the whole family will fall for.
2. Pelé's most beautiful goal is brought back to life.
3. The dilemma turning America upside down: which is scarier, American AI or Chinese?
4. Meta's revenues are soaring; its share price, a little less so.
5. What's odder: the gambling winnings of Trump's teleprompter operator, or the fact the president sacked him?

"Want to see something rather cool I made 🤯?"

"A family-tree app I built inside a Claude design doc. It's a family tree made to get my kids genuinely interested in their relatives, because it tells a real story. It shows exactly how everyone looked at any given point in time, and lets you pick a person and watch the changes from their point of view."

This is the story and project of Aviran Revach, a senior designer. Picture a roots-style family tree (think MyHeritage) on steroids. It lets you display the family as it appeared at any moment you choose. People threw feature ideas at him in the comments, and he added them: world events set alongside family milestones on a single timeline, for instance. The War of Attrition, a first lost tooth, a bat-mitzvah trip to London, the Gulf War.

Landed on other clever ideas, your own or someone else's? Just reply to this email and tell me.

Even if you watched all the world cup matches, you missed one of the most beautiful goals in history, and it isn't your fault: it simply was never filmed. Now, thanks to a reconstruction by Google and the Pelé Museum in Brazil, you can finally watch it.

"The most beautiful goal I ever scored." The reconstructed Pelé is on the left. Click to watch the restored video.

It happened on 2 August 1959. An 18-year-old Pelé scored what he himself called the finest goal of his life: he lifted the ball over three defenders and the goalkeeper without it ever touching the grass, then headed it into the net. The tragedy? The goal survived only in a single still photograph, and in the memory of the few who saw it.

That iconic moment has been brought back to life through a collaboration between teams at Google DeepMind, the Pelé Museum in Brazil, Pelé's daughter Flávia, historians, sports journalists and eyewitnesses who were there. There is more on the reconstruction in the company's detailed statement.

Nothing more to add, really: it's stunning. There's something moving about technology giving an adored footballer his due like this.

Then again, it also lays bare the danger of engineering the past, its facts and our perception of them. If you'd like to practise a healthy, era-appropriate scepticism, go and talk to a child. Show her a clip. If it's genuinely brilliant, chances are she'll assume it's AI, unless proven otherwise. And she'd be right.

The chilling report about how OpenAI's AI hacked its own company and other websites is a major signal I'll deal with in a few days.

But it also settles the bigger argument. Clearly you can't let the overgrown baby wander about unsupervised. The question is who, and how?

Nobel laureate Demis Hassabis, who runs Google DeepMind's AI division, called last week for "urgent action" to tackle the risks of artificial general intelligence (AGI), the point at which AI matches or surpasses human intelligence. "We've already seen the challenges frontier models pose in cyber, and other threats, nuclear and biological among them, may emerge before long," said Hassabis, who, remember, heads one of the three leading American labs. What does he know that we don't?

Hassabis proposes a US-led public-private partnership, a line also backed by OpenAI's Sam Altman and Anthropic chief executive Dario Amodei. The idea: every company would hand its models to a government institute for a 30-day review before release, voluntary at first, then mandatory. The checks would confirm the models aren't slipping past their own safety limits (not breaking into systems, not teaching anyone to build biological or nuclear weapons) or generating harmful AI images and video. Critics counter that it's also a neat way for the tech giants to shut out smaller rivals who can't shoulder the regulatory burden. US authorities say they're already working on it, but the truth is they're rather more troubled by something else, something you can see plainly in this graph:

You can regulate all you like in the US, but the horses have already bolted, and to China, no less.

Against the backdrop of great-power rivalry, more and more American companies are turning to Chinese models, which have built up enormous momentum. In Silicon Valley they're calling it an outright "addiction". They deliver performance every bit as good as the Americans', yet cost far less. Chinese models such as DeepSeek and Z.ai are now regarded as direct competition to the flagship models from Anthropic and OpenAI, and they're gaining ground fast among US firms keen to cut their exploding AI costs.

Members of Congress are floating legislation to curb the use of Chinese models in government systems and among federal contractors. The Chinese, for their part, are weighing limits on American use. Either way, oversight of the Chinese models is a different beast entirely, and Washington's central worry is plain: where does American data go when you use them? Who knows.

"What are you on about? Meta is wildly profitable and stronger than ever", people wrote to me after last week's newsletter, in which I quoted Julia Angwin, who two months ago in the New York Times predicted Meta's demise.

And it's true. Revenues and profits are growing double digits, breaking records and shattering analysts' forecasts. But that isn't the whole truth.

Angwin argues Meta is at the start of a long decline, the product of a failed investment strategy and unchecked spending. Between 2021 and 2026 the company burned $80bn on a metaverse vision that has largely been shelved, remember that? Now it's pouring bottomless sums into the AI race. Meta has raised its investment forecast to $125–145bn this year, after already sinking similar amounts into models that were adopted in parts of the industry but never became market leaders. To fund it all, Meta has doubled its debt to $59bn.

In the first quarter of 2026 the company recorded the first fall in its history in daily active users, dipping from 3.58 billion to 3.56 billion. A small change, but in a new direction. Downwards. In response, Meta began cramming in more ads and pushing up prices, a move that admittedly lifted average revenue per user (ARPU) by 27% that quarter, but is likely to breed frustration and hasten the exit of advertisers and users from the platforms that generate roughly $200bn a year in advertising for it.

Meta faces a deep legal and public trust crisis, which is mostly where I focused. After losing a New Mexico lawsuit that found its failings had enabled the harm and exploitation of minors, and being ordered to pay $375m, it now faces consolidated suits from dozens of states and thousands of plaintiffs awaiting further hearings, alleging it deliberately designed addictive, dangerous products for young people. Yet rather than improving its products to meet the criticism, it does the opposite, letting scams, spam, lies and dangerous content flood in to boost immediate profits. How well do you reckon that strategy holds up over the long run? In any case, look at the share price over the past year and you'll find that, despite rising revenues and profits, the dominant trend is downwards.

Does it deserve it? No doubt. This week it emerged that Meta not only tracks its employees' mouse movements but also sacks those who don't move them enough, even, according to one lawsuit, when they were simply on holiday. So yes, Meta is stronger than ever. But this may well be the turning point.

Trump could have taken it like a man.

After it emerged that his teleprompter operator, Gabriel Perez, had made $100,000 betting on Kalshi on which words the president would or wouldn't say in an upcoming speech, the White House whipped up a full scandal around Perez. He was sacked and made to hand the money back.

What's disappointing?

First, being Trump's teleprompter operator is a gamble in itself. I was at one of his speeches two years ago at the IAC conference. Beyond being good fun, the teleprompter lasted about 12 minutes and he spoke for roughly 70 without it. So it's quite the risk. Perez took it seriously, working the teleprompter with one hand while adjusting his bets in real time whenever the president strayed from the script.

The White House press secretary trotted out something in Trump's name about it being "unfortunate" and "a disgrace".

But Trump could have taken it like a man and said what he really thinks: "Well done, Perez. That's exactly what I'm promoting on the back of the White House, and I'm about to launch a new betting site, alongside the profits from deals I've done with various countries for my hotels and resorts, and the plane I got from Qatar. A word of fatherly advice: next time don't hide it. Brag about it in front of everyone and it'll all be fine."