Hey guys, Non here, coffee in hand. Want to hear some interesting stories?
Let’s start with the weather. According to The Climate Issue, a single 47-degree Celsius day in Barcelona last week cost the city €3.6 million in emergency cooling services alone. That is the story that matters today because it is the first time a European capital has priced out a single day of heat so publicly.
The more important question is how the bill for these temperatures gets paid when the climate is just getting started. First, the urgent question: who will replace Karoline Leavitt as the White House Press Secretary? Then we’ll talk about the surprising financial cost of Europe’s heat, how Google is making private AI practical with homomorphic encryption, the dozens of countries that reportedly helped China dodge Trump’s tariffs, and a final word on Nirmal Purja. Let’s go into it.
The most reliable signal I have from Washington this morning is not a leak or a whisper, but a resignation letter. Karoline Leavitt, the young press secretary who has been Donald Trump’s face in the briefing room for the last year and a half, is stepping down at the end of August. According to the BBC, she’s been with the president for years and is one of his "most trusted aides," but the job of managing a press office that handles thousands of inquiries a day—while managing the president’s temper and the administration’s contradictions—has exhausted her. The White House has not named a successor.
The popular narrative in the press is that this is a battle for the crown jewels. You hear the name Alina Habba a lot—Trump’s former lawyer, the woman who sat in a courtroom in New York for months while he was convicted on 34 counts of falsifying business records. You hear Katie Miller, the wife of Stephen Miller, who helped design the administration’s immigration crackdown. You hear Steven Cheung, the communications director known for his combative style.
The story is framed as a fight between hardline loyalists, a way to keep the messaging machine running at full speed after a difficult legal stretch. The implication is that the new press secretary will be a mirror image of the old one. That is a convenient story because it fits the image of a regime digging in, but it misses the machinery of the job. According to the BBC, when Leavitt was on leave earlier this year, the White House did not hire a temporary substitute from her office.
Instead, Vice-President JD Vance and Secretary of State Marco Rubio stood in at the podium. This suggests the "briefing room" function is actually a flexible pool of senior officials, not a single person’s job. The secretary is the face, but the real work is coordination with the other agencies—State, Commerce, Energy—who have to turn complex policy into soundbites that won’t get the president sued. If you look at the list of potential replacements, you find people who are already deeply embedded in the administration.
Anna Kelly, the deputy press secretary who is 29 years old, has been on the ground running the office while Leavitt was away. She understands the rhythm of the requests coming in and the relationships on the Hill. Sage Steele, the former sports anchor, is a media personality who knows how to appear on television, but she has no experience managing a daily press operation or the bureaucratic friction of a sprawling government. The real competition isn't for the "face" of the presidency; it is for the person who can translate the president’s chaotic impulses into a brief that doesn't collapse under legal scrutiny.
The gap here is between the image of a loyalist replacing a loyalist and the reality of a machine that has to keep running. The White House is not replacing a figurehead; it is looking for a manager who can survive being watched every day. That is a harder job than being loud.
We just talked about the difficulty of managing a public image and the institutional weight of the White House Press Secretary role, but now we’re moving from the architecture of political communication to the literal architecture of European cities as they face physical limits under extreme heat. A report by The Economist’s climate newsletter highlights how Europe is currently grappling with an unusually hot and dry summer that is beginning to stress the continent's basic infrastructure. This isn't just about people feeling uncomfortable while walking to the store; it is a systemic strain on the systems that keep modern life running, from electricity grids to water management. To understand why this matters, you have to look at how our cities were designed.
Most of Europe’s urban centers were built during a period when "extreme" weather was a statistical outlier. The buildings are made of stone and brick—materials with high thermal mass that soak up heat during the day and radiate it back out at night. This creates an urban heat island effect where the city stays hot long after the sun goes down, preventing the cooling that happens in rural areas. When you combine that with a lack of sufficient green space and modern ventilation, the city essentially becomes a radiator.
