How do you track something that moves roughly at the speed of light?
Think about swatting a fly – it’s almost impossible, depending on that fly’s health and derring-do – if flies were just a little bit faster, we’d have to wait for them to expire of their own volition in our homes.
Sometimes it seems like that’s what it’s like to track AI progress in 2026. The year just started, but we’re in a different spot than we were, say, at Christmas. Agentic AI is absolutely exploding, and humans are running after the LLMs, trying to catch up.
In that context, there was a great Imagination in Action event at Davos “under the dome,” as the world looks at the geopolitics, the business strategies, and the personal reactions to AGI as an evolving force.
So here’s some of what I took away from recent events, in pondering the aftermath of shows like CES, Davos-related goings-on, and general January observances.
The Big Themes
Across all of the many quotes that I received about the Davos Imagination in Action event, some of the most interesting revolve around enthusiasm for the event as something relatively unique. Here’s what my colleague Daniela Rus, director of MIT’s CSAIL lab, had to say:
“The AI Day in Davos showed what’s possible when Imagination in Action, MIT CSAIL, and Stanford HAI work together: the hype stays outside, the hard questions come inside, and decision-makers sit down with the people building the future of AI.”
I like that: a deliberately curated setting where hype gets filtered out, and hard problems get foregrounded. Sally Kornbluth was another who addressed this kind of approach, and various attendees talked about how IIA brought a focus on AI.
Then there were calls for AI governance, which I thought was a timely reaction. Alvin Graylin’s warning about a “winner-less zero-sum race” is one example:
“If the world’s leaders don’t wake up to the truth that AI must be treated as a public good and stop this winner-less zero-sum race we are on, we will destroy the planet we are trying to transform,” Graylin wrote. “The Davos AI Summit organized by IIA, MIT and Stanford is the apex of discussions around AI and our society. All the key players in the industry are in one place and discussing how to get our society to the next stage in a responsible way. That’s the spirit of Davos.”
In calling for additional governance on AI, people are implicitly asking questions like these: what are the enforceable rules, the accountability surfaces, and the feedback loops that keep incentive gradients from driving unsafe or socially corrosive outcomes? How do we keep AI safe, ethical, fair?
In addition, some of the talks at the event centered on what comes after the language-model era: agents, “world models,” and “Physical AI” show up as the next conceptual north star. The real leap, some experts suggest, is systems that can plan, act, and ground their representations in physics, biology, and lived constraints, while still being auditable. In a way, that mirrors the eventual emergence of new interfaces, and new ways of interacting with technology. Cue the household robots, if we can find ways to make them safe around adults, children and pets.
For more, see my interview with Yann LeCun, where the former head of research at Meta goes over LLM limitations and suggests we might be headed for this type of research in the future.
More Fun Fanfare
In the mix, too, are the many shorter and more personal expressions around the event. Some of these are interesting, too, as sort of raw testimony of a conference documenting the state of AI at the beginning of a landmark year. For example:
“I came with zero hard KPIs and no transactional business goals.”
· “In every conversation, I paid close attention—not to titles or outcomes—but to body language, voice inflection, and the sincerity behind each person’s mission.”
· “That experience gave me a level of hope and personal ignition that exceeded my expectations.”
· “I met one individual who told me he has been coming to Davos for 30 years… roughly 85% of the people here fit that description.”
I liked this one:
“Thank you for making billionaires blush, luminaries perplex, and small fish catch big fish.”
More:
· “A week of programming in one day. Insights and fortuitous collisions.”
· “It was mind-blowing and the talk of the town! Everyone was asking how I could get them in.”
· “I have no idea how you are going to top this next year…but I’m sure you will find out how.”
And a couple of callouts to some of our speakers:
· “Will I Am was amazing and great to meet.”
· “I will say that Ami Bhatt continues to be an absolute LEGEND.”
That’s a little more on the flavor of the event, as we conclude and go back to our everyday routines.
Moving On
So we have mandates, from this meeting of the minds. One is to maintain open research models. Another is to practice good governance, even if that means putting forth more effort, in AII initiatives. A third has to do with understanding the balance between human and AI skills, and making sure that humanity always has a place at the table.
Stay tuned for more.
