What is 'vibe teaming,' and how can it help solve the world's problems?

Jacob Taylor, a fellow at the Brookings Institution, recently asked his team to sit down with a piece of transcription software. He asked one of his colleagues — a development expert — what needed to be done to eradicate extreme poverty. And as that colleague explained, the transcription software took note. Taylor then slipped into the backend, creating a configured artificial intelligence model that soaked up his colleagues’ expertise on the issue. As a result, he created a platform that knew extreme poverty inside and out, and one that could communicate the policy dimensions, political nuances, and implementation strategies to solve the problem across the world. “In 90 minutes, what we were able to come up with, I think, kind of caught folks … off guard,” said Taylor, speaking at the Devex Impact House on the sidelines of the 80th United General Assembly. The result, according to a Brookings working paper, was this: The platform was able to come up with a strategy that diagnosed the problem, laid out its current trajectory, and identified what was off track, including binding constraints toward progress. “Growth is no longer lifting the poorest: the gains of the past were powered by structural transformation in China, India, and Southeast Asia, which masked persistent exclusion elsewhere,” the model stated. “The final mile is concentrated in high-fertility, low-growth contexts (e.g., northern Nigeria, eastern DRC, South Sudan).” The model then identified five “scalable strategic levers” — in other words, what might work to eradicate extreme poverty — including digital cash transfers, rural opportunity zones, localized employment guarantees, enterprise enablement grants, and nutrition and vaccine top-ups in fragile settings. The model then costed the strategy out, stating that with a $300 billion to $400 billion investment over five years — 40% of which would go toward cash transfers paid for by governments, the World Bank, and bilateral donors, and delivered by governments, nonprofits and mobile operators — extreme poverty would affect less than 3% of the world’s population. It also called for a “Global Compact to End Extreme Poverty,” one with a similar structure to the Global Fund to Fight AIDS, Tuberculosis and Malaria, and a financing platform that “aligns trust-based models … with public systems.” “We are not in a knowledge deficit. We are in a trust deficit. This strategy isn't about inventing new tools — it’s about removing the excuses. Let’s go from ‘why hasn’t this worked?’ to ‘what’s changed that lets us finally finish the job?’” the model concludes. The model was the result of a concept known as “vibe teaming,” Taylor said — an approach that uses generative AI as a collaborative partner rather than a replacement for people. The idea is to harness AI’s ability to understand natural language and handle routine tasks such as drafting or transcription, freeing up humans to focus on what they do best: deeper conversations, strategic thinking, and creative problem-solving. By offloading the repetitive work to AI, teams can spend more time on meaningful collaboration and push their ideas further. “[We need] to be smarter and better about how we build solutions, and yet, at the same time, need to think more human about the community engagement in not only the delivery of those solutions, actually, but the design and development,” Taylor said. “And so that feels like a circle that needs to be squared.”