Theories of Everything with Curt Jaimungal - I’m Giving Away $5,000 to Explain Hard Physics (and AI)

Starting point is 00:00:00 There's plenty of great introductory-level YouTube videos, but if you want to understand graduate-level physics or AI or philosophy, then those options get quite thin. For instance, where's the animated explainer video of geometric quantization or infinity categories? I'd personally like to help change that, so today I'm announcing the core contest. Core stands for competition for outstanding research explanation, and this year, Core 1 is our first year. The goal of core is straightforward to encourage the creation of outstanding video explanations on advanced topics and to hopefully raise the bar for specialized science communication online. This is especially for you if you're a grad student, a postdoc, a professor, a researcher, or anyone deeply familiar with advanced topics and you're keen to explain them.

Starting point is 00:00:48 In fact, if you want to give an animated explainer video of your PhD thesis or your recent paper, that's fantastic. Monetize those tiers. If it's not clear already, Core is basically an analog of the excellent Summer of Math Exposition, which was a contest to emboldened people to make explainer videos on topics where the submitter required just a tiny nudge to provide that activation energy toward creating something they may not have otherwise. I think the world is much better off, given the presence of the Summer of Math Exposition contest, and I'd like to do something similar for advanced physics, AI, and philosophy.

Starting point is 00:01:22 So, Core is an initiative by this channel, Theories of Everything, the podcast where I, Kurt Jaimungle, host explorations into these sorts of topics, and of course, there's no shortage of content about AI applications or physics news, like black holes are crazy. Yeah, we know, bro. Core is specifically for the technical, graduate level, foundational material, the theory that practitioners often skip over, even though it determines whether we actually understand what we're doing and why it works. works, what are the edge cases, and so forth. Now, to participate, just create a new public YouTube video explaining a concept within one of our three categories, theoretical physics, foundational AI, philosophy of physics. Now, if your topic sits at the boundary, that's fine, submit it, I'll even have a category for doesn't fit into boxes or what have you. The community, including you, will judge the best five entries overall, regardless of area. So let's get into

Starting point is 00:02:21 some examples, what would count? So for theoretical physics, maybe a video on the rigorous definition or difficulties of such of the Feynman Path into growth, maybe constructive QFT, maybe something about the Coleman and Dula theorem and how supersymmetry is seen as a loophole, BRST cohomology, asymptotic safety, cobordism, hypothesis, etc. For foundations of AI, there's plenty out there on how to use AI, but where is the compendious video explaining why neural scaling laws work or what PAC-based bounds actually tell us. You can also talk about the Lost Landscape Geometry, singular learning theory.

Starting point is 00:02:58 Elon Bernholz here has a thesis about language, LLMs, and cognition. Feel free to make a video on that, if you like, or Yosha Bach's theory of consciousness and how it relates to LLMs. It in no way needs to be tied to this channel theories of everything. By the way, these are just some examples that occur to me, so I'm showing you. There's also double descent and over paramatrization, etc.

Starting point is 00:03:18 For the philosophy of physics, you could talk about the whole argument and Defeomorphism invariance in GR, or structural realism, epistemic versus ontic, or the ontology of QFT, particles, fields, or algebraic structures. What about laws of nature? Is it humian? Is it governing? Great. Okay, so now you have some idea, and of course, these are just ideas.

Starting point is 00:03:38 Feel free to come up with your own. Right up front, you, the creator, keep full ownership and copyright over your video. Also, submitting to court doesn't prevent you from putting your work elsewhere. There's no exclusivity here. Even if there are multiple contests, you submitting to theirs, doesn't preclude you from submitting to mine. It's an open relationship. I'm not the jealous type.

