Taking care of entrepreneurs. Startup Embassy Founder. Hacker houses operator since 2012. https://xitter.cf/t.co/qpuK0BApcY SF. I write about coliving: https://xitter.cf/t.co/7sj1hNVzKp
Palo Alto, CA
Joined December 2009
- Tweets1.4K
- Following771
- Followers705
- Likes903
Pinned Tweet
Startup Embassy is the work of my life. For over a decade I've run it: a place in Silicon Valley where founders from all over the world come to live and build together, far from home, chasing things most people would call impossible.
I don't buy the lone-genius myth. What I've watched up close is the real thing: the doubt, the setbacks, the years away from the people you love, and the stubborn hope that keeps you building anyway. That is the only Silicon Valley I know, and it's the one I write about.
A few years back, a filmmaker named Kenji Yamamoto wanted to film inside a hacker house. He asked around. House after house said no. I was the only one who said yes.
I said yes for a selfish, human reason. I was ten thousand kilometers from my family, betting everything I had, and I thought: if I fail, or if I somehow make it, I want it on record. One day I'll have to explain to my kids why I was gone for so long.
I should never have let him in. Startup Embassy was for founders, only founders, and Kenji was a filmmaker. I broke my own rule. And I was completely wrong to hesitate, because Kenji turned out to be one of us. A dreamer risking it all on something uncertain, exactly like the rest of us. He slept with his camera on the pillow and filmed for ten years.
That film is called "The Dreamers and I." It follows a handful of us through the real thing. No highlight reel. Just the whole messy, human truth of trying to build something from nothing. It has screened at festivals around the world.
Here's the trailer.
These days I'm also part of Frontier Tower, but the job hasn't changed: taking care of the people brave enough, or crazy enough, to risk it all for an idea.
If that's you, welcome. You're among your own here.
Mama Carlos retweeted
Sra. Colau, según sus parámetros, #SoyPijo.
Mi padre murió cuando yo tenía 18 años. No heredé nada. Mientras estudiaba Ingeniería Industrial, trabajaba todas las horas que podía para ganar dinero.
Empecé de cero. He creado empresas, he vendido varias y me ha ido bien. Detrás hay incontables noches sin dormir, riesgos y sacrificios que usted desconoce.
Me fui a California con mi familia para seguir emprendiendo. Mi mujer renunció a su plaza de médico, ganada por oposición. Después de 14 años, hemos vuelto.
Y me encuentro con que una exalcaldesa de Barcelona considera aceptable hablar de “freír” a impuestos a personas como yo.
Pagar impuestos es justo y necesario. Financiar los servicios públicos exige que contribuyamos según nuestra capacidad. Pero una cosa es exigir una contribución y otra convertir en enemigo a quien ha prosperado.
Usted no sabe de dónde viene cada persona a la que llama “pija”. No sabe qué ha arriesgado, cuánto ha trabajado ni a qué ha renunciado. Su etiqueta borra todo eso para convertir el éxito económico en un motivo de desprecio.
Un país necesita personas que quieran emprender, invertir, crear empleo y contribuir. Si el mensaje que reciben es que se las quiere “freír”, algunas acabarán haciendo todo eso en otro lugar. Y perderemos todos.
No le pido que me exima de contribuir. Le exijo que deje de señalarme como enemigo por haber prosperado.
La envidia es muy mala consejera. Y alimentar el resentimiento desde una tribuna pública es una irresponsabilidad. Usted ha sido alcaldesa y conoce el peso de sus palabras. Debatamos sobre impuestos, pero deje de utilizar el desprecio como argumento político.
This gives me chills
Mary Phuong of DeepMind on AI loss of control risk.
"The last 2 minutes gave me chills.
A significant amount of people saying that AI companies are paying people to spread fear about these agents.
I have never really felt like [loss of control] was an actual thing to care about.
But seeing Mary and how her face trembles, made me reconsider my stand. She's either a Hollywood-level skilled actress, one of the best, or we should be scared. For real."
"Seems like Mary is already aware of more than she can admit and it haunts her..."
This one's about entrepreneurs, and it starts with my son.
Tomorrow he fights in the final of the Madrid boxing championship. The kid in the other corner is clearly better than him. My son has his puncher's chance, like anyone who steps into a ring, but he knows exactly whos waiting for him. That's him in the photo, worn out during his previous fight. To me he's already a winner, just for getting in there knowing all that.
But you do have to get in.
There's a phase zero to starting anything, and I have a name for the people living in it: coffee entrepreneurs. Every founder has been one, me included. You sit around a café table with people just like you, dreaming out loud about how cool it would be to build something of your own, to cut loose from the machine and the rat race and finally be free to do what you actually love, because god knows we all have ideas that could change the world. Days go by, then months, then years, and one day you look around that table and realize nobody has ever stood up and jumped. When I jumped, my son was still in his mother's belly. Now he's a fighter.
