Time has published a list it calls AI 100, and the first name on it is not a chief executive. It is Daniela Rus, who runs MIT's Computer Science and Artificial Intelligence Laboratory, cited among other things for her part in Liquid AI. Sam Altman follows, listed alongside his colleagues Mark Chen and Greg Brockman. Then Elon Musk. Then the twins Dario and Daniela Amodei, who founded Anthropic. Jeff Bezos turns up further down. The ordering is the most deliberate thing in the package: a university lab director placed ahead of the people whose companies set the market's terms.
Liquid AI is building a different sort of large language model. The models are made smaller and more efficient so that AI systems can run on edge devices with no internet connection. The argument for that is not only technical: a user can train their own model and work with it away from the attention of governments and businesses, without depositing personal data into the closed ecosystem of a large provider that shapes how the system gets used around its own priorities. The architecture relies, among other things, on input-dependent computation and a routing mechanism, which is meant to let the system better reproduce human-like, if simulated, cognitive processes.
That premise sits at an angle to almost everything else on the list. The rest of AI 100 describes an industry organised around scarcity — chips that must be bought, clusters that must be financed, services that must be subscribed to. Liquid AI's line of work is an attempt to make the interesting part of that stack small enough to walk out the door with. Time gave it the top slot, which is either a statement about research prestige or a bet on where the constraint moves next.
The cluster that follows — Altman, Musk, the Amodeis — is a fair picture of how the United States runs its side of the AI race. Market mechanisms do the organising, large corporations sell everything from coffee to medical care, and big business drives much of what happens. Anthropic has been in the headlines repeatedly over the past year, including over its standoff with the Pentagon, which is itself a useful marker of how far these companies now sit inside questions of state.
China arranges the same industry differently, with far stronger central state control, which is part of why open-source models get more of the attention there, even though Alibaba, Tencent and Bytedance carry considerable weight of their own. Running alongside all of this is the American effort to restrict not just GPU exports to China but China's use of remote cloud services — a separate story, and a live one.
The list widens from there. Ben Affleck is on it for his work with InterPositive, a company the South Boston actor conceived to apply digital twins in film post-production. Affleck, per the list, thinks cinema will be among the last fields AI replaces. Azalia Mirhoseini and Anna Goldie of Ricursive Intelligence are included for work on improving Google's TPU. Among model builders, Time names Yang Zhilin of Moonshot, and finds room to mention his Pink Floyd habit. Robotics is represented by Deng Taihua of AgiBot.
Read as a whole, the list is less a ranking of people than a diagram of a supply chain. There is the hardware layer, where Nvidia and others accelerate chip development. There is the model layer, where Liquid AI and Rus appear again. There are the user-facing services — GPT, Claude, Grok and the rest. Robotics supplies the physical shell, and a long tail of adjacent work fills in around it.
My own read is that the Affleck entry is the one worth pausing on, and not for the reason Time includes it. The claim that film will be among the last industries AI touches is the sort of thing incumbents in every industry have said in turn, and it is being made by someone who has just founded a company to put synthetic doubles into post-production. Those two positions are not contradictory, but they are not comfortable together either, and a list that includes both without comment is doing celebrity coverage rather than analysis.
There is also the matter of timing. Time offers the list as a prompt to think about where the industry goes in the third and fourth quarters of 2026. The third quarter is nearly spent. A list framed as forward-looking that arrives with a quarter of its own horizon already gone reads more like a snapshot of who mattered than a claim about who will.
Which brings the thing back to the name at the top. If the work Rus is associated with succeeds — models small enough to train and run outside anyone's data center, on a device that never phones home — then the chokepoints that make most of the names below her worth listing stop being chokepoints. Time put the person working to dissolve the moat above the people who own it.