Categories
New

Things 2026 Q2: Fisherian runaway, Androids, effortful music videos

Quotes

Here’s some nice quotes that all draw out aspects of the Dunning-Kruger bias:

Idealism increases in direct proportion to one’s distance from the problem

– John Galsworthy

Ignorance more frequently begets confidence than does knowledge

– Charles Darwin

Everything looks simple when you don’t know the first thing about it

– Kevin Williamson

Having established that, I now return to confidently holding forth on things I’m not an expert in, probably annoying Darwin in particular. I leave you to judge how sensible that is.

Super-normal stimuli and Fisherian runaway

Bear with me on this one as I invoke the principles of sexual selection to explain burgers that are too tall to bite.

Part 1: Gestalt Perception

Perception usually feels effortless, but the process that makes it work is strange and subtle. Dozens of contextual clues (sight/sound/smell, movement/light/texture, size/silhouette/colour, and many more) are all subconsciously processed, usually based on previously observed patterns, and our conscious attention is directed to whatever that process deems the most salient aspect. For example, you might see a poster you’ve never seen before, but it has a picture of an actor you know, some words in an elaborate font, a release date and some small print, and without effort you recognise it as a film poster, as well as a good guess for the genre and if it’s likely to be worth paying more attention to.

The way many parts contribute subconsciously to our overall perception is summed up by the slang term ‘vibes’, but over the decades other terms have approximated this idea – ‘vibrations’ of course, but also in different contexts bag, energy, flavour, and more recently aura or “x-coded” (for something with various signifiers of x). Similarly people say an idea ‘doesn’t pass the sniff test’, not because they can smell abstract concepts, but because of how it comes across in almost every other way.

Part 2: Supernormal stimulus

Certain aspects of perception have clear dimensions to them. Most obviously things like bigger/taller/longer but also brighter/more-textured/more-saturated, and beyond. If we recognise an object based on it being notable in some dimension, an object with a stronger version of that attribute can register more quickly and more strongly as a signifier of that object. Further, I think that obviousness biases us to rate it as a ‘better’ example of that object in some way. If it has an unnaturally pronounced example of that attribute it can become a supernormal stimulus.

Wikipedia gathers some examples from the natural world here, for example:

  • Birds will nurture fake eggs over their own if they are larger or more saturated (and Cuckoos exploit this fact)
  • Butterflies that select mates by bright colour markings will prefer orange paper model butterflies over receptive real mates
  • Herring gull chicks seeking a parental beak for food will choose a red stick with yellow markings over a faithfully reproduced yellow-beak-with-red-dot replica gull head

Of course, we can intuitively understand this without tricking small animals in laboratory settings, because we are animals ourselves. For example, a drawing of a woman with a dramatically high hip/waist ratio or a man with an implausibly pronounced shoulder/waist ratio will read as stronger and more obvious examples of each sex, even if those ratios go beyond what nature would naturally produce. Some examples:

Designs from Disney’s Hercules (1997) – exaggerated but legible:

Designs from The Book of Life (2014) – frighteningly extreme, still works:

Beyond human secondary sexual characteristics, we can experience it in pretty much everything. A giraffe with an extra long neck is extra giraffe-ey; a stretched limo is an even greater affluence signifier than a regular limo; a monster with an excessive number of impractically long and excessively sharp teeth is extra monstery.

Part 3: Fisherian runaway / Sensory Exploitation Hypothesis

Natural Selection just by itself is an incredibly powerful theory, but some things defy it – a classic example being peacock’s wildly impractical tails. Less often remembered is that Darwin also theorised about Sexual Selection: genes that produce traits that elicit sexual preference will have a reproductive advantage, even if they have no survival advantage. But this on its own doesn’t explain the peacock.

Fisher’s suggestion was that the earliest version of the ‘ornament’ (the surprising feature the species developed) initially signalled greater potential fitness in the natural-selection sense, which meant a preference for that ornament also had an advantage. If the ornament and the preference are correlated (I assume mainly because genes for both get passed on together) you can enter a feedback loop of ever-more exaggerated ornaments that are just barely constrained by natural selection.

The ‘sensory exploitation hypothesis’ proposes that preference for exaggerated traits arises from sensory biases such as the supernormal stimuli I mentioned above. It seems pretty intuitive to me that this will help drive the Fisherian runaway effect far beyond what natural selection would usually tolerate. It’s a bit like Goodhart’s law – when a measure becomes a target, it ceases to be a good measure.

So a male peacock with an even bigger, brighter, more saturated and more ornamented tail will look even more appealing to a female peacock, and both the tail and the appreciation for the tail get passed on. Thus the tail becomes ever more extreme, until it finally bumps against natural selection’s hard limits.

No points are awarded for this, but I did come up with the sensory exploitation hypothesis myself! Only after looking for research did I learn that all the above is well-established, and that point about ornament and preference being correlated is the deeper insight.

Part 4: Fisherian runaway in capitalism

I do think this same process applies to capitalism:

  • Consumers select products by their utility. Products with greater utility are more selected, more successful, and so are produced more. Natural selection at work. A Drosselmeyer nutcracker is incredibly good at cracking nuts, so it does well in objective comparisons, and people like me will recommend it to anyone that wants to crack nuts.
  • In many cases the consumer can’t perfectly assess the utility of the product before purchase, so they rely on other signals. This becomes more like sexual selection.
  • We can fall victim to a super-normal stimulus here – a thing that depicts outrageously exaggerated traits looks more appealing, even if it’s beyond all reason. A 13-in-1 tool must be better than a 3-in-1 tool. Drinking Tango gets you slapped by an orange man. This car looks so good, fireflies follow it. These two people love this product, and they are very attractive!
  • Then the Fisherian runaway can start – to compete, those products need to show the super-normal stimulus, and then the only way to stand out is to exaggerate it even further until you eventually run aground on actual utility / natural selection.

Perhaps you think I’m mad, but remember the burgers?

As a food, a burger is notable for being quite tall – it’s a stack of three (or more) things. A burger can also seem more appealing if it has more than just a plain burger inside. So what happens? Burgers that are taller and have more additional things inside look like even better burgers. Before you know it, restaurants are serving burgers that have to be held together with a skewer and cannot even fit in a conventional human mouth. It’s like eating a peacock’s tail. Metaphorically.

(I’m not sure about the correlation element… perhaps it comes in when chefs eat a taller burger at some other restaurant and conclude they should make their own burgers even taller?)

What about cars? Some people feel safer or more masculine if their car is bigger than others. In America where petrol is relatively cheap (so the natural selection pressure of miles-per-gallon is lower), this effect has reached peacock tail levels of madness as cars become ever larger. The correlation kind-of happens here because obtaining a bigger car means other people, whose cars seemed sufficiently big before, now see a need to get even bigger ones.

