Accelerate SF: Meet the former tech employees who want to use AI for social good in San Francisco 

San Francisco is a city of contrasts. It’s home to some of the most innovative and influential technology companies in the world, but also to some of the most persistent and complex social and urban challenges. From housing affordability and homelessness to transportation and public health, the city faces many issues that require creative new approaches and solutions.

That’s why a group of former tech employees have decided to launch Accelerate SF, a new initiative that aims to drive engineers to build AI solutions for San Francisco’s major public sector challenges. The initiative is hosting a hackathon on Nov. 4th and 5th, where participants will work on projects related to these challenges, and potentially continue working on them after the hackathon.

Accelerate SF was founded by Anthony Jancso, Jordan Wick, and Kay Sorin, who have each worked at leading technology companies — Palantir, Waymo, and YouTube respectively. The three have been promoting their hackathon at various AI social gatherings in San Francisco, and have received support from local stakeholders, leading AI companies, and large political organizations. Perhaps most notably, they have partnered with the Mayor’s Office of Innovation, which has helped them identify and scope real problems faced by public sector organizations in the city.

The hackathon is one of the first major initiatives by a large American city to incorporate artificial intelligence into the public sector. It follows the trend of other cities such as New York, Boston, and Seattle that have been exploring the use of AI for improving urban services and governance. However, Accelerate SF stands out for its focus on the use of large language models (LLMs), which are particularly good at finding and organizing large sets of complex data.

Jancso and Wick, who are engineers by trade, and Sorin, a seasoned marketing and events professional, all have experience in working with government organizations and solving their problems with software. They have individually noticed that most hackathons and AI initiatives are focused on solutions or technologies for the private sector, while the public sector is full of opportunities for AI applications that can have a positive impact on the city.

“We believe that the public sector is full of opportunities,” Jancso told VentureBeat, “because large language models are really good at applications that rely on analyzing a large amount of text. If you look at all of these government forms, and all of these government websites, they are excellent places to look for opportunities for AI.”

Some of the examples of AI applications that participants might be working on at the hackathon are:

• A new open book website that uses natural language models to answer questions about government spending in natural language.
• A simplified form for reporting car break-ins that uses a large language model to create a narrative of the incident based on the details entered by the user.
• A permit approval application that checks the permit application for errors and provides feedback to the user.

The hackathon will also feature speakers and judges from the tech and civics sectors, such as California State Sen. Scott Wiener and entrepreneur Kim Polese. These speakers and others will share their insights and perspectives on how AI can be used for social good and public sector improvements.

“I think there’s really a lot of concrete change that can be achieved, especially with so many brilliant, talented tech people in one city,” Sorin said. “I just encourage people to get involved and know that even making the smallest effort building one application can have a huge impact for everyone in the city.”

Accelerate SF has received funding from several sponsors, including Scale AI, Chroma, OpenAI, Anthropic, Replit and LangChain. These sponsors are also providing support, prizes, speakers, and tools for the hackathon. Scale AI, for example, is an AI company that is working closely with Accelerate SF on this project and is already building AI applications for local government as well.

The Accelerate SF hackathon is open to engineers who specialize in large language models and machine learning. Interested participants can sign up on Accelerate SF’s website or follow them on Twitter for more updates.