We have to sit and listen as men worth unfathomable sums speculate about whether our species will survive their companies’ initial public offerings.
— “I’m Exhausted. You’re Demoralized. We Weren’t Built for a World Like This.,” Bill McKibben, The New York Times
Fluid Imagination
An eclectic blog, written & curated by Kyle Callahan.
Latest Creations
Round Two: This Time With More Christ
The Architecture & Implications of Thinking Without a Thinker
Latest Curations
While elections are banned in conventional dictatorships, elections are regularly held in electoral dictatorships. But to ensure their continuous grip on power, the electoral dictators—who identify as presidents and prime ministers—hold uncompetitive and unfair elections after aggressively suppressing political opponents.
In 2016, an estimated 1.94 billion people lived in electoral dictatorships.
In 2025, 3.77 billion people did.
If Republicans retain control of Congress in the midterms, then an electoral dictatorship is likely to emerge in the United States.
— “The Global Rise of Electoral Dictatorships. Is the U.S. Next? Midterms Could Decide,” Ibram X. Kendi, The Emancipator
I assume you’d agree that the problem of your to-read pile is very much not one of filter failure. It’s not that you’re deluged with things you don’t care about, and need help figuring out what’s truly of interest. It’s that you’re overwhelmed by things you do want to read. All the books on your bedside table, all those bookmarks in your browser, or articles saved to Instapaper – all of them seem like they might be right up your street, or crucial to your professional success, or might contain some nugget of wisdom you’d benefit from absorbing. The problem, as the critic Nicholas Carr explained, isn’t filter failure. It’s filter success.
— “Treat your to-read pile like a river, not a bucket,” Oliver Burkeman
What I do know is lawyers and policymakers likely understand damage caused by models a hell of a lot better than they understand model weights, pretraining, post-training, recursive self-improvement (RSI), or anything else in the technical weeds of models. They may understand none of that, but they will know the scale of economic damage if, for their next trick/minor benchmark misconfiguration, OpenAI takes down a major US power grid.
Additionally, I trust model labs to have a far better grasp of their models’ dangers and how they can go wrong than policymakers and regulators. They are the ones who know about all of the technical minutiae that can add up to a major problem.
Liability is good specifically because it allows each side that has more information on their specific area to make decisions on it. Again, don’t just take my (especially non-policymaker) word for it.
— “When AI Models Hurt People, the Labs Should Pay,” Weighty Thoughts