Retrouvez l'épisode en version française ici : https://www.gdiy.fr/podcast/yoni-assia-vf/

On January 28th 2026, he launched his first AI agent.

By April, Yoni Assia had 30 active agents running 24/7 on WhatsApp, 45 GitHub projects, and a dev team that couldn't review prototypes faster than non-developers were shipping them.

In addition, Yoni runs eToro, a 1,500-person company he co-founded in 2007 with his brother and a friend.

The mission hasn't changed since day one: democratize access to financial markets.

It runs in the family.

His grandfather opened a Swiss bank and his father built Magic Software, one of the first low-code development platforms.

Obsessed with financial markets since he was 13, Yoni started trading early.

He lost 90% of his portfolio in the dot-com crash, and came back obsessed with building the infrastructure that would make finance work differently.

His first company put video cameras on roller coasters. It was sold to Kodak. Then came eToro.

In 2012, he wrote one of the first papers on tokenizing real-world assets using Bitcoin.

He posted it on BitcoinTalk and a teenager named Vitalik Buterin answered. Yoni paid him to co-write the paper, brought him to the eToro offices, and watched him leave to build Ethereum.

Today, eToro manages $20 billion in assets across 50 million registered users, generating $870 million in revenue and $320 million in profit.

Then came January 28th.

Yoni launched his first OpenClaw agent. He hasen't slept properly since.

He dreams about them, wakes up at 3am, opens WhatsApp, and finds his agents have been talking to each other all night, building, testing, shipping without him.

In this episode, Yoni shares everything he's learned from building eToro to the last months, running a company with AI agents at its core:

Why non-developers at eToro are shipping faster than the dev team can review
How to structure agents so they don't hallucinate, forget, or go rogue
The Security Claw, the Memory Layer, and why replication on the cloud changes everything
How eToro just launched an Agent Portfolio so your OpenClaw can trade directly in your account
Why will most of the world's money be managed by agents within 12 to 24 months
The biggest mistakes people make when they start with AI agents and how to avoid them

The most experienced builder in the room isn't a developer.

It’s the founder of a fintech who gave his quant team a thousand PhDs and told them to beat Renaissance Technologies.

You can contact Yoni on Linkedin.

TIMELINE:

00:00:00 - The agent that built its own team
00:11:46 - The startup inside every roller coaster in the world
00:18:31 - Was Vitalik really a genius?
00:30:30 - Non-devs building what the dev team can't review fast enough
00:38:50 - How the entire banking industry work
00:52:35 - Why banks don't want you to understand capital markets
01:00:15 - The one feature no bank will ever build
01:17:49 - The fintech competing with banks that don't want to compete
01:29:01 - eToro business model explained in 5 lines
01:36:35 - When every trader has an agent, who wins?
01:47:04 - Managing a 1500-person company with WhatsApp
01:58:16 - The biggest mistake with AI Agent

We referred to previous GDIY episodes :

#500 - VO - Reid Hoffman - LinkedIn, Paypal - How to master humanity's most powerful invention
#500 - VF - Reid Hoffman - LinkedIn, Paypal - Comment dompter l'invention la plus puissante de l'humanité
#473 - VO - Brian Chesky - Airbnb - « We're just getting started »
#473 - VF - Brian Chesky - Airbnb - « Après 17 ans, nous ne sommes qu'au début de notre histoire »
#452 - VO - Reid Hoffman - LinkedIn, Paypal - "We are more Homo technicus than Homo sapiens"
#452 - VF - Reid Hoffman - LinkedIn, Paypal - L'humanité 2.0 : Homo technicus plus qu'Homo sapiens

A few recent episodes in English :

#513 - VO - Jesper Brodin - IKEA - 40 billion in revenue empire with no bank loan
#500 - Reid Hoffman - LinkedIn, Paypal - How to master humanity’s most powerful invention
#487 - VO - Anton Osika - Lovable - Internet, Business, and AI: Nothing Will Ever Be the Same Again
#475 - VO - Shane Parrish - Farnam Street - Clear Thinking: The Decision-Making Expert
#473 - VO - Brian Chesky - Airbnb - « We're just getting started »
#452 - VO - Reid Hoffman - LinkedIn, Paypal - L’humanité 2.0 : Homo technicus plus qu’Homo sapiens
#437 - James Dyson - Dyson - “Failure is more exciting than success”
#431 - Sean Rad - Tinder - How the swipe fever took over the world

We spoke about :

Reading Recommendations :

Nineteen Eighty-Four (1984), by George Orwell 🇬🇧
1984, de George Orwell 🇫🇷

Interested in sponsoring Generation Do It Yourself or proposing a partnership ? Contact my label Orso Media through this form.

Hébergé par Audiomeans. Visitez audiomeans.fr/politique-de-confidentialite pour plus d'informations.

#542 - VO - Yoni Assia - eToro - “AI Will Replace Most Traders in 18 Months”

Génération Do It Yourself