A boxer has to climb into the ring knowing he might get his face broken, and a founder has to get up from that table and start walking. Every time we talk about entrepreneurship it's incubators, accelerators, VCs and the whole circus, and the one thing almost nobody mentions is the fucking risk: climbing up there scared and going anyway. Until we do that, we're still coffee entrepreneurs, however good the ideas on the table.
Tomorrow my son climbs into the ring. I'll tell you how it goes.
And you, stand up from the table. I've yet to meet a single founder who regrets it.
Congrats Manu!!!
Labelbox has raised $100M+ in Series E funding at $1.25B pre-money valuation, led by Infinitum Partners, with a16z, B Capital, Tau Ventures, Wipro Ventures, Snowpoint Ventures & others.
I'm excited about our products, Horizon for frontier AI labs and Recursion for enterprises, and where we're headed. I wrote more about what this means and what's next: labelbox.com/blog/labelbox-s…
Mama Carlos retweeted
One of the worst predictors of founder success we've tracked is how well someone pitches; the correlation between pitch quality and outcome was often negative.
Zuckerberg was so awkward in early investor meetings that VCs wondered if he could ever manage anyone, and Larry Page refused interviews and earnings calls for years. Bezos had to pitch 60 different investors to raise his initial $1 million.
Highly articulate people are great with the 30min process of telling a story to strangers, but this is a pretty different skill to building a company. Founders who spend years optimising to be persuasive have often optimised away from the building skills that compound over the 10yrs of execution that come afterwards; it’s incredibly hard to do both well.
The articulate founder will impress your partners but the awkward one will return your fund.
Mama Carlos retweeted
Villaroya lo explica muy bien.
Una intervención puntual hace huir a los capitales a otra parte de la economía. Si intervienes en un tipo de vivienda, se especulará en otro. Si intervienes en una norma se especulará con otra. Si intervienes todo el sector se especulará con alimento. Con lo que habrá que generar una nueva intervención. Y después otra, y otra, y otra.
Y cuando los dueños no quieran poner sus viviendas en alquiler porque sea antieconómico tendrás que expropiárselas.
Y cuando crees la ley de expropiación, la pondrán a nombre de sus hijos o de su mujer, y entonces tendrás que regular eso.
Y cuando dejen de invertir en viviendas, la construcción se hunda, nadie haga mantenimiento y el precio se dispare tendrás que obligarles a invertir en vivienda.
Y cuando no quieran tendrás que expropiarles los capitales. Y cuando eso ocurra se los llevarán fuera y entonces tendrás que prohibir que se los lleven fuera.
Y entonces viajarán con ellos fuera en forma de oro, joyas o lo que sea. Y tendrás que prohibirles que viajen.
Pero el resto del mundo seguirá avanzando y entonces tu problema será el resto del mundo y habrá que empezar a intervenir ahí. Ya sea convenciéndolos o conquistándolos.
Tiene razón Villaroya. Una intervención puntual no detiene al capitalismo. Hay que subyugarlos a todos.
Mama Carlos retweeted
In honor of GTA 6 and World of Warcraft Forever launching soon, I'm open sourcing a tool I built that lets you manage your AI agents while you game: Herdr HUD
I've been using it the last month while I play WoW and it's awesome
Basically it finds all your agents you're running in Herdr and puts them in a tiny widget that floats over anything on your screen.
When you click the floating button, your agents appear and you can see their output and message them
Anytime they message you, you get a small notification over the widget so you can see what they say then open the widget if you want to message them back
Allows you to game all you want without feeling guilty about not being productive!
I've literally got more work done while playing video games than most people do not playing them
Great for slower paced games like WoW or turn based games like Star Wars Zero Company or Slay the Spire 2
If you haven't used Herdr yet you need to. It's a free open source tool that runs in your CLI and makes it super easy to open and manage tons of coding agents at once across all your projects. This widget simply plugs into your Herdr instance
Link below for Windows, Mac, and Omarchy! And if you want to join my WoW Forever guild when that launches let me know too!
Installation instructions in the readme, as well as instructions for Herdr if you don't have that yet. Enjoy!
github.com/finna/herdr-hud
Mama Carlos retweeted
After 15 years of investing, we realised that truly exceptional founders have something impossible to fake; deeply unconventional lives.
We analysed 15,000 founders using five binary signals to measure this: quirky hobbies, early-life signs of exceptionalism, extreme life pivots, unusual geographies, non-linear careers. These sum to give a 0-5 score per founder. Whether someone started coding at age 10, speaks five languages, climbed Everest or quit a safe job to live in Chile, the signal we wanted was deviation from the mean.
