Not another tech show

 This Week’s Essay

The best venture and tech content in the world is produced for one audience.

Silicon Valley. Once a sun-drenched stretch of apricot orchards and prune farms south of San Francisco — a place so abundant it produced 30% of the world's prune supply.

The Valley exported itself to the world. The tools, the capital, the culture, the mindset. AI finished the job — what used to require proximity to Sand Hill Road now runs on a laptop in New York and São Paulo. The barriers fell.

And just like that, Silicon Valley became more of a mindset than a place. Marc Andreessen would probably disagree with me on that — and he'd have good arguments. But the evidence is stacking up on the other side. Nubank built the world's most successful digital bank from Brazil. MercadoLibre is the most valuable company in Latin America — a region of 670 million people, nearly twice the combined population of the United States and Canada. CloudWalk built a profitable fintech from São Paulo — ~$1B in revenue, $110 million in net income — and is now live in all 50 US states. The idea that innovation requires a specific zip code is getting harder to defend.

The market for tech and venture content is exploding — but it's almost entirely Silicon Valley-centric. The ICP of almost every major show — All-In, 20VC, BG2 — is the Silicon Valley founder, the Sand Hill Road investor, the US operator. Nobody is connecting the dots between what's happening in global tech and what it means for the founders building and investors deploying capital everywhere else. That translation gap is the obvious opportunity.

A few creators are starting to close that gap. Harry Stebbings has been using 20VC's reach to make the case for European tech — arguing that world-class companies are being built outside the Valley and that the capital and the narrative need to follow.

LATAM has been waiting for the same thing.

That's why we built the Debrief — a monthly tech show where Paulo Passoni and I break down the most important global tech and venture trends and put them in LATAM perspective.

On Tuesday we took the first stab at it and dropped our first 90-minute conversation. The feedback was consistent: deep but light, dense but fun, the kind of show you want to listen to twice. Fingers crossed we'll see immediate product market fit.

A couple of ideas that blew my mind — to give you a sense of what we covered:

Claude 4.6 was the real AI inflection. Bigger than ChatGPT.

ChatGPT gave AI a face. Claude 4.6 gave it hands. A model that answers questions is a better search engine. A model that completes tasks end to end is an agent. That threshold has been crossed. Every founder who's been comfortable saying "it's too early" just ran out of road.

The new recruiting question isn't salary. It's token budget.

The most productive people have already figured out their leverage changed. Token budget is now table stakes for top talent. The org design implications follow: flatter structures, PMs who build, CTOs who architect. Most LATAM companies are still running the 2022 org chart.

The bottleneck shifted from technical to creative.

For twenty years LATAM tech operated under one assumption: the talent is here but the engineering depth isn't. AI just made that irrelevant. Paulo built a fully functioning marketplace on a Saturday afternoon with no coding experience — in three hours. His hands were shaking when it worked. The new scarce resource isn't the ability to build. It's the imagination to know what to build.

The AI moat question.

When everyone runs the same models, product advantages compress to zero. Apart from distribution, the only defensible position is proprietary AI trained on your own data, optimized for your specific business goal. Vertical, not horizontal.

Two examples from companies you know. iFood built a Large Commerce Model on 14 years of Brazilian consumer data — 180 million monthly orders, 10 trillion tokens. Recommendation conversion up 66%. Push notifications driving 4x more orders. Revolut built PRAGMA on 40 billion transaction events from 25 million users. Credit scoring accuracy up 130%. Fraud recall up 65%. One model powering fraud, credit, and cross-sell at the same time.

Both are large incumbents with the data and the culture to move fast. That combination is hard to replicate. Smaller players will need a different playbook — give something away for free, accumulate behavioral data, then use it to build the model nobody else can train. Time will tell if a payment network can actually be free. But the logic is sound.

6 Hot Takes From The Debrief

1. On what's preventing companies from moving fast

What's stopping most companies from adopting AI isn't capability — it's anchor effects. "This is how we used to do things." Most companies haven't even tried the tools themselves. They're still waiting for the CTO to solve it. That's the wrong frame entirely.

2. On the role of the CTO in 2026

The CTO's job is no longer to build with AI. It's to create the architecture that lets everyone else build with AI. Your risk tolerance on cybersecurity should be a one out of ten. Your risk tolerance on an agent saying something wrong to a customer? Eight or nine. You need to be risking it.

3. On venture fund timelines in emerging markets

Venture funds should be 20 years — especially in emerging markets. The entire industry has anchored on the wrong lifetime. You start getting money back year seven or eight. MercadoLibre compounded for a decade before it delivered real value. The path is never a straight line up.

4. On Plata's valuation

Everyone says Plata is overpriced at 7.7x ARR. Nubank raised at 10x ARR all the way to $10 billion. At 100% annual growth, 7.7x becomes 3.5x in a year and 1x the year after. The valuation isn't the risk. The risk is cost of credit at scale. That's the only pre-mortem that matters.

5. On SpaceX and the public market problem

Elon built Tesla in the public eye and it was painful. Constant short sellers, constant pressure, constant need to manage perception. His reaction was rational: never again. Build SpaceX to a trillion in private markets and only IPO after. The problem is ordinary investors never get to participate in the upside.

6. On Nubank's US strategy

Nubank didn't go after Brazilians in Florida like Inter did. They sponsored Inter Miami. They're going after Messi, Beckham, and the affluent American — not the diaspora. That's a completely different brand bet.

What I’m Loving 

Read — PRAGMA: Revolut's Foundation Model (arXiv, April 9, 2026)

Revolut just published the full technical paper behind PRAGMA — the proprietary model we reference in this issue. Trained on 40 billion behavioral events across 25 million users. The paper is dense but the abstract alone is worth your time. This is what a real AI moat looks like in practice.

Listen — Scott Nolan: SpaceX, Founders Fund, and Rebuilding American Uranium Enrichment (Invest Like the Best, EP.467)

Scott Nolan spent 12 years at Founders Fund looking for the most important problems nobody else was funding. Then he found one so critical and so ignored he couldn't find a company to back — so he started one. The conversation covers what Peter Thiel taught him about investing, why being in love with an idea is dangerous for investors but required for founders, and why the single biggest bottleneck to a nuclear future is one nobody is talking about. Essential listening for anyone thinking seriously about where the next decade of infrastructure gets built.

Wildcard — Emma Grede x Jay Shetty (On Purpose with Jay Shetty, November 2025)

Entrepreneurship is relentless. The self-doubt doesn't go away, the path is never clear, and most days nobody is clapping for you. Emma Grede — co-founder of SKIMS, CEO of Good American, one of America's wealthiest self-made women — shares her blueprint on building from nothing. No formal education. No connections. No money. Go listen.

Thanks for reading,

Olga 

 P.S. If this issue was valuable to you please share it with a founder who needs to hear it. Let’s build LATAM’s next tech leaders—together

🎙 The J Curve  is where LATAM's boldest founders & investors come to talk real strategy, opportunity and leadership.