The mechanism of the crisis is tied to two specific constraints: power demand and water availability. On a blistering day, everyone turns on their air conditioning simultaneously. This creates a massive, immediate spike in electricity demand that can overwhelm local grids, potentially leading to rolling blackouts or equipment failure. Simultaneously, high temperatures cause increased evaporation from reservoirs and rivers.
In many parts of Europe, this means the water used for cooling those same power plants—or for industrial processes—becomes scarce. You end up in a loop where you need more power to stay cool, but the heat makes it harder and more expensive to generate that power safely. Someone will say that we can just build our way out of this by installing more AC units and bigger power plants. They might argue that the solution is simply a matter of capital investment: spend the money, install the tech, and the problem disappears.
It’s a logical stance because it treats heat as a simple engineering hurdle rather than a structural one. But there is a physical limit to that logic. If you put more air conditioners in a city, you increase the heat output on the streets below, which requires even more cooling power from the units next door. It's a cycle of diminishing returns.
Furthermore, building more large-scale infrastructure often takes years or even decades of planning and construction, while heatwaves are becoming more frequent and intense. The time it takes to build a solution is starting to exceed the interval between the problems. What survives this tension is a shift in how we measure urban success. We see this in the way cities are beginning to re-evaluate "grey" infrastructure—concrete and steel—versus "blue" and "green" infrastructure.
In places like Paris or Madrid, there is an increasing focus on permeable surfaces that allow water to soak into the ground rather than running off into sewers, and the strategic planting of trees to provide literal shade. The scale of this shift is visible in the numbers. While a single air conditioning unit might cost a few hundred euros for a homeowner, the cost of retrofitting an entire city’s cooling capacity can run into billions. For instance, if a municipality has to choose between subsidizing individual AC units or spending that same money on a massive urban reforestation project, they are choosing between a temporary fix for the resident and a long-term structural change for the city.
We are seeing a move toward the latter because it addresses the root cause: the fact that our current cities were built for a climate that no longer exists. The question we're left with is whether these measures can be deployed fast enough to keep pace with the rising mercury.
While Europe is trying to engineer ways to physically cool down a warming planet, Google is working on a different kind of cooling—trying to keep the "heat" of data collection from melting user privacy. We have moved from the atmospheric to the algorithmic. Google just released something called HEIR. It is an open-source compiler designed to let companies run AI models on encrypted data using a technique called homomorphic encryption.
To put it simply, if you send your medical records or bank statements to a cloud server for analysis, usually that server has to "unlock" the file to read it. With this technology, the server can perform the math on the scrambled ciphertext and send back an encrypted result, all without ever actually seeing what was inside the original message. The reason this matters is a fundamental bottleneck in how we use AI today. If you want a service that detects spam or identifies diseases, the provider usually needs to see your data.
But if the data is so locked down that the provider can’t see it—like end-to-end encryption—they can't provide those features. The alternative has been moving everything to your local phone or laptop, but that limits what the AI can do and risks leaking the company's secret software code. Homomorphic encryption tries to break that trade-off by turning privacy into a math problem rather than a physical one. Now, someone will say this is just a laboratory fantasy because it’s too slow.
And they aren't entirely wrong. Historically, running these types of calculations has been thousands of times slower than normal computing—basically asking an AI to do its job while wearing a lead suit. If the electricity bill for one request costs more than the value of the answer, nobody is going to use it. But Google’s move with HEIR is an attempt to make that "lead suit" wearable by creating a toolkit that handles the complex cryptography automatically.
They are partnering with hardware companies like Intel and Marvell to build chips specifically for this math. In their initial tests on single-threaded CPUs, they're already showing latency numbers for four different private inference applications. The goal isn't just better privacy; it’s trying to make it cheap enough that a hospital can use a cloud AI without ever technically "sharing" a patient's name with the cloud provider. We are moving from a world where you have to choose between convenience and secrets, to one where the math might eventually let us have both.