Rather than focusing on IQ or EQ, we call this metric the Outlier Quotient, or “OQ”. When forecasting founder success, it turns out that OQ was the single most predictive variable in our entire classification model, trained on ~70 different factors.
Our OQ score had zero correlation with having worked at a top-tier company or attending an elite university. The signals most VCs rely on are noise without signal. The best founders don’t signal like everyone else, don’t think like everyone else and certainly don’t build like everyone else.
Mama Carlos retweeted
Humans might have ~2–3 years left to make meaningful cognitive contributions before AI can do them all.
One of the priorities we need to get right before then is alignment.
Mama Carlos retweeted
The thing I freaking warned founders to engineers to VCs is here. Headline in a business paper:
“AI phobia is America’s new consensus”
“‘Stop us before we end humanity,’ the AI titans keep telling us. Also, ‘trust me and me alone, not my competitors’”
They poisoned the well and pooped in the hot tub. Two CEOs crippled by arrogance either purposely or cluelessly—injured the public perception of AI (and robots).
It looks only get worse and will launch Luddite political careers and elect a new president riding on the back of this new anti AI, anti Clanker donkey.
All of this could have been averted and now we are in a new Luddite timeline.
This stupidity will assure a) No IPO b) Market crashing post IPO burning down. All from avertable public comments and company positions.
This has a slight chance of being averted but past performance now shows future gains in doubling down in arrogance.
You reading this, as much as you may or may not love this tech, are in the 1% that by comparison hold AI’s promise and peril in an educated mind. I know this because somehow my feed reached you.
You represent a well thinking very small and easily trampled minority.
How did this happen?
Thank Dario. From his holier than tho thought pieces to his congressional and presidential emergency meetings to his almost daily fear marketing reports dressed up as academic papers he is the poster boy for the anti movement.
Quite unlike any CEO in history this guy has captivated investors to suppress logic for hopes of a big exit.
I do not dislike Dario personally, however his philosophies are childish and he has convinced LessThinkers to follow him over the cliff. No doubt all with the best of intentions to save us from our selves because they “know better”.
The outcome of course will not be what their cos-play of a philosophy: “we are like the nuclear scientists in the manhattan project making ‘the bomb’” now has the general hard working and very stressed out rest of the world, angry and freaked out.
The problem is clear when you drive across America and talk to common folks. THEY LOVED AI TWO YEARS ago. Now they have dozens of reasons to hate it. They certainly hate the poster boy and Sam (he is another story). But they will take the words of waning and fear seriously, as I expected and use this fear theater marketing from the LessWrong ivory towers as fuel to burn it all down.
I spent most of this summer talking to over 300 random folks around this country and NOT ONE supports the positive things we see with AI here on X every day.
NOT ONE. Bankers, doctors, nurses, lawyers, real estate, restaurant workers to high executives. In the promise this technology offered, flooded are the noise of Dario and now the press and politicians filling the mass minds with a guaranteed hatred if the tech. And no, no acceptance as Dario as savior.
The meeting the pope and having the Anthropic minion mini Dario give a speech has sent many Catholics off to full hatred of AI.
As I said, this can be readjusted but it takes very detailed coordination, communication and money.
I did the first half for free right here : identified the problem, much to my demise as many of my colleagues are having a party, rightly so, on the great daily achievements in the industry.
You folks have tolerated my obvious posting just about daily on the Anthropic game. Trust me Bro, it makes me no friends with the industry even if they don’t like the company.
I am speaking up here because we have 4 months. It will be hard but if understood we have a much better future of clear thinking. If unsuccessful we have a decade of decay and world AI dominance by China, because they have nothing like an Anthropic and if they do get one: BE VERY SCARED BECAUSE THEY HAVE SOMETHING VERY POWERFUL IN AI THEY ARE NOT TALKING ABOUT AND ARE ROLLING UP THE GATES.
This is the way of it.
I want to be very wrong here but my track record on the future says otherwise.
Pray for us.
Mama Carlos retweeted
Best of luck at @ycombinator demo day and it’s such an honor to have had you participate in Batch 0 of the @frontierheroes ! Huge part of building startups is story telling. I remember brainstorming the foundation of this story with @KimAdams during breakfast @Superherohotel and it’s so cool to see how it evolved! Love it and excited to be part of your journey!
We are launching on @ycombinator
We are the AI collections agency for late-stage portfolios.
We find hard-to-reach borrowers, understand their situation, and negotiate using AI voice, email, and SMS. Just last week we negotiated with 100k different people.
