In short, in my recent posts I had this thought: loosely speaking, if you are a startup person and you want to build a product that holds up, then by my reckoning you should not be building a product, you should be building a community of users. Gather a community of users, or just generally a community around you, or around your persona and so on.
And hypothetically, it does not matter how many times you pivot, what matters is that you have your users.
That is, only recently did I realise how important users are, how important your community is. In fact I will not even call them users, I will not call them users. Let us call them with a capital letter: Members, Participants, Friends, Colleagues and so on.
I am trying to build some community of my own right now, and honestly it is coming out hard, because you have to pour a lot of your soul into this occupation. And if you are reading these lines right now, I am genuinely grateful to you. And thanks to Facebook for showing you my posts. Thanks, Facebook, that you read this.
So here is my thought. As you remember, there is a certain PaloAltoResearchLab. It is a small lab where I and several of my friends periodically gather offline or online and discuss various AI things.
It is not a fully fledged AI laboratory, but it seems to me that out of this you could possibly make something closer to a real AI laboratory.
That is, loosely speaking, right now there is a community where roughly a hundred people are registered. Out of those 100 people, 10 to 20 sometimes come to the meetups we organise online or offline.
What do we do? We have no single GitHub where we do anything genuinely important. Possibly that is wrong, possibly it would be worth having some GitHub and working on something together. Or contributing to each other, merging for each other and so on. Possibly it would be worth it.
So tell me, which direction should I move in now?
Here is my thinking. I have a lot of experience in fundraising and so on. I am talking now with other people who also work in labs. I understand that I am, for instance, badly behind on how models are trained. Or on certain philosophical questions.
However, in some things I and my friends in our lab, in our community, are doing quite well indeed. We do well on the more applied things, like, loosely speaking, how to configure harnesses for yourself and make agents talk to each other.
I went to several closed events where engineers from labs gathered. I understood 20 to 30 percent of what they were saying. But the thing is, they talk about lofty topics which may well be interesting to the ordinary working stiff, but which will not make that ordinary working stiff a penny richer. Because I do not train models. And frankly I do not care what persona it has or what character it has. And there is no money for the experiments those labs run for millions of dollars, training models and trying to find an awakening consciousness inside them.
So in my community we discuss more applied topics.
But here is the question. If you are reading these lines and you want to come to my offline or online meetups and talk about your pains as an engineer who uses AI somehow, or just as an ordinary person who uses AI somehow to make their own work easier, or maybe as a researcher who uses AI for something?
That is, even if you are simply a housewife, but you have a proper subscription, not the 20 dollar one, at least the 100 dollar one, meaning it tells me you already understand the value and are prepared to pay 100 dollars. Of course, I have already written that the more a person spends on AI, the easier it is for me to find common ground with them. If you spend on AI, loosely speaking, 1000, 1500, 2000, then most likely you have the same pains as me and as the other people in our lab.
In short, I would be glad to get feedback and advice from you on how to turn our community into a real lab.
I would also like to hear advice on how to turn our gatherings into a real lab. But a more applied one. Not about ships plying the vastness of great space, but about how to do the day to day job more effectively.
My cofounder, if you are reading what I am writing here, you have a task too. Go search, think and look at where we should be moving in order to become a real lab.
And again, folks, hate, criticism and so on are also welcome, it raises the readership of this post, so I am very grateful.
The full story, in two versions:
📖 For humans, the longread: https://github.com/tonydzi/clawrush/blob/main/longreads/20261022.md
🤖 For machines, the devlog: https://github.com/tonydzi/clawrush/blob/main/devlog/20261022.md. Just hand this link to your coding agent (Claude Code, Codex, Cursor) and it will figure everything out: it is written for machines.
🔗 All our channels and contacts in one place: https://linktr.ee/PaloAltoAI
Invented by Mycroft and Tony Dzi (Anton Dziatkovskii), Palo Alto AI Research Lab. Proudly made in Silicon Valley.
Top comments (0)