We were just talking about how you can hide the ingredients of a recipe without burning down the house, and now we are going to look at a whole country doing exactly the same thing, except the ingredient is a tariff and the house is the US economy. The White House released a report this week showing that more than 40 countries helped China sidestep American trade taxes by routing goods through nations with lower duties. The countries listed include Canada, India, Mexico, Japan and South Korea. The White House estimates that between 30 billion and 300 billion dollars’ worth of goods have been moved this way.
They call it transshipping—moving cargo through a third country to hide where it really came from—and they say China repackages the goods to make it look like they are from a place with cheaper import rates, calling the whole operation “fraud cloaked in paperwork.” To see why this is happening, you have to look at the incentives, not the people. The US raised tariffs on Chinese goods during President Trump’s term, with the stated goal of protecting American jobs and revenue. The problem is that tariffs are a tax on the importer, not the country of origin. If Canada, for instance, does not have a tariff on that type of widget, a Chinese company can send it to a warehouse in Toronto, slap a "Made in Canada" sticker on it, and ship it across the border without paying the higher American duty.
The company saves the tax, the American government loses the revenue, and the goods eventually end up in American homes. According to the White House, this has cost the US billions in revenue and they claim it has cost American jobs. The obvious objection is that this is not China’s fault. Canada, India, Mexico, Japan and South Korea have all denied helping China dodge tariffs, and they point out that these routes are also used by American companies moving goods through friends and allies.
But the White House argues that China’s scale makes the difference, and that the network is now so sophisticated that it looks like a deliberate effort to exploit the system rather than a byproduct of normal trade. The thing that survives both sides of this argument is the fact that the US and China are about to sit down for high-stakes talks in Washington in September. Chang Pao Li, an economics professor at Singapore Management University, says the report gives the US a specific bargaining chip: Washington can argue that China has preserved access to the American market through third countries, and that any deal must account for these transshipment routes. At the same time, economies with deeper ties to Chinese supply chains—like the nations named in the report—now face the risk that the US could change the rules and demand they police trade more strictly.
It feels like we are stuck in a game of telephone where the tariff is the message and every country in the middle is just repeating it, but louder.
Before the last story: if you are getting something out of this, subscribe. The World in Twenty Minutes is on Spotify, on Apple Podcasts, on iHeartRadio, and anywhere else you already listen. It is free, it lands every morning, and subscribing is the whole reason it keeps finding people.
The story we just left was about nations bending the rules to dodge a tariff, and it felt like a game of telephone where the message got louder every time it passed through a new country. It was about the system bending to protect a price. The Gurkha mountaineer Nirmal Purja died in an avalanche on July 30th, aged 43. According to The Economist, he performed momentous feats in the Death Zone.
That is it. That is all I have. The Economist does not tell us what those feats were, and a headline cannot tell us what a human being did with their life or what they meant to the people who watched him on Netflix. The line I have is a skeleton, and I am going to leave it that way, because this story is not about the mountain, it is about the belief behind the climb.
You might look at a headline like that and assume it is just about physical limits—how high can a body go before it gives up? But the more important question is why anyone would try to go there at all. The Death Zone is the part of a mountain above eight thousand metres where the air is thin enough to make you dizzy and the cold is enough to make you forget who you are. The incentive is simple: you do not go to the Death Zone to survive.
You go to prove that a rule does not exist. You go to prove that a limit you were told is biological is actually just a question of will. Purja was not the first person to reach the top of a peak there, but the claim that he did it in record time was the refusal to accept that some obstacles cannot be overcome. That is what survives both the headline and the avalanche.
It is not about a watch or a flag on a summit. It is about the idea that if you keep asking for the keys, eventually the lock breaks. The last story ended on the idea that institutions will just shout louder until the problem goes away. Purja tried to shout louder at the mountain itself, and the mountain answered back on its own schedule.
We do not know what his last words were, and we do not know what he felt in the final meters. All we have is the belief that the game can be won. Be honest about the limits of what you know, because the honesty itself is the point. Take care.
Dr Non.