Starting point is 00:04:02 Now, submissions are open today, now, until March 31st, 2026, 1159 p.m. Toronto time. After the deadline, we'll have a peer review phase, and you'll be shown entries to score on a scale, say 0 to 5, guided by the question of how valuable is this entry to the space of advanced science explanation compared to what's already out there. It will be inspired by the one used in the admirable summer of math exposition. Again, I highly recommend you check out three blue, one brown. Now, this means that every single submission is guaranteed to be watched and scored by real people, your peers. This will surface the top entries for final judging by our panel of professors in relevant fields yet to be announced. Okay, so what makes a winning entry? We're focusing

Starting point is 00:04:49 on three rough criteria. Intellectual depth, which is roughly accuracy, substance, and rigor, insightfulness, which is more like, have you shown hidden connections, and have you elegantly simplified something without losing the core truth, an aha moment, and impact, the significance or relevance of this topic. So the three eyes, intellectual depth, insightful, and impact. When voting, you don't need to weigh each of these criteria on separately. You just ask yourself, overall, how much did this entry move the needle for advanced science explanation online? To be considered for prizes, we ask you to participate in the peer review phase. You can follow my substack at kurtjimungle.com. For more details, I'll place a section at the top for core

Starting point is 00:05:36 C-O-R-E with a full FAQ. I'm personally putting in $5,000 USD into the process. to put my own skin in the game in a sense, but also because I think the proliferation of advanced material on physics, AI, and philosophy will be far more beneficial in the long run. I'd rather fund this directly than wait for sponsors who may want to shape what winning looks like. This means the top five winners will each receive at least $1,000 USD. Now I say at least because there may be sponsors watching who want to donate to the prize pool. I'll also feature the winning videos and creators in a dedicated showcase right here on the The Theories of Everything channel. The only requirement to enter is that at some point in your YouTube video, you state that this is an initiative by

Starting point is 00:06:21 Theories of Everything with Kurt Jemungle and place a hashtag Core 1 in the title with a link in the description. You don't have to talk about how dashing and charming I am. Just a quick mention of the channel visually or audibly is fine. If you need inspiration and want to team up, then I just created this Discord, a core channel on Discord. The link is in the description. It's a place to connect with other potential entrants. Think of it as networking without awkward conference small talk. You can also find links to the full rules at kirtjemungle.com. All links, including the submission form, the Discord invite will be there.

Starting point is 00:06:56 Note that on substack I also write frequently about physics. I talk about philosophy. I give my various thoughts on topics like language and I email out podcast episodes ahead of time. You don't have to sign up on my substack to qualify for anything. You can feel free to just bookmark the relevant pages without subscribing. Now, a word of advice. You may think, but Kurt, what if my topic is too technical?

Starting point is 00:07:20 Okay, that's great for two reasons. Number one, the fact that is technical implies that there are very few people who already have videos on the same topic, which means you're filling in a gap. Now, that's fantastic. That's what this contest is about. Number two, if you're even asking the question like, hey, Kurt, what if my topic is too technical, well, that means that the topic is so specialized it's grabbed you. That in turn indicates you have a passion. For instance, if you want to talk about why decider space is more difficult to realize

Starting point is 00:07:51 a CFT duality than anti-Dicitor space, well, that's super interesting. There's not even a synoptic video about ADS CFT, let alone DSCFT. Now, if you want to talk about something seemingly obscure regarding philosophy, such as constructive empiricism, then that's fantastic. In fact, I encourage it. You may say, look, Kurt, I don't have the answers. Cool. Firstly, if it's an extremely advanced topic, then almost no one has all the answers, but you do want to be accurate, of course, so feel free to lay bare your research and be honest about what you don't know. You'll even see that on this channel, I have various examples, like I went through Agrippa's Trilemma on this Can Physics Explain Its Own Laws video here? Or I talk about AI scaling laws with Max Tegmark and

Starting point is 00:08:36 Jeffrey Hinton or the latest in cosmology with Roger Penrose. There's even a lecture on indexicals and Hempel's Dilemma, or this is my attempt to explain string theory here. There's nothing wrong with genuinely showing your research. So, if you have expertise in an advanced area and a passion for explaining it pellucidly, I hope you'll consider creating something for Core 1. If you're a professor who wants to judge, reach out. If you're a company who wants to contribute more to the prize pool, get in touch.

Theories of Everything with Curt Jaimungal - I’m Giving Away $5,000 to Explain Hard Physics (and AI)

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.