WIP video about Networks - Kindly Requesting Feedback
Added 2025-09-27 08:52:11 +0000 UTCDEADLINE FOR FEEDBACK 29/09 - 1:00 PM UK TIME
We've got an exciting video coming out very soon!
We’d love to hear your feedback and suggestions as we wrap up this video all about Networks!
It's rough around the edges as we still need to shoot some MOTS (Man on the street) and finish final anims etc. but as always we appreciate you watching and reviewing this cut!
Is there anything wrong? Any other questions? Anything more you'd like to see? Explanations you think don't do a good enough job or lack depth? Let us know!
Thank you so much!
-Team Ve
Comments
As someone who pays to be here, I was a bit miffed to have to sit through the paid advertisement...
Kate Proctor
2025-09-30 02:53:02 +0000 UTCThanks Christian - we took your advice and decided it wasn't worth a reshoot. If it was just in a voice-over rather than presenter-to-camera section we probably would have redone it though! James, writer for this video
Veritasium
2025-09-29 17:31:05 +0000 UTCThanks Charles - will pick up that interruption in the final cut! And yes - we really tried to bring out that link to social media, without over-explaining it at the same time
Veritasium
2025-09-29 17:26:47 +0000 UTCgreat point - and we did at one point consider splitting the video up! But on balance decided that we could give the key takeaways - all of which are linked by the underlying theme of small world networks - in 30 mins. So we opted for the single video in the end. James, writer for this video
Veritasium
2025-09-29 17:25:00 +0000 UTCwell spotted - we've updated the 4 times to 3 times! we didn't manage to implement the "web two" feedback - on balance we felt this was probably ok! James, writer for this video
Veritasium
2025-09-29 17:23:27 +0000 UTCThanks Timothy - great feedback which we've incorporated into the final cut James, writer for this video
Veritasium
2025-09-29 17:22:07 +0000 UTCThanks Michael - as you say Dr. Strogatz tells the story so well, and with such clarity. James, writer for this video
Veritasium
2025-09-29 17:21:44 +0000 UTCreally interesting point. In our disease model, we use an infection factor which quantifies the likelihood of one infected person infected another. Perhaps a similar parameter (an evangelism factor?!) for information transfer would work. I suppose we'd also need to consider a receptivity factor... and maybe even an immunity factor there too! Food for thought! We won't be able to squeeze all this into this video, but thanks so much for sharing.
Veritasium
2025-09-29 17:20:44 +0000 UTCThanks Lionel - we're tweaking the end of the video to reveal those results you mention more clearly (and the sound issue!) James, writer for this vid
Veritasium
2025-09-29 14:36:50 +0000 UTCyes great point - in the first simulation we cover that Watts & Strogatz did - and in simulations by other researchers - shortcuts generally undermined cooperation. The key difference with the subsequent experiments Watts did was that these were no longer simulations - they were played with real people. And it turns out real people don't behave exactly like a simulation! We've taken this feedback on board and are gonna draw this out more explicitly 🙏 James, Writer for this video
Veritasium
2025-09-29 14:28:18 +0000 UTCI’ve always been confused by two concepts which sound similar but which are completely different: “degrees of separation” and “degrees of freedom”. Here’s an explanation which helped me understand the difference between the two concepts: https://chatgpt.com/share/68da3234-adc8-8002-b2de-44af0e12e8bf
Stephen C
2025-09-29 07:19:39 +0000 UTCAt 11:42 out of nowhere Derek Just says ing.
Liam
2025-09-29 05:09:40 +0000 UTCExcellent
Beth
2025-09-29 03:15:04 +0000 UTCHahaha I think the key would just be to make something quickly and not use the resources we have at our disposal, which is entirely possible. But a lot of our episodes are being watched on TV now (I think it's the most common method of watching them). So some people appreciate the slick production values. I think I know what you mean though. You want a sense of raw, immediate, authenticity. We'll try to find ways to insert that back in.
Derek Muller
2025-09-28 20:16:46 +0000 UTCthanks for your detailed and timestamped feedback!
Derek Muller
2025-09-28 20:12:21 +0000 UTCI wonder how the influence of evangelists can be included in the models. I don't mean religious evangelists, but don't intend to exclude them either; the fact is that some people, by proclivity, by the strength of their beliefs or clarity of vision, by their willingness to play others as patsies, or simply by their eloquence and persuasive powers, are unequal (and favored) participants in any communications game. (I know that the list I just wrote is not comprehensive.) But for social networks, how do we model the presence of powerful individuals beyond nodes and edges?
