Disembedded is a series of experiments in doing economic geography publicly and at a distance. The first will run from August 23rd to 28th 2026: I will try to write a quantitative paper in six days, alone, from a beach on Lantau Island near Hong Kong, with AI tools and no local human collaboration, and every hour of it will be livestreamed. I call the series Disembedded to emphasize that I will be temporarily weaning myself off the local patterns of interaction that economic geographers assume are helpful in my normal mode of academic production.
I see these as experiments in applying the tools and spirit of the creator economy to academic scholarship, a way to find out what I can take from it that would make my own practice better. If it strikes you as fundamentally weird, I agree. But perhaps that is because the dominant system of academic production is out of touch with the rest of the world of content, and not because it is wrong for academic material to be content.
Below I define creators and propose a few ways our field might improve by embracing the creator economy more. None of this is exhaustive or a final word. It is meant to anchor the project in issues you might care about. I write from the perspective of an economic geographer, but nothing here should be any different for other social science researchers.
Drawing on a recent review by Peres and colleagues, we can agree that the creator economy is the part of the economy where individuals use digital platforms to create and monetize creative works under their own personal brands. As geographers we often get more meaning from where something happens than from what it is called, so it may be just as useful to say that the creator economy is what is happening on TikTok, Instagram, YouTube and other prominent platforms.
What is striking is how comprehensively this mode has displaced linear release, seemingly across every form of content except the scholarly. Music, television, journalism, education, and instruction have all been reorganized around it, and film is now going the same way. In 2026 two YouTube creators, Kane Parsons and Curry Barker, made the jump to theaters with Backrooms and Obsession, and now the film industry is abuzz with speculation that they have changed Hollywood forever.
For its part, the scholarly production model has been fundamentally unbothered by the platformization of creative production. The recent trend toward preprints, some of them on platforms like ResearchGate, Academia and Google Scholar, has been welcome, but these have mostly sped up the consumption of journal articles. They resemble markets for finished papers more than they resemble other platforms. The best papers and books still aspire to be published by the best journals and presses, most of which existed in the fully analog era. Of the top five economics journals, Econometrica and the Review of Economic Studies are jointly the youngest at 93 years old. Economic geography is doing no better: our flagship journal was founded in 1925. It seems weird to me, in a bad way, that no top publisher has emerged since the early twentieth century. Either our established publishers are extraordinarily fast at adopting publishing innovations, or the publishing model has not fundamentally changed.
I submit that the field can get better at what it wants to do, and find bigger audiences, if it embraces three aspects of the creator economy: radically transparent research methods, interactivity, and the distribution of content via personal brands.
A key difference between the creator economy and linear release is transparency about production. Creators show the work being made; linear release shows only the result. The extreme version is the livestream, a completely unedited product that leaves no gap between backstage and on stage. But even when they are not streaming, creators use longform formats, podcasts, and social media to bring the audience into what is happening backstage.
In academia, radical transparency about personal lives serves no obvious purpose, and could alienate academics from their audience even further. A fully transparent research and writing process, however, does seem to be the logical final step in the current movement toward transparent research.
In the wake of the replication crisis, many academic communities have moved toward greater transparency: preregistration, open data, open code, preprints. Each step is welcome. Each is also glaring in how far it stops short of transparent research practice. We have the tools to capture every stage of the research process and preserve a complete record of how a paper was made. We have simply chosen not to.
The arrival of AI has produced a range of unsatisfying institutional responses, from the naive (outright bans) to the obscure (disclosure requirements nobody reads and nobody can verify) to the unenforceable (rules premised on a detection capability the enforcer does not have).
A recent example from geography's leadership: the editors of Progress in Human Geography now require every submission to carry a detailed declaration of AI use, or non-use, addressing preparation, writing, and formatting separately. The declarations will be published alongside the articles. It is a well-meaning policy, and it is unlikely to deter anyone actually taking the cognitive shortcuts it worries about, particularly since the same editorial instructs that manuscripts should "at most, reflect only light use of AI in their preparation and writing." The editors concede the point themselves: they recognize the difficulty of policing AI use, and acknowledge that there are no clear lines demarcating the zone of irresponsibility.
Videographic transparency may be the better path. Rather than requiring researchers to declare their AI use, why not record the research itself, making every keystroke a matter of record? This would shift the burden of demonstrating good practice onto researchers, where it belongs, rather than adding a further verification task to the workload of editors who are already overwhelmed. It would also, I suspect, produce a healthier relationship to AI as a research method than the current message, which amounts to: tell us everything, and it had better not be much.
