1/ For nearly 350 years, science has communicated itself through one object: the paper. A linear narrative, frozen as a PDF, written for a human reader. We've come to treat that format as the medium of science itself. It doesn't have to be. It's a historical artifact. 🧵

The potential for catastrophic effects from the AI boom demands robust deliberation and real democratic governance. Localized initiatives like data center moratoria won't get us there. jacobin.com/2026/04/ai-data-…

Toozology is a difficult and esoteric science, and so it must be approached with the greatest care. But the Toozologists have not reached their goal, a true and correct interpretation of Tooze. They have not appreciated his recent contribution at the right hermeneutical level. Questions have been registered pragmatically—is this ‘blueprint’ feasible? if feasible, would it actually be ‘leadership’ (ie hegemony)? The extent to which this represents another major revision to his thinking has not been seen. In particular, it must be placed within the context of two other Toozological observations: first, his soft spot for Kindleberger’s hegemonic transition theory of the 1930s; second, his hatred for Arrighi and disdain for the neo-Gramscian idea that the history of capitalism can be narrated as a sequence of hegemonies, such that we could characterize the contemporary period as an “interregnum.” Hitherto these positions have been reconciled by observing the “radical novelty” of the early twentieth century—both imperialism and the problem of attempting to construct a global order “after imperialism” were radically novel problems; ditto America, a self-contained continental superpower protected by two oceanic moats. There was a transitional crisis caused by the failure of liberal hegemony in the early twentieth century, but this was because of that conjuncture’s radical novelty. The uniqueness of that conjuncture means the ‘interregnum’ concept developed to explain it is not appropriate for our time. There will be no hegemonic transition from America to anyone else, Tooze had maintained; instead we have polycrisis. Now, he has a blueprint for How China Could Still Win hegemony. Apparently, then, whether or not we are living in an interregnum has not yet been determined. It is an open question. It may be the case, according to Tooze’s recent column, that we do in fact have a hegemonic transition on our hands, if only the CCP wills it. Thus we draw the conclusion that there are other uniquenesses that can lead to a conjuncture correctly being characterized as a hegemonic transition. The early twentieth century uniqueness was not itself unique. Whether there were also nonuniquely unique circumstances characterizing the conjuncture during which the Italian city states and the Dutch Republic came into capitalist leadership roles appears to be equally open to further research and investigation within the bounds of correct Tooziological doctrine as a result of this innovation.

We are actively reviewing applications for a Performance Engineer role at Aeolus expensive-legume-bf0.notion.… I'll be at NeurIPS later in the week, so DM me if you're interested in chatting more about the role or our approach to modern differentiable modeling more broadly!

Clearing out my old office and finding transparencies from my talks from 25 years ago...

🚀 I am hiring a Founding Engineer for Inkboard Inc. 🚀👀 I’m confident there is no other role on the market like this - and for the right person, I believe this could be a career defining opportunity. General Requirements: - Good vibes. 😁 - You're a prolific generalist engineer who hates being confined to just code. - You’re rarely blocked: you learn new skills on the fly to get the job done. - Have worked/Are looking to work on atoms and bits rather than just the latter. - Our customers (and we) love nerdy/quirky/joyful things - if you do as well please mention it when you reach out. - (Bonus) You’ve built products people love. More info in the linked tweet below!

Recently published! Simulating the air quality impact of prescribed fires using graph neural network-based PM2.5 forecasts 👉 bit.ly/3CFlejg By Kyleen Liao, @JBuch7, @karadlamb and @PierreGentine #airquality #wildfires #forecasting

I've thought of this as one of the stronger consistency tests for AI NWP models -- glad there's a detailed study out now!

Are there any examples of differentiable modeling being applied to ecological systems, or discrete-state systems, more broadly?

Horrifying but not that surprising if you have spent any time in New Delhi.

Update your Zoteros/bibliography managers ppl: there’s a new ERA5 reference in town!

I've tried my best to tune this out, but the response of sea ice to anomalous sea surface and atmospheric temperature perturbations over the past two years is a scary phenomenon.