Paul Weiss
2025-09-28 19:58:41 +0000 UTCThanks for your detailed feedback - and for running the AI review. I just want to respond to a few of your points: Overall: yes comments change the video before it goes out. We can easily re-record voice over, we can cut sections out. In cases of factual error we will re-record even on a tight timeline. In our last video, for example, we totally overhauled the ending based on a Patreon comment. 1. We would love to give more time to review. This is something we’re working on. In the past it has sometimes been 12-24 hours. I know 48 hours is not much better but we are working on increasing it. Each video is like an intricate meal, prepared by multiple chefs: the writer, animators, editor, and presenters. It’s challenging to make it all come together well before the deadline. We could provide previews earlier but then the video would be less complete. We try to get the video to a reasonable stage before sharing it for feedback. 2. Noted, I’ll see if we can beef up the definition of clustering. 3. Agreed on Prisoner’s Dilemma - this is partly because we have a full video on it, but I’ll see if we can expand in this video. 4. That would be a good idea. I’ll see if we can arrange going forward. It would probably have to be an AI transcript so it might contain errors. 5. Good idea and thanks for doing this. 6. This is always done in the YouTube video description. 7. I believe the sims were vibe coded and links will be available in the video description, maybe we should call this out. 8. OK, we’ll tackle this. Again, we really do appreciate the time you’ve taken to provide feedback and we’ll do our best to implement the above notes.
Derek Muller
2025-09-28 19:51:07 +0000 UTCNice video. One general thought I had was that in the real world of human connections, what leads to a small network is not just the flat "small network" effect, but also the hub effect mentioned later in the video. After all there are "multipliers" in society who tend to know more people across more diverse networks. At 31:28 it is mentioned that "Watts ran experiments where players could decide who they play with". Unfortunately the results of these experiments aren't presented in the rest of the video. Production topics: 11:42 overlap of sound bits between different scenes
Lionel Pöffel
2025-09-28 12:20:13 +0000 UTCExcellent, informative, eye opening; a new nugget to think about. Also, Dr. Strogatz is a great math n science communicator. This content was captivating. Go with it.
Michael G Rourke
2025-09-28 03:25:17 +0000 UTCI would find it really helpful if each person (including yourself) had a “name tag” on screen (long enough to read) as they are introduced. BTW, this show wasn’t “my thing” when it started, but I watched all and enjoyed it a lot. I will be watching it several more times too. My pastor is also an account with a major international company (he’s a “ numbers guy”) and a “hub”. I think he”ll enjoy it! Sorry if there are mistakes here; my iPhone only displays one line of text at a time.
Timothy-Douglas Alvey
2025-09-28 03:13:55 +0000 UTC@12:05 - "almost 4 times" should really be "3 times" - unless the graphic is incorrect (graphic shows 20k -> 57k ) @15:43 saying "the web, too" sounds like "web two" (which is after this time period you're talking about) -- I would use "also" instead .. but this is verging on a nit-pick ;)
Gnare
2025-09-27 21:00:57 +0000 UTCI liked the video but some how it felt like 3 or so networking videos that did not tie together very well. My response is an over all feel. I think Ronan M with the help of AI does a good job of giving you a very detailed review, (Interesting use of AI). His review I think says the video needs to be much longer or maybe broken into more pieces?
Kat Seibert
2025-09-27 19:24:27 +0000 UTCAnother informative video! Thank you! At about 11:42 there is an interruption that might be better edited out. This story seems to hint at an idea that I have, and that is that we do not yet understand how to live and thrive with "social media". Sometimes it is an echo-chamber of comments for all the reasons that this video states. Your advice on seeking positivity is encouraging in that we might figure out how to turn off the screens and live a little. Thank you!
