Are We Bold Enough? TEDAI Kicks Off a New Era of AI

Yesterday was Day 1 of TEDAI in San Francisco. “Are we bold enough?” is the core question of the conference, and from the 17 talks of that first day, I would say we are bold for sure.

Here are a few highlights of day one:

On the bold side:

• Sam de Brower's answer was that when we can overcome the worries and disillusionment,  2026-2027 should be the time for “enlightenment,” and that AI is “close to being a new species,” and James Joaquin qualified AI as an “exotic mind-like entity.”

• Eric Zelikman believes that we can create AI that is truly human, that can “empower people instead of replacing them.”

• Rafael Rafailov believes that we can get closer to AGI once we train AI how to evolve, which he calls “meta learning,” once we get “the right rewards and experiments.”

On the challenge related to the amount of AI energy and computing power required, answers came from many perspectives:

• Aysse Coskun believes that we will need AI to become a “conductor” that can manage and optimize the energy use of multiple data centers, ensuring flexibility in energy usage and optimizing clean power.
  

• Janet Egan argues that the need for data centers to support the growth of AI creates critical geopolitical issues: those who own the data centers and therefore computing power (currently, the US dominates) will have an upper hand due to their valuable power. “Nations that have data centers will write the rules,” so “geographic diffusion is essential.”
  

• Philip Johnston's answer was bold: he proposed placing the data center in space, and he is launching his first one next week. It requires a four-kilometer-square data center that can be launched via a reusable spaceship (which keeps costs affordable). The benefits of a space center include not requiring land, battery storage, or water (a major issue, as much water is needed to cool down the data center on Earth), and needing many fewer cells because it is close to the sun and always receives energy. He believes that this model is economically viable and a great alternative.

On the impact of AI and jobs, all the speakers agree that it will have an impact, but minimize the long-term effects. Vlad Tenev made a historical review of jobs and believes that this is just a new iteration on new jobs like we have seen since the caveman, with each innovation making some jobs obsolete but creating many new ones.

And on the challenge of leadership, May Habib emphasized that only CEOs who embrace AI, “get their hands dirty and don’t delegate” and become “architects of radical simplicity” will succeed. To succeed, they need to fundamentally redesign their organization, structure, workflow, and champion human potential. Those who delegate AI to IT miss the whole point that AI is not just a technical tool but requires rethinking all aspects of an organization.

On the personal level, it was great to see friends (like Hope Timberlake in the picture)  and meet new ones.

Excited for Day 2 of the conference with workshops and panels.