Pietro Benedetti Italian · 25 years old

A dating app where you never scroll through profiles. Instead, you chat with an AI for a few days — it learns how you think, what you value, what you're actually looking for — and then it finds your match. No swiping, no games. Just a conversation and a name.

Upload your conversations with someone — a friend, an ex, someone you lost. The AI learns how they write, how they think, how they respond. And then you can keep talking to them. It's not them. But it's close enough to matter.

Most investment tools are built for people who already know what they're doing. This one isn't. Answer a few honest questions about your life, your goals, and how you'd feel losing money — and get a clear, personalised picture of where your first savings should go.

The AI industry looks like a contest of rivals — OpenAI vs Anthropic, Microsoft vs Google. This paper argues that framing is misleading. Beneath the visible competition, every model ever built funnels through the same foundry (TSMC, ~92% of advanced chips), the same lithography machines (ASML, sole supplier), and the same high-bandwidth memory (SK Hynix, ~62% of shipments). Add to that a financial layer where suppliers fund the customers who buy their own products — Nvidia into OpenAI, Amazon into Anthropic — and what looks like plurality turns out to rest on a handful of shared nodes. The appropriate unit of analysis is not the firm but the network. And networks with few shared organs have few places to absorb a shock.

Open PDF ↗

150 years of Italian emigration, from 1876 to today, built into an interactive timeline of six distinct waves — each with its own set of destinations. Argentina, Brazil and the US dominated the first wave. The postwar boom went to Germany, Switzerland, Belgium. Today it's mostly the UK, Germany, Switzerland, Spain. Same country leaving, almost never the same map.

Open video ↗

Using microdata from the 2023 ISTAT household survey (34,968 adults), I find that separated and divorced Italians face 69% higher odds of being current smokers than married people — after controlling for age, sex, education, income, and geography. The result holds across four model specifications. Whether divorce causes smoking, or smoking causes divorce, the data can't say. But the association is real, large, and hard to ignore.

Open PDF ↗