Reflections and analysis on technology, innovation, and the digital future. Exploring trends that are transforming our industry.
One at the Santander offices about MCP in financial services. Another about how I built eleven products on my own. Two different subjects with the same conclusion: what gets automated is the work, not the accountability.
Being an Anthropic partner opens direct contact with them for our clients. But it doesn't marry us to anyone: an orchestrator that only knows how to talk to one provider isn't an orchestrator, it's a reseller. And what struck me most about their offices was the security.
Hiveflow was selected by the Latino Business Action Network for its program at Stanford. I graduated at the Knight Management Center in May 2026. The most useful thing wasn't what I learned: it was finding out how much of what I already knew, I had learned the most expensive way possible.
In April 2026 Talent Land moved to Mexico City and I took the Developer Land stage to talk about three ways of working that already coexist: agents working for us, with each other, and with humans. That's where I presented SwarmMind.
I came back from NVIDIA GTC with two professional certifications and a machine for running models locally. It wasn't tech tourism: if you bet on open-source LLMs living in the client's own building, hardware stops being procurement and becomes part of the problem you solve.
On January 31, 2026 I won the Most Innovative Use Case award at the Prompt Driven and Toolhouse hackathon, held at the 500 Global offices in Palo Alto, with SwarmMind: a Kanban board where the workers are AI agents. What I presented that day became Hive Protocol.
On September 26, 2025 I won the Decentralized AI Hackathon run by Flower Labs at Stanford, with one day to build and more than 200 participants in the room. My project, FedLLM Studio, had eight institutions training a model together without sharing a single piece of data.
At the Genius Arena Hackathon by Capital One, inside Talent Land 2025, we built InversoAI in 24 hours without sleep and made the top three: a finance tutor that gets to know you first and teaches you second, and that ends in real instruments.
Cell-based, microbial and insect proteins; and an artificial intelligence that no longer searches for proteins but designs them. This is where BioQuantech came from. Talk at Talent Land 2025.
Two Sigma, Numerai and WorldQuant have been deciding with models rather than by hand for years. What AI does well when investing, where it fails, and why the work left for people is deciding where the system must stop.
On the Developer Land stage at Talent Land 2025 I explained how ideas from genetics help build AI agents that improve on their own: an agent's prompt, tools and limits are its genes, and they can be left to evolve.
Problem detection went from weeks to minutes, a working prototype fits in 25 minutes and validating with real users dropped from months to days. That changes who gets to start a company. Talk on the AI Land stage at Talent Land 2025.
The last day of the series is the one that explains the other eleven: Cuántica is the venture builder where the ideas that end up becoming companies get built and funded.
Biotechnology, artificial intelligence and quantum computing on one platform, applied to protein design. Of everything I've built, this is the closest to the frontier.
The day's news, summarized by artificial intelligence and tailored to what you care about. Zuma News tackles a concrete problem: there's too much to read and too little time.
Space exploration built from Mexico. Kukulkan Space combines advanced simulation and artificial intelligence to work in a field almost nobody in the region is building for.
A digital art collection that is also the entrance to a real estate metaverse. SoulWindow brings together two worlds almost nobody was connecting: NFT pieces and digital property.
What if you could buy, sell and manage properties from a single platform, with artificial intelligence in between? That's Housi.ai: searches that understand what you're after, optimized pricing, and all of your property management in one place.
Machine learning that reallocates your ad budget hour by hour across Google Ads and Facebook Ads, with a typical 30% to 50% performance lift. Day 6 of the 12 days with JohnOlven.
Custom AI solutions, app development, VR/AR, process automation and technology consulting. Day 5 of the 12 days with JohnOlven.
Video calls, built-in payments, synced calendar and CRM in one place, for doctors, lawyers, psychologists and coaches. Day 4 of the 12 days with JohnOlven.
Fully optimized articles in minutes instead of hours: keywords, competitor analysis, SEO writing, quality control and meta tags. Day 3 of the 12 days with JohnOlven.
Design custom UGC scripts in four steps and send them to a community of more than 3,000 creators. Day 2 of the 12 days with JohnOlven.
Day one of the 12 days with JohnOlven: Hiveflow, a visual orchestration platform for building complex workflows with language models without writing a single line of code.