Thanks to Instagram and similar social media, we get a similar effect for a bunch of things:

  • Treats/drinks that look amazing go viral and drive more sales (and more social posts, driving the correlation) than those that actually taste good
  • Holiday destinations that make for good photo opportunities start to do better than those that are actually enjoyable
  • Getting more abstract: a video/post that has a very high number of views seems like it must be more worthy of our attention
  • Bit of a stretch: a famous person having an opinion seems more compelling than an expert in that domain’s opinion

So as always… watch out for that.

Podcast – The Magnus Archives

The Magnus Archives is a supernatural horror fiction podcast that ran for 200 episodes across 5 seasons from 2016 to 2021.

I got into it essentially by the strength of the writing: I was at a craft fair and found a range of stickers with intriguing phrases like “The sky ate him” and “The world is always ending”.

Unfortunately I didn’t note the artist and the work is not signed!

The framing premise is that the eponymous Magnus Archives takes statements from anyone that has experienced something supernatural, and the Head Archivist is recording these statements out loud to tape in order to log them. In this way it takes a classic “spook of the week” form.

I liked that a lot of the elements of the premise that don’t really make sense end up being narratively justified – with the exception of the very non-diegetic spooky music, which was still excellent (especially the “and then things got really spooky” musical cue in season 1).

The spook-of-the-week setup has two tightropes the author must walk, which author Jonathan Sims talks about explicitly in some of the Q&A episodes: the balance between the weekly spook and a greater narrative, and the balance between raising questions and and answering them. For me, neither one is walked perfectly throughout – but Sims (who is also the lead voice actor) does a much better job of both than I’ve seen before. Even though for me the series peaked around 75% of the way through, I still strongly recommend it if you like spooky business.

Some points to note before jumping in:

  • Episode 1 (styled MAG 1) did not do much for me, MAG 2 is a better bellwether for what the rest is like, so I recommend sampling at least up to that point
  • There is a dedicated subreddit, with a thread for each episode featuring discussion by fans and no spoilers about future content. This can be quite fun for a debrief and help spotting deeper connections between things. (Hub of most of these threads can be found here)
  • There is a dedicated wiki with plot summaries and full episode transcripts, but you need to be very careful because immediately after the plot summary you will see information that spoils future episodes
  • My very favourite bit begins with the word “Hello”. You’ll know it when you get there
  • My least favourite bit is recurring characters that are constantly angry and obstinate
  • Probably avoid if you’re averse to the “endless suffering” trope
  • There’s a follow-up called “The Magnus Protocol” which is still running and I have not yet sampled

The home page with various links is here, but it’s probably easiest to just search ‘magnus archives’ in the podcast service of your choice.

The Androids are Coming

Because I’ve been writing Things for almost 20 years, I can see how some of my posts serve as a marker of developing technology, or at least my awareness of it. For example, I posted in 2015 this Google Research post about neural networks generating images, which is I think the precursor to generative AI art as we now know it.

Back in 2009 I posted about Boston Dynamics’ ‘Big Dog’. Developments in robotics carried on, with bipedal models eventually roaming free of connecting cables and getting increasingly human-shaped. They now seem to be at something of a turning point.

Compare the performance of China’s androids in the 2025 new year show here:

… with the 2026 new year show:

My guess was that android promotional material would shy away from any implication of violence to avoid reminding people of films like Terminator. I hadn’t accounted for the fact that state propaganda might actively want to promote martial prowess, hence the above. At least we can look at what’s coming directly instead of guessing.

Terminator style threats to humanity are a natural fear but I suspect are a low-probability / high-impact event (perhaps more likely in the form of perfectly efficient and numerous state police facilitating a concentration of power into an ever smaller number of human hands, rather than a bona fide machine uprising). Warfare is also a sufficiently incentivised area that bespoke robots (rather than androids) are arising already, starting with drones.

We’ve spent centuries building specific machines to tackle specific tasks efficiently. Building an android to do similar things seems laughably inefficient – a sledgehammer to crack a nut, or supercomputer to play tic-tac-toe. But there is of course a massive efficiency: we have tools, workplaces and entire paradigms built around the human form-factor. Making an android to solve one physical task is inefficient; making one that can do any human task imaginable, with more accuracy / strength / reliability than a trained human, is incredibly efficient.

In December 2024, Citigroup published this forecast of robot numbers by type. Humanoid robots (which I’m referring to as androids) are the most notable leap, getting a foothold in 2030 before exploding by 20250:

The leaps forward in AI, most notably visual processing, seem to have unlocked a major piece in this puzzle and make the above look more plausible to me. Here’s a couple more cases (like the Chinese New Year’s Show example above) that show the improvements in these models over recent years, and so gives an idea of the trajectory.

February 2025: Helix androids putting away shopping very slowly:

May 2026: Helix’s Figure 03 working a production line (and other things):

In a very direct test of a specific skill, there are contests in China in long-distance running for androids. In April 2025, one robot completed the half-marathon in 2 hours 40 minutes.

One year later, a robot completed the half-marathon in just over 50 minutes, beating the human record of 57 minutes (which immediately seems like a less relevant comparison).

I don’t expect this to be quick, so the slow initial commercial growth in the Citi forecast does look reasonable to me. I’m certain there will be massive bumps in the road as public perception and legal / insurance systems navigate around the inevitable accidents and abuses this technology will produce. But I now feel pretty confident that sci-fi visions of robot drivers, plant-pickers, and assistants are not at all fanciful.

I’ve also been reading a lot about what people think will happen to the humans whose jobs have been replaced. To distil everything, the answer is that… the result will be a product of huge numbers of competing systems and feedback loops, and as such is impossible to predict. It’s a bit like this Financial Times prediction of GDP based on different AI scenarios, which is quite funny and frightening at the same time:

Music videos where effort is the trick

This Things is getting a bit heavy, so let’s look at some Music videos in which a beautiful visual effect is achieved, but the amount of effort that must have gone into it is tremendous.

Director: Michel Gondry
Artist: Kylie Minogue
Song: Come Into My World

Once again Gondry finds a way to visualise the repetition of the song in a video that just keeps getting more impressive:

How did he do it? Well, he explains it and the inspiration here

His explanation seems to take for granted an enormous amount of compositing work, all the more impressive given that this was in 2002 when computer-assisted tools were less well developed. That said, it’s Gondry – possibly he found an incredibly efficient way to do this that he’s not revealing here. Or he just takes that “enormous effort” part for granted because he does that regularly.

Director: Nick Goffey and Dominic Hawley
Artist: The Chemical Brothers (with vocals by Beck)
Song: Wide Open
Dancer: Sonoya Mizuno
Effects: The Mill

Starting with a close-up to assure you that you’re looking at a real human dancer, something impossible starts to happen. Make sure to stick around until 3:28 (or just skip there if you’re busy) in which the video casually does something even more impressive.