Mama Carlos retweeted
Grateful to the @Icelandeclipse community for the stage and the trust and one of-a-kind-experience. Got to meet some incredible people and to share what we're building at @frontierheroes and @frontiertower - our SF landing pad helping the world's best founders land, raise, and build in the US. If you're a founder eyeing US market entry or investor seeking to co-create something truly unique together with us, let's talk. 🇮🇸
Mama Carlos retweeted
Hermes Agent by just crossed 200K GitHub stars in just a few months. I tested every claim in the @NousResearch project: memory, skills, cron, sub agents, all from Hermes Desktop App and WhatsApp.
The whole system is one loop:
message in → agent run → tools fire → loop ends → memory saved → skills grow
What surprised me:
No embeddings anywhere. Skills and user facts are plain text files.
Chat history is one local database. And it still compounds: the longer it runs, the harder it is to leave. The moat is memory, not the model.
I caught two gaps too: it never saved a non-complicated skill without being asked, and a cron bug. Details in the video.
Watch it, then save the harness diagram below. 🔖
You Can Build Anything. You Can Learn Anything. 💪
Mama Carlos retweeted
Harvard and MIT Researchers Simulate the Entire Planet with 8.3 Billion AI Personas
It is big.
In a development that feels straight out of science fiction, a large collaborative team led by researchers from Harvard University and the Massachusetts Institute of Technology has unveiled MatrAIx: a population-scale AI simulation infrastructure designed to model the behavior of virtually every person on Earth.
MatrAIx centers on Persona 8B, a dataset containing 8.3 billion unique digital profiles. Each persona is defined across 1,290 categorical dimensions that capture everything from demographic background and psychological traits to spending habits, technical literacy, behavioral quirks, and lifestyle preferences.
I am using Persona 8B quite a bit and it is interesting.
The system is not merely a static database. Researchers bring these personas to life as agents powered by frontier large language models, including. These agents can then be dropped into four distinct digital environments: surveys, AI chat interfaces, live web browsing, and native desktop and mobile applications.
How MatrAIx Works
Persona 8B was constructed using a sophisticated dependency graph that preserves realistic correlations between attributes. Some records are synthetically generated while others are carefully extracted and grounded in real human data from biographies, reviews, surveys, and consented self-reports.
For practical research use, the team has released a high-quality coreset of approximately one million personas.
Once activated, the agents interact with real digital products and systems. Researchers have already run more than 18,000 evaluation trials across 1,010 tasks spanning commerce, software, finance, healthcare, and more than 20 other domains. The system records granular behavioral signals — how long an agent hesitates after a price increase, when it abandons a broken checkout flow, how much latency it will tolerate before closing an app, and whether it continues after an AI assistant fails.
Strong Validation Results
In a controlled study of 400 trials measuring adherence to ten behavioral attributes across all four environments, the agents successfully expressed or correctly suppressed their assigned traits 91.5 percent of the time. Human judges also rated the quality of the human-grounded personas highly (average 4.135 out of 5).
These results suggest that MatrAIx can generate coherent, demographically and psychologically consistent simulated users at a scale previously impossible.
Implications for Research and Business
Traditional market research and user testing are slow, expensive, and limited in sample diversity. MatrAIx offers a complementary approach: the ability to simulate how billions of different types of people might react to a new product feature, pricing change, interface redesign, or AI system — overnight, on a single server.
Product teams could stress-test ideas before expensive real-world launches. Researchers studying human-AI interaction could explore rare behavioral edge cases. Policymakers and social scientists might model the downstream effects of new technologies across highly diverse populations.
The project is open source. Code is available on GitHub, a project website has been launched at matraix.ai, and the one-million-persona coreset is being released for broader research use.
MatrAIx is not intended as a replacement for real human feedback. The researchers emphasize that it is a powerful tool for exploration, hypothesis generation, and large-scale stress testing.
As the underlying language models improve and the persona models grow more sophisticated, the fidelity of these digital populations is expected to increase further.
For now, the system represents one of the most ambitious attempts yet to create a usable, population-scale digital mirror of humanity a simulation infrastructure that lets us ask “what if” at planetary scale before deploying them in the real world.
I went to Silicon Valley to build a company. I ended up building a community. It took me years to understand that the community wasn’t a side effect. It was the thing.
Mama Carlos retweeted
Here I talk about Nomad families and how they are navigating travel as well as minimumviablesociety.com/ thanks so much @MrTimothyAllen for having me
He planned for 20 families.
500 applied.
That's not a waitlist. It's a signal.
New episode with Daniel Thompson on why families, not lone founders, are how you actually fill a Free City👇
youtube.com/watch?v=xXHkRh3Q…
