29 Apr

Posted at 10:32h in Artificial Community Intell
In AI, a world model is the internal map a system uses to predict and plan.  Large language models like GPT rely on theirs to “guess” what might happen in text—or, increasingly, in real life. As we automate harder tasks, that map becomes critical: without a coherent picture of reality, an AI can’t reason or act reliably.
 
But there is no single, optimal map. We have to choose which features to highlight. Do we mirror the world exactly as it has been, or emphasize the future we want? Take mobility: if a driverless car must swerve on a cliff road, does it privilege the passenger or the pedestrian? The decision encodes values—speed, resilience, fairness—long before any code runs.
 
Much talk of AI imagines a glossy, placeless future which feels alienating. Pull.City’s world model, on the other hand, is tuned for local agency: making nearby businesses and organizations more present, strengthening neighborhood ties, and letting communities steer the technology that serves them. That lens guides us in how we model the world—and reflects, of course, the kind of communities in which we want to live.

20 Feb

Posted at 20:17h in Artificial Community Intell

Last night, during a dinner discussion about which colleges to visit during his spring break, my 16-year-old caught me off guard. Instead of suggesting schools, he pointed out that the most successful people he could think of had dropped out of college. He listed tech entrepreneurs who have become household names: Bill Gates, Steve Jobs, Larry Page, Sergey Brin, and Peter Thiel. Before he could propose that maybe he didn’t need college either, I asked him, “What else do all these guys have in common?” “They were entrepreneurial,” he said. “I’m not sure that’s entirely true, and what does being entrepreneurial even mean?” I responded. 

One thing is clear—they were all significantly supported by elite institutions. Harvard University provided Gates with access to advanced computing resources and networks that supported his founding of Microsoft. Although he dropped out of Reed College, Steve Jobs continued learning through auditing classes and leveraged his Silicon Valley connections, notably through the Homebrew Computer Club, to launch Apple. As Stanford PhD candidates, Larry Page and Sergey Brin utilized the university’s resources and Silicon Valley connections to develop and launch Google. Similarly, Peter Thiel’s education and networking at Stanford University were crucial for his co-founding of PayPal and early investments in Facebook. Does Thiel really have the authority to tell young people not to go to college? Would we even know his name if he had not attended Stanford? It wasn’t just academic institutions that propelled today’s celebrity entrepreneurial dropouts with the support they needed to thrive but the entire ecosystem in which they lived. Silicon Valley became what it is today because of the billions in government research and development tech grants invested there. These entrepreneurs had access to education, capital, and legal systems that would prevent others from claiming their intellectual property.

Over the course of my career, I have had the privilege of working alongside entrepreneurs in places like Haiti and Afghanistan, where survival is a daily exercise in entrepreneurial spirit—one that requires initiative, verve, and resourcefulness. There is very little, if any, access to academic institutions, capital, and sometimes even basic infrastructure, such as functioning roads and ports. I think of Magalie Dresse, the Haitian entrepreneur who owns Caribbean Craft, which manufactures handmade Haitian crafts for export to North American retailers. When she received her first large export order, which required her to build a staff, she quickly realized that her employees lacked basic financial literacy because they did not have bank accounts. She found a way to collaborate with a Haitian non-profit focused on literacy training. This partnership helped her employees learn essential skills, making the business more efficient and improving their personal lives. Magalie’s business is one of the rare ones that has continued to exist through earthquakes, pandemics, and government coups. It’s an example of the power of directly investing in and empowering local people from the ground up.

Witnessing entrepreneurs in Haiti and similar environments underscores the critical need for AI solutions that prioritize their needs. While AI has the potential to significantly boost the productivity of small businesses, the current tech landscape predominantly serves larger corporations with a top-down approach. Instead, we are building AI solutions that serve small businesses that leverage the local community. We are guided by our experience working with true entrepreneurs who not only overcome real risk but would, incidentally, jump at the chance to go to college.

27 Jan

Posted at 17:17h in Artificial Community Intell

The idea that OpenAI—or any tech giant—will dominate AI stems from a fundamental misunderstanding of how transformative technologies evolve. In the AOL era, many believed the internet would simply replicate existing media power structures: centralized, corporate-controlled, and dominated by a few players. That’s why the AOL/Time Warner merger seemed so logical at the time—it was a bet on the internet becoming a bigger, digital version of cable TV.

Similarly, today’s assumptions about AI dominance rest on the notion that whoever builds the most powerful models will control the market. OpenAI and Microsoft evoke comparisons to AOL and Time Warner, while Meta’s and Google’s models are also breathtakingly advanced and expensive, requiring billions of dollars in compute and years of research.

However, the recent emergence of DeepSeek, a Chinese AI model developed for a fraction of the cost of these systems, has sparked discussions about possible Chinese dominance in the space. This, too, reflects a misguided assumption.

The real lesson is that lower-cost models are inevitable. They signal a shift toward a more open and distributed AI ecosystem, much like the early internet. Just as launching a website in the 2000s cost a fraction of what it took to produce traditional media, AI is becoming increasingly accessible to innovators at home and around the world. This process has the potential to unlock entirely new industries and far more interesting possibilities than an automated AOL chat room.

