Having watched [[Jaimungal & Carroll (2024)]], I wanted to both nuance and generalize an essential point that I believe Sean Carroll was making: >I don't care who you are. I don't care whether you're an actor or a person on the street or a PhD in physics. That is entirely irrelevant. But I do care how much work you've done, put into thinking and learning about physics as it is understood. You're asking me to put aside time in my life to read your theory and pay attention to it. Well, first, read my theory and pay attention to it. My theory is called the Standard Model of Particle Physics ([[Jaimungal & Carroll (2024)|Jaimungal & Carroll, 2024]], [00:06:00](https://youtu.be/Js5YcGrr_UY?si=4pQdqenEmuNO1y7K&t=360)). The context for this quote is that many lay people with an interest in science (including deluded hobbyists and genuine cranks), and occasionally a few scientists who stray outside their domains of expertise (or become senile), produce more radical claims that can be humanly debunked. Hollywood actor Terrence Howard is one such example of a lay person who famously "attempted to debunk the Pythagorean theorem, claimed he can kill gravity, said he does not believe in the number zero, and claimed he remembers the events of the day he was born", according to his Wikipedia entry. Ironically, science communicators such as Sean Carroll are partially responsible for this: as advanced physics get popularized to a lay audience, more people fall prey to the [[Dunning-Kruger effect]]. Likewise, LLMs increasingly enable lay users to acquire just enough basic knowledge in a given field to believe they can make meaningful contributions. A common example is patients with no medical training who second-guess their physicians using AI-generated self-diagnoses that lack clinical context, such as family genetics or travel history. The tension at the heart of Sean Carroll's quote is as follows: when evaluating specialized claims, such as novel physics theories, experts are faced with a dilemma. - On the one hand, a claim should be examined on its own merits, not those of its author. Hence the motivation for the double-blind peer review process, by which the author's identity is kept hidden from the reviewer (and vice versa) to avoid interpersonal biases. - On the other hand, the time that experts can allocate to reviewing claims is finite, which implies that worthwhile claims must somehow be triaged before they get reviewed. A crude approach to doing that would be to randomly sample $n$ of every $N$ claims, giving each one a fair (albeit low) chance of being reviewed. On average, though, this will mostly yield ideas that are near the mean of the distribution, which is to say that they are likely to be wrong. A more sophisticated approach might be to let specialized AI agents perform the first review and filter out claims that are not worth a second review by human experts. Sean Carroll advocates for a *proof of work* approach. At face value, this is sensible. To challenge Einstein's theory of general relativity, for instance, a person needs to understand it *at least as well as Einstein did, and presumably better* (to be able to see past it). That is not achievable at the hobbyist level. But how do we assess how much work somebody has put into their claim? This is where the Sean Carroll interview stops short of providing guidelines. For better or worse, a common proxy for proof of work is a person's academic and professional credentials. A lay person who watched a few YouTube videos, and perhaps read a physics textbook, might understand enough to overstep into Dunning-Kruger territory, but should not be assumed to possess enough proficiency in the field to be relevant. The likelihood that they can correctly identify novel ideas that push the boundary of knowledge, when they are themselves nowhere near that boundary, is infinitesimally low (not zero — but also not worth spending scarce review resources on). On the other hand, somebody who studied physics to postdoc level, has published in peer-reviewed journals, and works full-time as a professional physicist, meets the necessary, *although not sufficient*, knowledge prerequisites to start challenging existing theories (at least in their domain of specialization, or immediately adjacent to it). This is why I find it difficult to hold Sean Carroll's view that we should not care about who authored a claim, no matter how noble a principle that is. The practical reality is that *proof of work* is very highly correlated with somebody's academic and professional credentials, to the degree that one's *identity* and *credibility* in a given field are functionally indissociable. And I suspect Sean Carroll *does* care about who authored a claim — and that the distinction he makes is purely superficial semantics. Srinivasa Ramanujan is one notable exception that comes to mind, being a child prodigy in mathematics with little formal education. Had G. H. Hardy and J. E. Littlewood paid no attention to Ramanujan's submission in 1913, some of his groundbreaking theorems may not graced our world. But what makes Ramanujan so famous is precisely the exceptional nature of a case like his — where someone lacking the usual credentials was yet capable of pushing the boundaries of knowledge. I do not believe Ramanujan's example invalidates the fact that giving attention to far more uncredentialed individuals with extraordinary claims would be a net benefit, given that it would divert scare expert attention from those with legitimate claims.