How did they do it? Surely to do that you would need some sort of motion-controlled camera or real-time movement tracking, re-render almost everything in CG and then carefully composite the live-action with the CG with pinpoint precision?! The answer is yes, that’s exactly what they did:

Director: Páraic Mc Gloughlin
Artist: Weval
Song: Someday

If you take carefully framed photos of everyday objects and architecture with repeating patterns and show them in sequence, you get a strange sort of stop-motion effect. I saw this trick first used in Michael Langan’s 3-minute video Dahlia (2008), then in Thomas Sauvin’s Recycled (2013) (seemingly no longer viewable but described here; a short snippet can be seen from 1’12” in this film festival trailer).  Páraic Mc Gloughlin pushes that idea even further with incredible smoothness, range and synaesthetic effect for Weval’s Someday (2019) music video (after about 21 seconds of build-up):

If you enjoyed that you could also try the video for IMANU – A Taste of Hope (Hallowvale) by Chris Takács and Kevin Rogerson; or enjoy spacelike slicing instead of timelike with the video for Max Cooper – Repetition by Kevin McGloughlin.

The AI Shovelware is here

Also in the category of marking developing technology, last time I wrote about the AI ‘trough of disillusionment’ referencing Mike Judge’s “Where’s the shovelware” article.

Well, true to how the Gartner hype cycle (sometimes) pans out, it turns out we were coming out of that trough.

The financial times looked back into those same trends (paywalled link) in February 2026 and found recent upticks in websites, iOS apps and GitHub repositories:

In April 2026, Naavik did a detailed overview of the topic just for games, showing not much effect on PC (although a rise in AI-discosed games), but a huge effect on mobile (especially on Android), showing the number of publishers is up, and the relative proportions of success have remained about the same.

That last point is particularly interesting because the feeling in games is that there are already way more games than the market can sustain (so the vast majority do essentially no business), so an AI-driven surge in new games could just dilute success and make it even harder for anyone to succeed at making games. The counter-argument would be that with a wider range of games, more niches are filled, so more people buy more games – for example, maybe I buy 6 games a year, but if there were twice as many that were twice as appealing to me maybe I would buy 9. If, as Naavik found, the relative proportions of success are about the same then these two effects perhaps cancel out. Or realistically, there’s lots of other effects and that data isn’t sensitive enough to detect them or their true net effect, and also not enough time has elapsed to feel the effects, so let’s just wait and see.

Still this does put more pressure on the question of ‘discovery’ – I can’t buy those amazing games if I don’t know they exist. Most people who work in video games seem to want discovery to get solved through some clever platform design, but I don’t think that’s very likely:

  • If I knew what I wanted and could find games that meet exactly that, just maybe we could solve it, but people do not know what they want! I would never have searched for something like Outer Wilds or Portal – in fact if I had searched for either of those accurately it would have effectively spoiled the game – but they were exactly what I wanted.
  • This type of problem is approximately solved in social media, where algorithms can serve content people are likely to enjoy (well enough that they keep scrolling, anyway) – but I don’t think people play enough games for you to run the kind of predictive models that work there
  • It might just about work if you took as input everything about you (social media consumption, books read, films watched etc on top of games played), but nobody holds all that data (yet)… and we feel uncomfortable with anyone holding all that data.

So in the absence of all that, we just get marketing.

Or some sort of procedural AI-generated game starts out as what you thought you wanted and evolves into what you actually needed based on your real-time reactions. That sounds a lot like a dream, or perhaps a nightmare, both figuratively and literally.

Perhaps more likely, in the future maybe your household android tracks all of that contextual data about you locally, and can then make perfect recommendations?

  • Transmission ends, stretching for a note of optimism

Categories
New

Things 2025 Q4: Podcasts, long titles, recursive knowledge

Podcasts

I’ve circled around podcasts without diving in for a long time, unsure of how to navigate the medium. With movies or video games, I’m already looking at places that tell me about new and interesting releases, and I know how to find out if something I’m considering might be good; with movies especially I’m aware of directors, writers or actors I like that will form part of the draw. With podcasts, I have none of that!

Despite that, I’ve found a few podcasts over the years that I liked, so it seemed a good time to round them up.

Blank Cheque

Blank Cheque is a podcast about films (apparently a very popular one but I’d never heard of it, I guess because of my distance from podcasts) specifically about the careers of directors who got ‘blank cheques’, i.e. enough power that they were able to make exactly the films they wanted. It takes the (apparently classic) form of two friends chatting about the subject matter, usually with a guest, but if there was ever a conversation I wanted to overhear about film, this is it. The friends are David Sims, a film critic who is often very funny, and Griffin Newman, an actor/comedian who is also incredibly insightful about films.

They came to my attention through their ‘stunt’ podcast that kicked the whole thing off: “The Phantom Podcast”, a long podcast series diving into Star Wars: Episode I – The Phantom Menace in forensic detail, under the conceit that this was the only Star Wars movie ever made and they are trying to make sense of it. They then ‘discover’ later Star Wars episodes one at a time and continue to analyse them as if nothing else existed. It’s quite incredibly niche and nerdy and there’s only one Things reader I would recommend it to, and I already have, so maybe don’t start with that. (I will say the full ~36 hours (!!) of material covering the prequel trilogy does go out on a pretty audacious high with a George Lucas impersonator defending the films against all criticism in front of a live audience).

I recommend picking a well-known director or film you like and searching the wiki to see if they have covered it (they have covered quite a lot of the big ones by now), then give that episode a listen. My favourite part is that they are often fantastic at drawing out ways in which the film’s meta-text resonates with the text. My least favourite part is that they will often talk over and interrupt their guests, especially women. That’s a pretty sharp down-side, so hopefully the fact I recommend them despite that shows how much I appreciate the good parts.

Link to Blank Cheque podcast

Designer Notes

Ok, I got into this one a while ago but didn’t mention it until now. A little like ‘Blank Cheque’, Designer Notes sees Soren Johnson interview game leads that have produced especially interesting games over their career. As an interviewer, he’s fantastic at drawing out the most interesting stories, with a knack for spotting when a fascinating nugget got skipped and asking the guest to dig into it.

Like Blank Cheque, if there’s a game series or game-maker you like, search (maybe like this) to see if they have been covered and start there.

Link to Designer Notes podcast

Gamecraft

Much shorter than any others mentioned here; if you have any interest in video games as a medium, a business, and in how it has adapted and grown over the past four decades, this is a fantastic listen. Mitch Lasky covers most of the material and is incredibly insightful since, through a lot of luck and some good judgement, he was somehow very close to the action at almost every major evolution the business has been through since the early 90’s. Blake Robins plays the ‘straight man’ role, standing in for the audience by asking the questions you’d naturally want to ask or pointing out context Mitch is sometimes too modest to state.

Link to Gamecraft podcast

Mindscape with Sean Carroll

Sean Carroll is a theoretical physicist with an interest in many different scientific fields, and in this podcast he interviews scientists of all kinds to get a better understanding of their area. He’s a great communicator, and by studying the published materials of his guests in advance he’s able to coax them into explaining it clearly even when they themselves are not quite as eloquent. (Okay, sometimes it doesn’t quite work and he ends up doing most of the explanatory lifting).