12 Dec

Posted at 19:18h in Artificial Community Intell

The journey to co-founding Pull.City reflects a deceptively simple-sounding lesson I stumbled upon many years ago and whose implications I am still seeking to fully understand: communities are strongest when they tackle challenges on their own terms. While this may seem like an obvious truism, in practice, it calls for subtlety, patience, and a commitment to truly understanding the unique circumstances of each community.

Looking back, I have personally worked on and witnessed many well-meaning projects. The most successful ones all have one thing in common: those built from within endure. In Haiti, after the 2010 earthquake, local artisans charted their own paths to sustainable livelihoods, crafting solutions rooted in their own expertise and traditions. In Afghanistan, working with Roya Mahboob and the Afghan Girls Robotics Team revealed how technology and education could spark transformation—when shaped by those closest to the challenge. In Iraq, Yazidi leaders drove advocacy efforts for justice and recovery, creating a movement that resonated globally because it was born from their own experiences and needs. These efforts helped elevate Yazidi activist Nadia Murad, a survivor of unimaginable violence, to becoming a Nobel Laureate.

Each of these efforts succeeded because the people most affected were the ones defining the solutions. These lessons shaped Pull.City’s mission: to give communities tools to solve their challenges on their own terms. This philosophy took on new urgency during the pandemic, when businesses in Brooklyn’s Navy Yard pivoted to manufacturing medical supplies by connecting with neighbors and finding untapped local resources. The strength of these networks showed how much more communities can achieve when they have the means to collaborate effectively.

Pull.City was designed with this in mind. Using AI, we help local businesses connect, share resources, and grow together. It’s not about directing them but about building the tools that let them define and pursue their own futures.

The deceptively simple idea that communities thrive when they lead themselves continues to guide everything we do. It’s a principle that demands humility and ongoing reflection, but it’s also one that offers the most meaningful results. Through Pull.City, we aim to bring this philosophy to life and support communities in building their own paths forward.

05 Nov

Posted at 18:09h in Artificial Community Intell

Artificial General Intelligence (AGI) is a concept that envisions machines reaching human-level intelligence, learning and understanding almost anything independently. The implications fo AGI are provocative, even frightening because of the power it implies: a machine so advanced it could outperform human thought processes.

While AGI remains a sort of obsession among AI doomers and boosters alike, I’ve found thinking about what I call “Artificial Community Intelligence” or ACI more interesting. ACI is fundamentally different in its purpose and design. AGI seeks to reach the pinnacle of intelligence, to match human capability across a range of tasks. ACI is designed with community in mind. It’s purpose-built for the unique challenges and needs of living together, especially in urban spaces.

While AGI is modeled on individual human thought, ACI is modeled on the networks that define human life in communities. It’s not so much about doing things for us but about enabling us to do more together, connecting dots that we might otherwise miss. And while true AGI may or may not emerge in our lifetime, ACI is possible now, grounded in the same social forces that have allowed city neighborhoods, towns, and villages to thrive for centuries.

28 Oct

Posted at 16:41h in Artificial Community Intell

Years ago, when I was working at a prominent internet startup, the founder ordered the team to log into a chat room while investors ogled at a demo of “online community” upstairs. During the early days of the internet, I felt using “community” in this manner was wishful thinking. Since then, the term “community” has been stretched thin by social media platforms. Describing likes and comments as “community engagement” may have seemed innovative once, but it now feels insufficient, even sad. Today, of course, there is urgent talk regarding how we need an idea of community that is more than an Instagram feed, especially for young people. As AI is being trained on the aggregate of this vast, sometimes antisocial, dataset, we face new risks and, perhaps, the ultimate irony when it comes to the word “community.” The void stares back.

Fortunately, an alternative is available and I first got a sense of it when COVID hit. At the time, the worst was feared for New York. Businesses in the Brooklyn Navy Yard pulled together and manufactured medical gear in vast quantities. It was an impressive pivot. I was working on migrating a massive procurement platform for a consortium of manufacturers in the Midwest. Being based in the Brooklyn Navy Yard, I wanted to help local businesses. Speaking to businesses at the time, I heard a common refrain. They were set up to source from overseas and were struggling from supply chain disruptions. In many cases, businesses eventually discovered they could meet their needs locally—sometimes from someone just downstairs or around the corner. It became evident that communities needed better internal visibility and help to operationalize around proximate businesses. Unfortunately, the “community” of most internet platforms did not seem designed to maximize these kinds of connections, but atomize them.

“Pulling together” to powerfully respond to a challenge, in a manner that strengthened our community,  was the genesis of Pull.City. This work accelerated with the emergence of powerful AI. Given our focus on community, our use of AI proved novel. Debates concerning alignment and the misuse of data mattered less, given that we were engineering our platform for community control. Pull,City builds tools to make local businesses more visible and accessible to each other—reflecting the kind of proximity-based connections that define our neighborhoods.  Today, we are excited to be working closely with local businesses in the Brooklyn Navy Yard and DUMBO to build a platform that is transforming, for the better, what online “community” can mean.