I want our community to incentivize innovative and responsible use of AI so that we can produce better knowledge about human and economic geography. For inspiration I look beyond the academy to the wider creator economy, where, without the ability to impose hard rules, communities are enforcing their own standards of AI use.
Something like this is already happening. In July 2026 Substack, facing the same flood of generated text, declined to ban anything. It integrated an AI detector, let readers scan any post, and added an optional "How I make this" statement where writers describe their own process. The stated principle was that the use of AI is not necessarily a problem but a lack of transparency around it definitely is. The platform imposed no threshold and no penalty. It made process visible and left readers to sort it out, and they have: an analysis of a few thousand posts found that the writers who switch detection off are, on average, the heaviest AI users. The setting became a signal, and the community started reading it.
Platform tools allow researchers to get much higher quality feedback from the community while research is being produced.
Under the analog system, external feedback is available only at dedicated moments in a research project, and almost all of it arrives after the paper is finished. That timing creates real inefficiencies. Researchers must reconstruct a process they completed months or years earlier in order to answer a reviewer, and they have almost no opportunity to improve the work while it is still moving.
Digital platforms gave content creators access to timely information they could act on immediately. Comments, likes, and subscriptions all send signals that can be folded straight back into the work, reducing what Richard Caves called the "nobody knows" problem in creative production. It is far easier to anticipate what will work when the market has already reacted to a prototype, whether a mixtape, a product image, or a trailer.
To be clear, I am not looking forward to the day when academics ask their followers to smash the like button. As I start this project I am not confident I will charm a single viewer into giving me feedback at all, and I am curious to find out. But I am broadly committed to the idea that a community of well-informed people might benefit from having more of that information earlier.
A creator is a personal content brand, with an identity and, in many cases, a personality. If personal brands strike you as unbecoming for academic researchers, consider how personalized our profession already is.
We tell ourselves that scholarship is about facts and concepts, and then find ourselves unable to discuss a fact or a concept without naming the person who came up with it. Tenure is an entirely individualized process even though the work in a candidate's vita usually is not. Visit the website of any department or research institute and you are guided toward the profiles and accomplishments of individual academics rather than teams. And we hand out more individual awards than almost any industry I can think of, including professional sports.
Double-blind review is one thing. Beyond peer review, let us not pretend this is an anonymous line of work.
If we accept that this is a field where production is already personalized, it is strange that so few academics have embraced the creator economy directly. What would doing so look like? Principally, it would mean that the academic's own brand, rather than their university or their publisher, becomes the main channel through which readers encounter their work. It would also mean treating legitimate academic production as something broader than the set of things that carry an impact factor.
A few economists have already done this. Paul Krugman, Tyler Cowen and Alex Tabarrok all reach large audiences directly, and Noah Smith migrated out of academia into the creator economy entirely and no longer holds an affiliation. They are outliers in economics, but they are a road map. So is Andrew Gelman, a political scientist who, unlike the others, covers a narrow range of topics alongside a set of co-authors at Statistical Modeling, Causal Inference, and Social Science. The closest thing I can name in geography is the GLaD podcast, which is excellent, but not a case where the producers' personal brands dwarf their institutional affiliations. Apologies to any economic geographer I am overlooking, but I think it is fair to say we have not leaned into this.
Those examples point to a second feature of an academic creator economy: it involves much more than articles and books. Short-form posts, links to interesting work, podcasts. Academics are perfectly capable of working in these formats, to the extent they have something interesting to say. But the incentives to publish in top journals, and eventually to hold positions at those journals, mean such work stays a side project until it succeeds and often long after.
This is a missed opportunity, and particularly so for economic geographers. We are a varied and heterodox tradition that prides itself on reading more widely than mainline economists do. Why, then, are we cultivating an audience only among the readers of academic presses? Should we not want to say something to at least a few of the hundreds of thousands of people who subscribe to Paul Krugman's Substack? Things are not going so well for us that we can afford to turn them away. Connecting with those audiences does not require abandoning peer-reviewed research. It requires deciding that reaching them is part of the job.
Disembedded is a set of experiments along these lines. Each one will leave a fully transparent archive of a research and writing process, using technology to make a contribution to the field. The first will use AI tools to speed up and improve my ability to fill a research gap, and it will hopefully benefit from peer feedback delivered through the tools of streaming platforms.
If it fails, let that reflect on my own abilities as a scholarly content creator rather than on the promise of economic geography in the creator mode.
Patrick Adler · Department of Geography, University of Hong Kong
23 August 2026