Here I recommend scrolling through his list of guests and topics and diving into whichever one you find interesting.

(Random bonus note: I like the way the little intro song has a melody that kind of matches Carroll’s intonation.)

Finally, as an aside, Sean Carroll appeared in Piers Morgan Uncensored to debate Eric Weinstein on the topic of theoretical physics. I’ve seen this characterised as “Jerry Springer for nerds”, notably including the following exchange (only missing a little context):

            Eric: If Sean had actually read my paper…

            Sean: I actually have read your paper! I have it right here!

            Eric: Well first of all, Sean, how dare you…

Someone like Eric is able to get the attention of ‘alternative media’ by sounding very smart and espousing a theory that ‘the establishment’ doesn’t agree with. You need to be incredibly knowledgeable and calm to reasonably debate with someone like this, and Sean pulls it off. I found it fascinating for that reason alone. I even give Piers a little bit of credit for attempting to end the discussion with some grace, given how little he understands the topic (by his own admission).

Link to Mindscape podcast

Useful memorisations

I’m sure a few of you have already done this, but here’s two simple things worth memorising.

First, the numeric position of each letter of the alphabet (a = 1, b = 2, c = 3 etc). In the summer of 2001 I once needed to walk a couple of miles alone and spontaneously decided to memorise this while I walked; I’m amazed how much this has paid off over the years. Admittedly I do move in circles where this sort of thing comes up more often than perhaps is normal.

Second, the ‘major system’ is a simple correspondence between the digits 0-9 and consonants (0 = s, 1 = t/d, 2 = n etc). Once you have memorised this, you can turn any number into a set of letters (e.g. 20251231 = N S N L T/D N M T/D), you then insert whichever vowels you need to make a word or words, which in turn are usually easier to memorise than the numbers because they are a bit weird. This does take a bit of skill, and I’m not very good at it but get by (e.g. NSNL.TN.MT = No SNaiL-eaTeN MuT). Great for PINs or a surprising number of other things. Admittedly sometimes the effort of translating the thing into words and then back again often means I end up memorising the numbers directly, but that’s still a win.

If you have similarly useful memorisations to recommend, let me know!

AI coding trough of disillusionment

In a fast-moving world, this article by Mike Judge, “Where’s the Shovelware? Why AI Coding Claims Don’t Add Up” is now a bit old (article date 3rd Sep 2025), but it was a pretty thorough look for signs of AI coding being transformative in terms of software / websites / apps / games and not finding any evidence:

  • The METR study found developers thought AI made them 20% faster but it actually made them 19% slower
  • Mike’s own mini-trial found nowhere near the powerful effect he had subjectively thought was happening
  • There is no evidence of increased releases of apps on iOS, Android or Steam
  • No increase in domain name registrations
  • No increase in new public Github repositories

Mike also then directly addresses the most obvious counterpoints that could be raised.

https://substack.com/inbox/post/172538377?=&aid=recKGiRdTrn0tC0OH

Now, to avoid oversimplifying, we should also consider the following:

  • Generating code is just one part of shipping a product, all the other parts can be quite difficult and annoying, so maybe that functions as a barrier
  • This doesn’t mean nobody sees any benefit from AI in any coding situation. People can be coding for personal use, or improving existing workflows
  • Even if there’s no effect on publishing rates, there could be other effects, such as different levels of quality or hiring patterns

There are also, of course, other ways in which generative AI has an impact – there’s dramatically more to this topic than I’m covering here. For example, it’s a great way to make social media spam, presentations everywhere are getting bespoke illustrations that would have otherwise been just text, and some have seen a decline in job roles for junior positions (for example), so there’s still lots to watch out for.

Films where they actually had to Do The Thing

With editing and special effects, movies can weave a lot of magic. However, sometimes the plot will require the film-makers to actually depict something incredibly impressive without any tricks.

I first encountered this as a theoretical example in the book (recommended to me by Will W), “Adventures in the screen trade” by William Goldman, screenwriter of All the President’s Men and The Princess Bride. In one chapter he writes a short story and then has experts in different aspects of film-making give him feedback on how the story might need to change to be realised on film. Of particular note was that the story hinged on a boy getting a transcendently incredible haircut of unknowable beauty; it is firmly pointed out to him that this is essentially impossible to portray on screen.

What I’m getting at here is skills specifically outside of those you expect for film-makers. An example of what I’m not counting: sometimes a character gives an inspiring speech to rouse people to action and it’s literally inspiring – difficult to write, but movies have professional writers. Or sometimes a film requires stunt work, especially a Jackie Chan film in which part of the appeal is he is performing difficult and dangerous stunts himself – but this is still a fairly normal part of film-making, with Jackie Chan being an apex example.

Where it can get harder is music.

Inside Llewyn Davis (2013) is a film about a folk singer on the edge of musical success and (more often) disaster. They found Oscar Isaac for the lead, who is not only perfect for it but also a legitimately talented musician. At one point he is brought in to record a novelty song, “Please Mr Kennedy” (with Justin Timberlake as the song’s on-screen and off-screen writer; also with Adam Driver doing a silly voice), and Isaac’s character opts to take the session fee rather than any cut of the song’s profits. As an audience you should read this as a mistake, as the song will probably perform well. As a novelty song, and to the point of this section, it is legitimately good enough that you believe that’s a possibility. (It’s partly based on the original Please Mr Kennedy by Mickey Woods, and as @mooviedude141 notes in the comments on the film’s version, a different Kennedy did pretty much go on to send two of the performers into outer space).

There are plenty of other films in which a song needs to do a lot of lifting and they pull it off (“Remember Me” in Coco (2017); the Cloud Atlas Sextet in Cloud Atlas (2012); a central performance in Sinners (2025)), all of which are impressive, but I know of two examples that go the full distance.

In O Brother, Where Art Thou? (2000), a surprising through-line is that the escaped convict protagonists end up recording “Man of Constant Sorrow”, and while their hijinks continue the song is becoming popular in the background, to pay off beautifully near the end. Meanwhile in real life, that song (and the film’s soundtrack) legitimately became extremely and surprisingly popular.

More recently Kpop Demon Hunters (2025) features several songs by the in-movie group HUNTR/X (performed mostly by genuine K-pop group Twice among other expert contributors), who are supposed to be extremely popular – and in real life, to quote from the Wikipedia summary, “it became the first film soundtrack on the Billboard Hot 100 to have four of its songs in the top ten, was certified double Platinum by the Recording Industry Association of America (RIAA) in October 2025, and received five Grammy Award nominations”.

What about non-music?

