I've always considered myself a technologist — I have both the skills & tendency to create things using technology. Sometimes, I exercise this tendency via solo runs and it's quite fun. But god, very few things beat the synergy and collective intelligence that collaboration unlocks. Surround me with truly ambitious people and you'll automatically 10x my abilities.
A few months ago, I decided to take a break from active employment — mostly to rest and pursue my interests. But, before this, I had been exploring a product that could help companies reduce their observability costs using ML. Anyway, after a few weeks of solo experimentation in my home office, I realized I could accomplish more in a space with more ambition, vision and synergy.
So to scratch my itch, I commenced search for a residency/fellowship where I could safely spend the next few months of my sabbatical. I had a few ideas and directions in mind. I had some weak goals. All I needed was a creative space that provided some synergy and a safety net to fail.
Unfortunately my search led nowhere. [1] All the spaces I found placed an extreme emphasis on outcomes (weekly demos, final demo day) rather than the process (collaboration, research & experimentation).
One rather stark observation I made was that these residencies seemed very meritocratic and anti-credentialist from a distance, but once I got closer, it turned out they were just filtering by the usual “high signal” heuristics (ivy, iq, olympiads, etc). I have my own philosophy which is essentially that, quota systems lead to systemic bias.
So I gave up on the residency. The next option was to frequent a cafe, but my gripe with cafes is that the entropy that exists there is tangled. You have to crack it open. You need to chat with everyone to get good signals, while not being a nuisance. Cafes don't have “high focus work area” / “social area”. Lastly, cafes only provide temporal cohesion. That architect in the corner has only 45minutes to finish their coffee, if you don't make a conversation within that time period, you lose the butterfly effect forever.
And so while there’s alot of beauty in organically untangling the connections, what if we could just engineer the serendipity. Sure, there’s alpha in discovering a city in the 1800’s without a map, but think of how efficient digital maps make us in 2025. Imagine a cafe where you knew everyone’s interests and specialities. A cafe with implicit cohesion. A cafe where there’s consistent relationships. A cafe with diverse frequencies of ambition and creativity. A cafe where the serendipity is accelerated. Maybe, we don't have to imagine.
Thing is, today I came across 1646, a residency whose mission aligns with my vision. From the bat, they have experimental embedded in their name, and they make it clear their residency exists solely to provide a space for research/work/reflection, not for a final product. The only catch here being, it’s an art residency, not a technological one.
Lately I've developed a lot of interest in how humans organize at large scale. Once I discovered the literature for social exchange theory, I couldn't help but notice how conventional "high signal" heuristics accelerated the expansion of the gap between societal clusters. If we want to meaningfully tap collective intelligence, we have to strive for diversity. Not by biasing for a particular race, but by actually rethinking our incentive structures & filters. For residencies, this translates to optimizing for diversity & experimentation, instead of output. [2]
Hear me out.
- What if we created an experimental residency in Ghana (or anywhere in the world)?
- What if we brought together ambitious, opinionated but low-ego individuals with varied potential and gave them a safe environment for them to collaborate and experiment?
- What if we made it multidisciplinary and increased the surface area for collective intelligence?
- What if we created an experimental space for creative technologists and called it 1957.residency?
It's very clear I think alot about the butterfly effects of experimental spaces. Not only on individual level (micro) but also on societal clusters (macro) and even the global economy.
Least we can do is try, so if you're interested in accelerating how the right people find each other across large networks, let's chat.
[1] Recurse Center seemed quite decent, for what I was looking for. However I felt they were still endgoal oriented. I wasn't trying to push my abilities. I just wanted to do science.
[2] I'm well aware most of these incentive structures exist because of the profit oriented kanker called capitalism.
Agency is a non-diverse way to evaluate potential. Agency is the fusion of ambition and the internalization of the fact that you don’t need permission. The children of rich people don't need permission, and even worse, they learn from a very early age they don't need permission. Find ambitious kids and help them internalize that they don't need permission.
The idea is to train a graph neural network to generate embeddings for traces being emitted by your telemetry system. After, you can cluster these embeddings and perform parallel sampling on a large scale. I built some models but I eventually found out GNNs generalize very poorly. But, according to research done by netflix and stripe, transformer technology could be applied here so there's still potential.