Charles S. Cook
2025-09-27 18:20:04 +0000 UTCJust some thoughts: Overall, please answer if any of our comments will change your already filmed video? Some thoughts: 1. If you want a review, give us more time to do it. We effectively had less than two days to support your request. 2. You talk about clustering as if it’s obvious. I have an intuitive grasp, but a definition and example would greatly help. 3. Give more thought to how the Prisoner's Dilemma is presented, as you rushed through it, as also pointed out by Christian, below. 4. Include a timestamped transcript like those on YouTube. 5. Run your transcript through an AI, asking it to report on clarity, truths, and errors. 6. Provide links to the various articles referred to in the video that are downloadable. 7. What sim program did you use? I want to run it also. 8. At the 24:00 minute mark in the video, you verbally state a play against one opponent while showing many opponents, which is confusing coming off a one-versus-one scenario. 9. At the 31:43 point, you state that the choice of players means cooperation flourishes. How about the leftover non-cooperators? Here is the Gemini AI Review based on a transcript I created from your video. Gemini: Of course. Here is a review of the provided text from the Veritasium video on networks, focusing on clarity, truths, and errors. Overall Assessment The text provides an excellent and engaging overview of several fundamental concepts in network science. It successfully translates complex academic ideas (from sociology, mathematics, physics, and computer science) into understandable narratives using compelling examples. The clarity is generally high, and the core scientific truths are well-represented. There are a few minor errors and oversimplifications, but they do not detract from the overall quality and educational value of the piece. ________________________________________ Clarity The text is exceptionally clear for a general audience. It employs several effective strategies: • Storytelling: It begins with a relatable story (connecting a falafel salesman to Marlon Brando) to introduce the abstract concept of "six degrees of separation." • Analogies: It uses simple analogies, like people arranged in a circle, to explain complex network topologies (clustered vs. random). • Concrete Examples: It grounds abstract concepts with real-world examples, such as the Kevin Bacon game, the C. elegans nervous system, Chicago O'Hare airport, and HIV prevention strategies. This makes the implications of network structure tangible. • Visual Language: The description of the simulations ("wiggle around and like find their location") helps the reader visualize the dynamic processes being discussed. ________________________________________ Truths (Key Scientific Concepts Covered) The video transcript accurately covers several foundational truths of modern network science: 1. The Small-World Phenomenon: The core idea that most people in a large network can be connected through a surprisingly short chain of acquaintances is true. This was famously explored by Stanley Milgram in the 1960s and is often referred to as the "small-world problem." 2. Clustering and Path Length: The text correctly identifies the two key features of real-world social networks: high clustering (your friends are likely to know each other) and short average path length (the "six degrees" phenomenon). 3. The Watts-Strogatz Model: The explanation of how a few random "shortcuts" can drastically reduce the average path length of a highly clustered network is an excellent summary of the seminal 1998 "small-world" model by Duncan Watts and Steven Strogatz. The statement that just 1% of shortcuts can drop the path length by over 80% is consistent with their findings. 4. The Strength of Weak Ties: The text correctly connects the "shortcuts" in the network model to the sociological concept of "weak ties," popularized by Mark Granovetter. These acquaintances act as bridges between different social clusters, providing access to novel information and opportunities. 5. Scale-Free Networks and Preferential Attachment: The description of hubs and the "long tail" distribution of links is a good introduction to scale-free networks. It correctly identifies the two key mechanisms of the Barabási-Albert model: growth (the network expands) and preferential attachment (new nodes are more likely to link to already well-connected nodes). 6. Hub Vulnerability: The text accurately portrays a key property of scale-free networks: they are robust against random failures but highly vulnerable to targeted attacks on their hubs. The example of the O'Hare airport shutdown causing cascading failures is a perfect illustration of this principle. 7. Network Structure and Cooperation: The discussion of the Prisoner's Dilemma on networks is accurate. Research has shown that in clustered networks, cooperators can survive by forming groups, whereas in more random networks, defectors can more easily exploit and dominate. The insight that online environments can resemble random networks and thus discourage cooperation (the "keyboard warrior" phenomenon) is a valid and interesting application of this theory. --------------------------------- Other Comments -- Errors? 1. Oversimplification of Network Models: o The video presents the Watts-Strogatz (small-world) and Barabási-Albert (scale-free) models as the primary ways to understand networks. While foundational, network science has undergone significant evolution. Real-world networks often exhibit properties of both models and other complex features not captured by them. o The statement that preferential attachment naturally leads to the hub-and-spoke structure is true, but the text rightly adds the nuance that it isn't the only mechanism. This is an important qualification. 2. Causality in Cooperation: The text explains that in Duncan Watts's experiment, network structure seemed to have no average effect on cooperation because clustered networks were prone to tipping to either all-cooperation or all-defection. This is a subtle but crucial finding. The video handles this well, but could have emphasized that while the average outcome was the same, the volatility and predictability of behavior were vastly different depending on the network structure. A clustered world is a more fragile, "all-or-nothing" world.