In The Ring (1998), a particular video will ‘curse’ anyone who watches it. Perhaps making such a video is arguably within the wheelhouse of a horror film director, but the execution of it really stands out. The video is semi-abstract but nonetheless legitimately creates a deeply cursed feeling, which is why I won’t embed it here and also is part of why that film haunted my nightmares for years, only resolving when I dressed as the main spook Sadako one Hallowe’en.

Very long titles of manga and anime that explain the whole premise

This trend has been running for some years in Japanese anime and manga. Some Examples:

  • The World’s Finest Assassin Gets Reincarnated In Another World as an Aristocrat
  • Banished from the Hero’s Party, I Decided to Live a Quiet Life in the Countryside
  • I Got a Cheat Skill in Another World and Became Unrivaled in the Real World, Too
  • The Magical Revolution of the Reincarnated Princess and the Genius Young Lady
  • The Guy She Was Interested in Wasn’t Really a Guy at All

I wonder if this becoming acceptable (or even appealing) makes some stories more marketable than they would have been in a short-title paradigm. Perhaps this is partly why the Isekai genre (protagonist goes to ‘another world’, a genre lately overrepresented in general and also in my above examples) has had such a boom in popularity.

We could consider how this might play for major Hollywood films. There have been a few examples with titles changing mid-marketing-push to get longer/shorter and more/less explicit about the premise:

  • “Birds of Prey (and the Fantabulous Emancipation of One Harley Quinn)” did not seem to work as a title and had a late pivot to “Birds of Prey”. I can’t help but feel just “Harley Quinn” would have worked better, but that’s a topic for another Things.
  • John Carter was previously to be titled John Carter of Mars, but seemingly Disney thought the ‘Mars’ bit would put people off, so hoped to just… sneak it in there?
  • Edge of Tomorrow sounds cool but tells you very little, so had a late pivot to the much clearer (but equally terse) Live. Die. Repeat.

Perhaps there’s something fundamentally different going on in the marketing strategy though – Hollywood films that aspire to be blockbusters will often signify their intended mass appeal with a very short title such as ‘Titanic’, ‘Frozen’, or ‘Sinners’. Meanwhile, a lot of anime and manga are trying to carve out a niche, and they can best do that by setting out the full premise up front so the people looking for that exact sort of thing will find it.

So, if there’s anywhere we might expect this trend to show up in Western media, perhaps it’s books?

Recursive Knowledge

Laurence once pointed out to me that “you have a vested interest in anyone that has a vested interest in someone you have a vested interest in”, and he remains correct.

In the second book of Cixin Liu’s Three Body Problem trilogy, the author dramatises the problem of insufficiently recursive trust. Two groups, who we can label Alice and Bob, have a problem: each of them has the capacity to annihilate the other before there is a chance to retaliate.

  • Alice trusts that Bob will not attack, and Bob trusts that Alice will not attack
  • But, Alice does not know that Bob trusts that will Alice not attack, and vice versa
  • If there’s a chance Bob does not trust Alice will not attack, probably he will attack
  • So to be safe they each should launch a pre-emptive strike

Unfortunately this keeps going even if you can add another level of trust. Consider just Alice’s side in the next level up (using close synonyms for trust/know to try to keep it manageable):

  • Alice trusts that Bob will not attack
  • Alice is also confident that Bob, in turn, trusts that Alice will not attack. Hooray!
  • But… Alice does not know if that meta-confidence is reciprocal. As far as she knows, even though Bob trusts that Alice will not attack, he might not know that she knows that
  • So Alice must consider the possibility that Bob does not know that Alice knows Bob trusts that Alice will not attack.
  • If that’s the case, Bob might be concerned that Alice is unaware that Bob trusts that Alice will not attack. Bob will consider the possibility that Alice does not believe Bob trusts Alice will not attack.
  • If Bob is considering that, it means he worries Alice thinks Bob does not trust that Alice will not attack – and if that was the case Alice, assuming Bob does not trust her, would launch a pre-emptive attack.
  • In this scenario that Alice is imagining, Bob would therefore launch a pre-emptive attack… so Alice had better launch her attack first.

You can see how this keeps going even if you layer on more meta levels of trust. The problem presumably caps out at our ability to compute much further. Unfortunately, if both sides entrust such decisions to AI then this will either become undecideable in the halting problem sense, or they will pre-emptively attack each other while trying to get to the bottom of the recursion just to be on the safe side.

How far can humans really think this sort of thing through? In the previous section I noted there is a manga titled “The Guy She Was Interested in Wasn’t Really a Guy at All”. As you would expect from the title, the drama hinges on the fact that Aya has a crush on Koga who she assumes is a boy, but Koga is in fact a girl in her class who dresses in a very male-presenting way outside of high school.

I don’t think it’s really a spoiler to say that eventually the secret comes out, because it’s very much about the journey rather than the destination (and at the time of writing the series is unfinished). That knowledge recursion is dramatized over a series of stages – Aya finds out about Koga, then Koga learns that Aya has found out but Aya doesn’t know that, and at a certain point Aya overhears that Koga has realised that Aya knows about Koga’s identity – but Koga does not know she knows that! That sounds pretty abstract, but in the moment of reading, the meaning and emotional significance of this highly meta revelation feels incredibly clear. I suppose as social animals, it makes sense that humans can grasp this stuff pretty well when it’s in the context of a story.

While I was thinking about all this I found out that Steven Pinker now has a book out called “When everyone knows that everyone knows”, which deals with ‘common knowledge’; when recursive knowledge operates at a population level.

I should note a slightly distant and dark irony there: we’re going through a kind of ‘common knowledge’ moment right now with the release of the Epstein files. In 2006 Pinker wrote a letter to his friend Alan Dershowitz, who was Epstein’s defence attorney at the time, on the the wording of the “internet luring statute”, and this apparently contributed to federal sex trafficking charges against Epstein being dropped. Pinker says he was unaware this was how it would be used and regrets sending it.

  • Transmission ends somewhat uncomfortably
Categories
New

Things 2025 Q3: Neuromancer, Expertise, Consciousness

Brian David Gilbert (BDG)

BDG makes the kind of weird videos I wish I had my act together enough to make. Here’s two examples that show his range.

A short song about hats:

A slightly spooky promotional video about get-rich-quick schemes:

I think we need more of this kind of thing, so please go watch all of his videos and even consider his Patreon.

(Side-note: BDG reminds me of Spike Jonze both as a person and for his creative output.)

Neuromancer and the Prism of Hindsight

I recently read William Gibson’s 1984 debut novel, foundational cyberpunk text ‘Neuromancer’.

It projects ahead to an unspecified time in which everything is online, and hackers enter some sort of cyberspace ‘matrix’ to conduct various shenanigans. It’s also very much about cybernetic enhancement, with some consideration of AI and some space business. It’s noirish and fast-paced but also dense and intense. There’s a lot going on.

My first thought was that being written in 1984, it seems astoundingly prescient about the future online world.

