A Virtual AI Institute

Having switched between academia and startups many times, I see we’re at the point where, in certain areas of human endeavor focused on research, technological disruption leads to social disruption. Both startups and academia are techno-social organizations. The most advanced Computer Science labs, such as MIT’s Media Lab and Berkeley AMPLab, fused with startups and their ecosystem — venture capital, marketing, industry events, training — early on. The pull of hybrid labs affects the whole university. Social media is a source of data and experiments in humanities, and using it at scale requires partnerships with CS. In a way, a CS becomes the CTO of a project/research direction, and the domain expert is the CEO. Of course the analogy is not complete as the time scales for grants are different, efficiency criteria are different, and so on.

When building Bay.Area.AI meetups and the wavebreaking AI By the Bay sequence — oh.hai.ai, self.driving.cars, ai.vision (March 6–8, San Francisco) — I am relying on advice and example of multiple experts in the field, the full-stack AI leaders who built MIND stacks form the ground up and now define the AI strategies of their companies and their customers. I need a standing body that our communities can turn to for practical advice that is battle-tested.

In industry, examples of research organizations are Gartner and O’Reilly. Gartner maintains its Magic Quadrant, that is used for buying decisions, classifying technology across several axes. O’Reilly does an admirable job in keeping up on trends supporting its publishing, learning, and events businesses.

Since I run about a dozen pragmatic communities, meetups and conferences, I’ve been thinking of a new kind of a research approach: what works and what’s needed. People come to meetups to learn, and the nature of their questions and conversations is similar to scientific debate. The truth is very simple — will you use it in production?

Cicero Institute is a virtual organization, aligning top practitioners in AI around real-life adoption of the MIND stack and evidence-based projection of trends as expected by those who build the current systems. I call them the full-stack AI leaders — folks who can both build AI from the ground up and then apply it at industrial scale, shaping the strategy.

Some of the best full-stack AI leaders come from the research engineering communities By the Bay. Cicero Institute for Human AI focuses on human aspects of AI, starting with what works in industry context (a human-first problem), how to make sure AI follows human values (Human-Compatible AI, Human-Centric AI, and so on). If you’d like to be an advisor affiliated with the Institute, please get in touch. If you want to support the Institute and collaborate with us, there’re two ways to do it:

• Support By the Bay meetups and conferences — sponsor the learning video series and the events. We don’t have a parent vendor company and rely on sponsors and partners to continue our mission
• Support the Framework.Foundation, the parent non-profit, directly. We align various open-source projects to converge on reference SMACK and MIND stacks that can then be used for commercial support, consulting, and other models working well with board-prioritized foundations such as Linux Foundation, CNCF, Scala Center, and others.