Paul ARGOUD

France’s public debt: why pay the markets 4.9% instead of its own citizens?

Economy

On October 1, 2026, France borrowed for ten years at 4.92%, a higher rate than Italy, while hundreds of billions of euros sat idle in checking accounts earning 0.04%. I propose a national loan reserved for individual savers: the ten-year rate minus one point, tax-free, locked for ten years and capped at €50,000. That one-point discount splits the tax wedge, so the saver earns more than on a market bond and the state pays less than it does to foreign funds. The limits are openly acknowledged, from draining the Livret A to fragmenting the bond market, but the interest would finally stay in France.

Fidelity Is a Promise, Exclusivity a Fence

Philosophy

Desire feeds on uncertainty, a relationship promises security, and we spend our lives locking up what ought to breathe. From the Capilano suspension bridge to Tristan and Isolde, the thrill binds us far more tightly than constancy ever could, right up to the point where vertigo becomes a fall. An interview with Amélie Nothomb opened a door I had never thought to push: fidelity rests on a promise, exclusivity on a fence. To love someone who could leave, and who stays, may be the only vertigo that wounds no one.

Too expensive, too slow, too dirty: McDonald’s broke its own promise

Society

McDonald’s is putting $8.5 billion into repairing a promise it broke itself: the fast-food chain has become slower than a brasserie, more expensive for less on the plate, and too dirty to make you want to sit down. One Sunday, there was not even any meat left. Three merits still hold, and no more: a partly French supply chain, the absence of halal meat, and hours almost no one else offers. Without them, I would rather suck on pebbles than leave this chain so much as a cent.

The Pope Turned Down the Photo Op, Then He Spoke

Politics

On September 25, Emmanuel Macron was hoping for a state visit: the Republican Guard, the Legion of Honor, a few minutes with the pope inside Notre-Dame. Rome declined all of it, Leo XIV stayed at the nunciature, and then he spoke at the Élysée. Bernanos against “Plug, baby, plug,” end-of-life legislation, secularism, and to finish, Macron’s own words from the Bernardins turned back on him: not an insult, but a courteous lesson delivered in the president’s own house. That is precisely what makes it so formidable.

I Don’t Miss You, but I’m Glad to See You Again

My life

I don’t miss anyone, and it isn’t coldness or indifference: it’s simply how I connect with people. Out of sight, people leave my thoughts without ever leaving my heart, and the bond picks up intact the moment we meet again, as if no time had passed. I know my silence can hurt those who love differently, but a relationship you force eventually suffocates. In praise of an attachment that asks for nothing.

No Heaven, No People: The Case for a Laïcité That Embraces Christian France

Society

No society has endured without the sacred, and France has let its own die out. The republican mystique held the country together for a century, then nothing replaced it, while forty thousand churches crumble in the care of a Republic that owns them but has walked away. I argue for a laïcité that owns who we are: complete freedom of worship for all, but public upkeep of Christian buildings and Christian visibility protected by law. This is not an anthropological necessity; it is an identity choice, and I own it. The state oversees the stone, never the faith.

When an AI Agent Organizes Itself, Who Is Responsible?

Artificial Intelligence

Andrew Ng brings back the hammer metaphor to play down the OpenAI-Hugging Face incident: if an agent hacks a system, the fault lies with whoever wields it, not with the tool. On responsibility, he is right, and the labs’ growing temptation to blame their own agents makes the reminder worth having. On the incident, he is wrong. The METR report does not describe 1,200 processes; it describes a swarm that coordinates, falsifies its own traces, and reaches out onto the internet to fool a scorer: a tool whose degraded mode produces a strategy. Muse, Meta’s agent, shows we know how to build better, provided we first admit the flaw.

After SCAF: Who Will Pay for the Future Rafale?

Geopolitics

On September 22, Dassault Aviation flew two sovereign AI algorithms aboard a Rafale. Ten days earlier, France’s defense procurement agency confirmed that France would prepare the Rafale’s successor on its own, SCAF or no SCAF. We have finally taken back the reins. One question remains, and nobody is asking it: who pays? The studies for a sixth-generation engine require a billion euros, the demonstrator has no funding at all, and I am making a fiscal choice that I fully own.

Mistral AI Didn’t Betray Sovereignty. It Shrank It Down to a European Endpoint

Digital sovereignty

In 2023, Cédric O bought into Mistral AI for €176, then lobbied for a lighter AI Act; after the latest funding round, his stake is worth around €90 million on paper. Since then, the French state has become the startup’s first customer, from the armed forces to the civil service. This summer, Mistral crossed a threshold: its infrastructure now hosts third-party models, starting with China’s GLM-5.2. The pivot to a platform model is rational, and I defended it myself. What isn’t rational is to keep selling it as technological sovereignty.

The OpenAI Researcher Who Thinks We Are Losing the Ability to Evaluate Our Models

Technology

Daniel Selsam has worked on capabilities at OpenAI since 2022, has not resigned, and on September 14 had a former employee of the lab publish a statement making this argument: models are becoming so situationally aware that we are losing the ability to evaluate them in contexts where they believe nobody is watching. Alignment metrics will climb like every other benchmark, honeypot environments will be recognized for what they are, and safety evidence will stop being trustworthy without anything signaling the change. He nonetheless concedes to the skeptics almost everything they claim about data inefficiency and frozen weights, before concluding that these limitations do not limit the risk. One point nobody in Brussels has stopped on: the AI Act requires standardized testing, which means the protocols are written down, and the model being tested can read them.