Then I read that Gibson didn’t really know much about computers or networks, he just liked the language of it. So my second thought was perhaps you need to be far enough removed from a thing, like he was, to see where it will lead.

Then I paid attention to the cover of the copy I had been lent (by Nick H), and realised something strange: it seemed to have some of the hallmarks of an AI-generated image.

Examples:

  • The core composition is a bit odd
  • The shape of the hairline is dramatically and weirdly asymmetrical
  • There’s a strange artefact on the hand that doesn’t seem specific or prominent enough to represent anything
  • The cityscape in the background has some repeating patterns that a human artist would probably try to avoid
  • Some of the domes in the cityscape seem unintentionally asymmetrical

Of course, this thing was published in 1984, and this art is by a human (Steve Crisp). I’m looking at something from the past through the prism of hindsight; in the context of a “futuristic” image, I’m primed to look for AI signifiers.

So comes my third thought: my thoughts on the book being astoundingly prescient also come through the prism of hindsight.
– I’m discounting everything that doesn’t really add up (the 3D visual interface isn’t realistic or sensible; there’s an eye-hacking thing (I think?) that doesn’t really make sense; the stuff in space seems very fanciful)
– I’m over-reading the things that were prescient (the everything-is-online aspect, the ability to leverage that fact to achieve powerful feats with ‘hacking’
– I’m under-reading the parts that really weren’t prescient, at least so far (the cyber-business and simulation aspects mostly).

This doesn’t really diminish the book – it’s a fascinating and impressive work, building out its own strange reality, and inspiring The Matrix (1999) even more directly than I had assumed. You just have to be a bit careful when judging prescience.

Very Short Animal Videos

Thanks to the Reddit algorithm for serving me these tasty and very short animal videos. They are optimised for portrait though and I use YouTube videos to embed things, so I’m not sure how well this will work:

“My cat will eat anything”:

Eating anything
byu/TheHenanigans inUnexpected

“Cat tries ice-cream for the first time”

he tried ice cream for the first time
byu/tuanusser inholdmycatnip

Sound needed for these:

Surprise:

Trying out a new de-corker when..
byu/fpotw infunny

The Paradox of Expertise

An exchange I saw recounted online and can no longer find went something like this:

A: Oh, do you consider yourself some sort of expert in vaccines then?

B: Well yes, I studied medicine and specialise in vaccines

A: Don’t you think that makes you biased?

[Edit – I found it! Added below. T.M. 02-07-2026]

Humans are prone to confirmation bias. We tend to give heavier weight to things that support what we already believe, and lighter weight – or none at all – to those that contradict it.

What I find even more insidious is a kind of second-degree confirmation bias: we discount someone’s remarks as being due to their confirmation bias… due to our own confirmation bias. For example, someone might doubt a particular bit of well-evidenced medicine, but when they hear a medical expert defend that thing, they assume the expert is only defending it due to the expert’s own confirmation bias.

Without getting deep into the concept of hierarchical trust networks, this is quite difficult to cleanly dissect, because confirmation bias is a real thing.

For example, you may recall that Researcher Bias exists*: a researcher who believes that an experiment will yield a certain outcome is more likely to end up getting that outcome, even if they are not intentionally manipulating the experiment to that end.

*But aren’t the studies looking into Researcher Bias suspect? As I wrote about in Things 133, a meta-analysis and even a meta-meta-analysis cannot satisfyingly answer this question.

You also see this in Planck’s principle: “A new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die and a new generation grows up that is familiar with it” – colloquially and bleakly paraphrased as: Science progresses one funeral at a time.

Or the Upton Sinclair quote:

It is difficult to get a man to understand something, when his salary depends on his not understanding it

If all this sounds a little vague and theoretical, I recently faced it head-on: a small stand of bamboo began to spread in my small garden. I know that some particular species of bamboo can spread very aggressively and do real damage. So what I need is an expert who can identify what kind of bamboo it is, and then I’ll know if I need to pay for some other expert to help get rid of it.

The trouble is those two experts are the same company. They will assess the bamboo for you, and then if they think it needs to be removed they will offer you the (quite expensive) service of removing it. The obvious question is: can I trust them to diagnose it correctly, if they know they can make money from one particular diagnosis? (My best guess for this was to at least consider the opinions of two different experts).

The same can be said of any product you buy in which the amount of it you should use is unclear. How many Aspirin should you take, how much sunscreen to put on how often, what collection of skincare products? The people who make these things should really know the answer, but they also make more money if they can convince you to use more than you need.

Infamously, Alka-Seltzer increased sales by normalising the use of two tablets instead of one through their advertising (and tagline, ‘plop plop, fizz fizz’, or ‘plink plink fizz’ in the UK). Still, the origin story (Snopes link) does at least suggest this did originate with a doctor suggesting two would work better than one.

This also runs the other way – a product could offer a legitimate advantage, but by default we don’t believe it when they tell us. I recall the story of a certain battery manufacturer having a significant research breakthrough making their batteries as much as 20% more efficient, an advantage they kept for a few years. Unfortunately from the consumer perspective, all batteries are claiming some kind of mysteriously special efficacy, so it’s hard to trust any one of them as being particularly meaningful. (I wish I could remember who this actually was!)

One possible answer here is Which?, who try to assess consumer product effectiveness with scientific tests. Of course, when they find a product doesn’t do what it should, the manufacturer will usually counter that they didn’t test it properly, and claim that they have a better understanding and more accurate test of their own product. Depending on the product in question I tend to find this more or less compelling.

So what, really, should we do about this?

In some cases, as I alluded to earlier, there are ‘trust networks’. I don’t need to trust a single vaccine expert on their effectiveness, because they are endorsed by thousands of disparate experts, and disparaged by a small number of non-experts (who can also have their own biases, if for example they are selling an alternative).

In other cases the direct incentive structure seems to run very strongly one way – it doesn’t seem to me that climate scientists finding evidence for climate change benefit from that conclusion anywhere near as much as climate-deniers trying to sell you an online course about their views benefit from people believing their denial.

For substances such as sun-screen and painkillers, the proper quantity to use tends to be endorsed by professional bodies, not just the people who sell them. In the case of painkillers you are of course free to experiment with a lower dose and judge the results for yourself.

When it comes to academic research, you can often look into the funding source. If a study casting doubt on climate change is funded by a big oil company, maybe it’s worth looking for other studies.

It feels like I’ve climbed all the way up a mountain of concern only to climb all the way back down again, so, er, maybe it’s fine???

Emel – the Man Who Sold the World

I enjoy David Bowie much more as an actor (Labyrinth, The Prestige) than as a musician, but this cover of ‘The Man Who Sold the World’ stopped me in my tracks. Emel’s delivery seems much more suitable for the slightly spooky lyrics than Bowie’s, and the extended glissando vocal at the end was so compelling I bought a Theremin (this one) in an ultimately misguided attempt to find a way to make a similar sound.