Ronan Mandra
2025-09-27 17:54:03 +0000 UTChe can't wait to talk about citations
L
2025-09-27 14:54:25 +0000 UTCCompletely isolated communities would break this, yes, but I would argue that it is really really difficult to fully isolate :). Traveling merchants on land and sea historically have been a thorn in side of isolationists for millenia, always finding a way to poke through isolation boundaries. Sea traders in particular can provide a lot of shortcuts with all the ports they visit. You know a baker, who knows their supplier, who trades with a merchant ship, etc.
chromicacid
2025-09-27 14:53:15 +0000 UTCAfterthought: - Can persuasion be modeled as a small world network between correlations? True, one correlation does not imply causation, but persuasions always seem to be a convincing series of correlations so that multiple correlating links are created. Not sure I like this idea. Feedback: @1:36 "anything malevolent now has conduits that it didn't used to have" If we make the world more connected, surely people will know each other and wars will end. Just like making nitroglycerin safer. @9:33 paper Reference? Note: I'd like to find out what "KBG" is in the figure title, but I'm having trouble finding the paper. Googling shows either a genetic disorder or autocorrects to soviet secret police. @10:32 "73 days" Are you accounting for Madagascar? If you haven't played Plague, Inc, nm. @14:50 Could you interview a data analyst at Incogni? I'm wondering if personal data propagation through data brokers benefits from small world networks. Seems like a relevant ad. @16:51 "because they resembled the hub of a bicycle wheel" Video not showing a bicycle wheel. @18:58 "screen saver" Yes please. @27:37 "the public goods game" Can we get a rules summary and how it differs from the prisoner's dilemma? The small world prisoner's dilemma was a negative outcome and got a rules breakdown with animation, but the small world public goods game (positive counterclaim) did not have placeholders for either. Audio: @11:41 "need less than one percent" Audio: Accidental insert? 0.1sec @11:54 "ranked the 63rd" Audio: Misalignment with video, about 4 seconds early. Video showing rank 63 doesn't show until ~4 seconds later. @17:21 unintelligible Audio: Sounds like "but the girls". Hope your subtitles are prepared to clarify. Thought experiment: - Modify the prisoner's dilemma by having a defection also increment an animosity counter - The animosity counter modifies the opponent's behavior by increasing the likelihood of defection (thereby increasing the likelihood of only gaining 1 point instead of 5) - Cooperation would decrease the animosity counter by 1, regardless of the actions of the opponent
chromicacid
2025-09-27 14:44:55 +0000 UTCI think that's why he referenced the game theory video for more info
chromicacid
2025-09-27 14:43:09 +0000 UTCI doubt very much I have ANY degree of connection to an aborigine in the Amazon basin who has never been contacted by the outside world. I'm sure most of those connections are ONLY local clusters. So the Small World hypothesis depends completely upon breaking geographic isolation boundaries, something that rarely happened in the old world. For instance, there would have been an extremely low likelihood of any small world connection between the indigenous people of the Americas and a poor dirt farmer in pre-Columbian days. It's also unlikely that anyone on earth has a small world connection to Grok, that famous guy on a world in the habitable zone in the Alpha Centauri cluster. So the Small World depends on random connections that cross natural boundaries. And that's how I've always thought about it. Now I'll finish listening to the video, because I'm sure you point that out. ;)
Mike Mitchell
2025-09-27 14:29:40 +0000 UTCI'm confused. I thought it depended on the shortcut rate whether people cooperated. What's different between those earlier results and the ones presented at the end of the video?
Christian Leichsenring
2025-09-27 11:52:15 +0000 UTCI know you want to rush the explanation of the Prisoners Dilemma because everybody and their mom already knows it but as it is in the video, it's maximally confusing to anyone who doesn't. A chest? Points? Prisoners? What's the scenario here exactly?
Christian Leichsenring
2025-09-27 11:24:35 +0000 UTCGeneral remark, I wish you forgot everything you learned when making documentaries for TV networks and made YouTube videos again. I don't know how you'd do that though. Maybe some heavy drinking or hard drugs?
Christian Leichsenring
2025-09-27 10:59:33 +0000 UTCItalian and German pronounce the z like in pizza and nazi. As most people implicitly know but promptly forget when it comes to any other word in these languages. So the newspaper is pronounced "tseit" not "zeit". The "die" and the "ei" were pronounced correctly though, props for that. No, not worth a reshoot but, you know, next time.
Christian Leichsenring
2025-09-27 10:54:37 +0000 UTCAt around 11:45 there seems to be a bad cut. You start describing the next section before the current section has finished.
Holger Wolff
2025-09-27 09:07:39 +0000 UTC