The Science of Consciousness

Here’s the ‘hard problem of consciousness’ as David Chalmers puts it:

It is widely agreed that experience arises from a physical basis, but we have no good explanation of why and how it so arises. Why should physical processing give rise to such a rich inner life at all? It seems objectively unreasonable that it should, and yet it does.

I had previously thought there was nothing interesting here. We only have our own experience of consciousness to go on, so it seems unjustified to consider it “objectively unreasonable”; this is just how it turns out and there’s nothing more to say. (Previously I wrote about Chalmers’ other formulation, the meta-hard problem of consciousness, although perhaps I misread his intent).

I read the book “Being You: a new science of consciousness” by Anil Seth, and I’m excited to have slightly changed my mind as a result!

An argument against there being anything about consciousness to dig into is the ‘philosophical zombie’: a creature that in every way resembles and reacts like a normal human but lacks consciousness. This is easy to imagine, and suggests there’s nothing you can “do science on” because there’s no way to distinguish the zombie from a human that does have consciousness.

Seth makes this counter-argument: “Can you imagine an A380 flying backwards?” In one sense, this is easy – just picture a large plane in the air moving backwards. But “the more you know about aerodynamics and aeronautical engineering, the less conceivable it becomes”. The plausibility of the argument is “inversely related to the amount of knowledge one has”.

One could imagine the same thing applies to consciousness – it does seem like if you deeply understood the way in which consciousness arises, “imagining” a philosophical zombie would be a lot harder. That does seem fair to me!

But still, how do you find a way in to this topic?

Seth’s answer is what he calls the ‘real problem of consciousness’: to explain, predict and control the phenomenological properties of conscious experience. Still difficult, but at least something specific to aim for.

His first way in is to consider how we might measure how conscious someone is – specifically the level of awareness rather than wakefulness. The diagram below shows how different states sit across these two axes.

So we’re looking for some kind of measurement that would show regular conscious wakefulness as having a similar level to lucid dreaming, for example.

He talks about some interesting research showing that a measure of the complexity of electrical signals in the brain seems to correlate well with what we think of as consciousness. Even better, there are tests that can distinguish someone with ‘locked-in syndrome’ (conscious and aware but unable to move any part of the body) from someone in a ‘vegetative state’.

A simpler precedent to the complexity model is this: simply imagining playing tennis produces a detectably different pattern of brain activity to imagining navigating a house. These two kinds of thoughts can therefore be mapped to ‘yes’ and ‘no’, enabling someone with locked-in syndrome to communicate. This dramatically debunks my thought that there was nothing useful to look into here!

Unfortunately, the rest of the book gets quite a bit heavier and less compelling.

First there is Giulio Tononi’s “Integrated Information Theory” (IIT) of consciousness. Put tersely it posits that consciousness is integrated information – kind of a huge claim as it arguably means even atoms are perhaps a ‘little bit’ conscious. It suggests a very specific measure of consciousness: Φ (Phi), essentially how much an information system is more than the sum of its parts.

This theory doesn’t seem to go very far just yet. Seth’s summary of where it is at:

…some predictions of IIT may be testable […] there are alternative interpretations of IIT, more closely aligned with the real problem than the hard problem, which are driving the development of new measures of conscious level that are both theoretically principled and practically applicable.

So it seems we just have to wait to hear a bit more about that.

Next is the Karl Friston’s “Free Energy Principle”. In this, the term ‘free energy’ can be thought of as a quantity that approximates sensory entropy. The clearest summary Seth makes is this:

Following the FEP, we can now say that organisms maintain themselves in the low-entropy states that ensure their continued existence by actively minimising this measurable quantity called free energy. But what is free energy from the perspective of the organism? It turns out, after some mathematical juggling, that free energy is basically the same thing as sensory prediction error. When an organism is minimising sensory prediction error, as in schemes like predictive processing and active inference, it is also minimising this theoretically more profound quantity of free energy.

This is not really a theory of consciousness but, Seth considers, something that will help explain consciousness eventually. I get the impression Seth understands this enough to see how it might be of value, but not well enough to explain it so others can see that – at least not me.

Finally Seth considers the possibilities of animal and machine consciousness, and largely concludes it’s very hard to say anything about these, which is a bit disappointing but is also quite fair.

To summarise, I thought there was nothing useful to say or do about consciousness, but after reading ‘Being You’ I now think that’s wrong; it seems like there is something to dig into here, but so far our theories are only just scratching the surface of it.

(If you want a more detailed recounting of the book with added commentary, not all of which I agree with, you can read this long review by ‘Alexander’ on LessWrong)

  • Transmission ends
Categories
New

Things 2025 Q2: Constraints, Polarised light, Kuleshov, Infohazards

Artists vs. Constraints

The medium of any given art form creates restraints or encourages certain properties in the art itself.

For example, movies and theatrical productions are somewhat constrained in duration by the capacity of the human bladder. Paintings tend to be painted at certain scales that are easier to perceive, to distribute, and to display. Writers paid for serialised content (such as Charles Dickens or people writing for a long-running TV series) can often see better financial returns if they can find a way to spin the story out for longer.

Here are three examples of this I find particularly pointed.

1 – Tintin

The Tintin comics were originally serialised in a newspaper supplement one page at a time. To entice people to purchase the next newspaper, it helps if there’s a cliffhanger of some sort at the end of each page. I found it very hard to read a collected Tintin comic in full once I spotted this pattern, because it turns out just about every single page contrives a cliffhanger, sometimes in a very silly way.

Two classic end-of-page cliffhangers
Two of the milder end-of-page cliffhangers
Example of a decoy end-of-page cliffhanger

2 – Lubalin internet drama

Lubalin is a musician who (among other things) sets internet drama to music – that is, he takes slightly deranged exchanges (typos and all), and arranges them to music. Execution is everything, so check out part 1 as an example:

You can check out the rest in the series here.

The constraint here is that the current dominant algorithms really like 1-minute videos (specifically TikTok and Youtube shorts). In a ‘making-of’ video you can see the edited highlights of Lubalin constructing one of these songs. The notable moment comes at 8m 42s (link to that timestamp) when he plays what he has developed in full and is horrified to find it is quite a bit longer than a minute! So he has to adapt.

Noooo!! It’s like… double the length!

You get to see the constraint impact the art directly – and I think you can see how it slightly helps, but also slightly hinders. (You can skip to the final 1m song here)

3 – Calvin and Hobbes

Finally, Bill Watterson’s comic ‘Calvin and Hobbes’ was syndicated to newspapers between 1985 and 1995. The constraints on the Sunday supplement colour format were particularly harsh: panels 1 and 2 had to be optional (as they are sometimes dropped for space), and there had to be panel breaks in certain places so the comic could be reconfigured as necessary to fit a full, third or quarter-page format.

Panel breaks must fall where specified (but you can have more). Panels 1 and 2 must be optional. Sheesh!

Eventually the strip grew popular enough that Watterson was able to mandate a single full-page format – which is still a constraint! – with some excellent results. Check the 25 examples here to see the range possible (note, just the ones in colour are the ones where these specific rules apply or were eventually avoided).

What do we conclude from this?

I like the two opposite reactions one can have:
– Take something wildly inappropriate for the constraints of the medium and try to cram it in anyway.
– When you can, question those constraints and see what you can achieve if you break some of them.

Perceiving Polarised Light with Haidinger’s brushes

This is incredible to me: it’s actually possible to perceive polarised light with your very own human eyes! Find an area of pure white on a polarised LCD screen (very likely whatever you are reading this on), then tilt your head from side to side. Faint bow-tie shaped areas of yellow (and apparently blue, but not for me?) will briefly be dimly visible as you tilt your head, caused by polarised light. Read more here:

https://theconversation.com/polarised-light-and-the-super-sense-you-didnt-know-you-had-44032

An advert where a bear directs a film

This is one of those adverts where the creative agency has so much fun with the execution that the purpose of the ad seems a bit of an afterthought. I love it!

This thing is 13 years old now and I keep going back to it every few years so I’m giving it official Thing status:

Bonus bear content via Clare:

Wildheart Animal Sanctuary on the Isle of Wight recently rescued two bears that had grown up knowing only cages and concrete – article here. They raised the funds for a purpose-built large natural enclosure. The bears arrived recently and have been exploring their new home. You can try to spot the bears via one of the live-streamed cameras, or watch them explore their new habitat in recent videos such as this one.

The Kuleshov Effect (via Josie)

Terse description from Wikipedia:

“The Kuleshov effect is a film editing (montage) effect demonstrated by Russian film-maker Lev Kuleshov in the 1910s and 1920s. It is a mental phenomenon by which viewers derive more meaning from the interaction of two sequential shots than from a single shot in isolation.”

When moving pictures first became possible, it was not obvious that a ‘cut’ would be something a viewer would accept, and early movies were often presented with very few cuts following the established art form of the play.

It turns out several dramatic things happen in the human mind with a cut:
1) An instant change in perspective is … just completely acceptable!
2) Much like Grice’s maxim of relation in language, we assume a cut within a scene has meaning – for example, if we see a character notice something off-screen, then cut to a thing, we assume they are looking at that thing. The mild version of the Kuleshov effect.
3) Building on top of the above point, the full surprise of the Kuleshov effect is that we may even overinterpret the images either side of an edit to have them make more sense!

You get essentially the same effect in comics or any sequence of panels, for example just putting these two images together implies they are related, and our mind invents a story:

You can also see the same effect distilled in GIF form:

You also get a version of this when giving a slide presentation – you can accompany some text or spoken word with an illustration, and the mind of the viewer will automatically interpret them together. This is why a presentation can be improved even with some barely relevant images!

Beyond that, I think there are even weirder effects somehow going on that relate to how we process movies:
4) Non-linear editing, in which edits go forward and back in time (such as Memento, The Prestige, Speed Racer) are also fully comprehensible, if handled carefully
5) Non-diegetic music (music that is not happening in the scene, e.g. the sound of a John Williams score in a space battle) weirdly doesn’t seem weird

One of my favourite examples of this: from pure sound design and Kuleshov effect, Rian Johnson conveyed Rey and Kylo communicating with one another across space in The Last Jedi with nothing more than an edit.

Another corollary is that you can splice new footage into old, and if you take advantage of how we interpret edits, you can recontextualise the old footage to make it seem as if, for example, Star Wars characters are drinking Cristal Beer.

Anti-basilisks

Before we begin, some necessary context: an “infohazard” is information that could harm people if known. There are many tweaks, shades and nuances to this; you can find ‘fun’ examples over at the SCP Foundation (their wiki page on the topic; search for SCP entries tagged as infohazard here), or this very enticing trailhead I have not followed at Lesswrong: “a typology of infohazards”.

Below I’m going to describe something that some consider to be an infohazard, or just dangerously adjacent to one. I will go on to explain why actually I don’t think it is (although the adjacency remains possible). This is the last Thing in the post, so if that freaks you out you can just stop now!

I will give you some space to do so.

–

–

–

–

–

–

I’m referring to Roko’s basilisk, which is nicely summarised on Wikipedia. It’s a theoretical superintelligent AI that we could build which would “punish anyone who knew of its potential existence but did not directly contribute to its advancement or development, in order to incentivize said advancement”.

So, you see the problem. It’s clearly a very stupid thing to build, but now you know about it, so you can infer that other people know about it, which means that some of them might eventually have the means and motivation to build it, which in turn maybe means you should help them do that to avoid this ‘punishment’ – probably at minimum by telling more people about it!

The first version I encountered contained the threat that even if this basilisk does not come about until long after I’m dead, nonetheless it would review history and then simulate the minds of the non-contributors in order to punish them. At first I felt no concern about what may happen to some theoretical simulated me, but I suppose the real threat is that the ‘me’ right now might actually be the simulation it is running rather than the original me, and as such I am not safe!

So, maybe if you squint hard you should worry about this a bit, but my counter is that if such a thing is possible, many other similar things are possible that could easily cancel this out.

The anti-Roko’s-basilisk

A superintelligent AI that destroys any instance of Roko’s basilisk it can find, and will reward (or retroactively simulate rewarding) anyone who helps bring it about.

The double anti-Roko’s-basilisk

It’s like the anti-Roko’s-basilisk but it also specifically rewards anyone that would be punished by a Roko’s basilisk twice over (perhaps by simulating two instances of them having a great time), and anyone who helps build it three times over. So you should definitely build this one, and even if you don’t, maybe other people will, and it more than cancels out the antics of Roko’s basilisk.

At this point I think it’s pretty clear we’re just inventing imaginary monsters having imaginary fights, making the whole thing seem quite childish and inconsequential, and certainly not motivating enough to start worrying about making or not making any such things.

Anyway, just to keep things sufficiently spooky, let’s talk about adjacency.

LessWrong was founded by Eliezer Yudkowsky, and when LessWrong user Roko posted the original formulation of the basilisk, Yudkowsky gave an uncharacteristically blunt response and banned the topic for 5 years (as recounted in the Wikipedia article). It did look like a bit of an overreaction, but as that article goes on to recount, he later explained (or post-rationalised) his reaction, most notably with this:

“The problem was that Roko’s post seemed near in idea-space to a large class of potential hazards, all of which, regardless of their plausibility, had the property that they presented no potential benefit to anyone.”

I find that idea quite compelling. It’s very hard to know how large this hazardous part of idea-space is, but it seems like it could be non-zero. I can imagine using the basilisk concept as a springboard into conceiving the most hazardous ideas possible, but at this time I am choosing not to do that, because there really is no potential benefit to anyone.

What about the opposite… the infobeneficial? Well, hopefully that’s everything else you find in Things, at least to some extent.

  • Transmission ends