The Mouse Dystopia

Here's a rather nice little video on Universe 25, a.k.a. the Mouse Utopia :

This is something I've come across many times over the years. I'm sure I've read at least one long read on it, though I'm not going to trawl through everything to find that again.

Anyway, the basic story is simple. A bunch of mice were given unlimited resources, perfect conditions, and the result was an absolute disaster. Not only was the result a population collapse to total extinction, but the mice demonstrated weird pathological behaviours : some became constantly aggressive, others retreated from mouse society altogether, some became obsessed with personal grooming. Mothers ate their own young and reproduction completely ceased.

The popular reading is that comfort kills. We need something to strive against, the thinking goes, something difficult to prevent us sliding into sloth and indulgence.

A more sophisticated interpretation, as the video presents, came from the original investigator. What went wrong was not excessive comfort, says this alternative interpretation, but a lack of clear social roles. The mouse society collapsed because they had no meaningful activities to structure their little lives around; in effect they failed from sheer ennui.

Both of these are tempting readings from the brief popular description I presented. It's certainly hard to see people taking literally deadly selfies and not see more than a little resemblance to the grooming-obsessed mice who beautified themselves to death. This is part of some deep appeal in the human psyche to a state of decline. Like the myth of Atlantis, there's something perversely romantic about the notion of a society which reaches the heights of glory only to fail and never return.

And yet, it's all bollocks. All of it.

After some prolonged discussion with ChatGPT, I've come to the conclusion that what's really going on is likely far, far simpler, and any resemblance to modern Western society is entirely coincidental. What U25 tells us about excessive comfort and the need for meaningful social roles is, I think, exactly nothing. This doesn't mean that these things don't matter. Quite likely, they do. But it does seem that U25 can't tell us a bloody thing about their effects either on humans or, importantly mice. The experiment appears to be yet another case of a popular finding having serious flaws and being massively over-interpreted.

Major caveat : ChatGPT provided many links to support its assertions, but I haven't checked them. I would expect a small percentage of these claims to be wrong, so by all means, do go through the full linked discussion and check them for yourself. However, unless it's made a really serious mistake, it seems very unlikely that the overall claims here are anywhere near as flawed as what really went on in the hell-hole that was Universe 25.

With that in mind, it looks like this a utopia only in an incredibly narrow and extremely stupid sense. Yes, it provided the mice with unlimited resources, and yes, in principle it could have supported a larger population than the maximum value it reached before the collapse. But two absolutely crucial factors are missing from the popular narrative that fatally undermines any meaningful attempt to map the mice to men.

First, there were no good controls for density. Earlier and later experiments did hold the population at lower levels by removing mice, but only when the density had reached a level which was already known to be above comfort levels : these too saw similar behaviours in the mice to U25. At no point does there appear to have been any attempt to either hold the population stable at a very low level, or even at a moderate level but during the phase while the numbers were still increasing (i.e. before the self-destructive behaviours began to happen).

This blows the idea of "excessive comfort" right out of the water. The mice were given little stimuli beside other mice and food. No attempt was made to provide enough space for the mice to live without having to constantly deal with other mice : sure, they always had enough to eat and drink, but they had little way to meaningfully escape from other mice when conflict arose. The idea that this is "excessive comfort" is absurd : it would be more like being locked in a restaurant with thousands of other people and expecting a flourishing society because you could have as much to eat as you wanted. Hate your neighbour ? Who cares, another steak will make you feel better !

This is not quite to say that sheer, simple, mouse density was the dystopian driver. It might well be that with a suitable rearrangement of the furniture, more compartments and better division of the space, U25 could have indeed sustained its nominal maximum mouse population. But it does look highly credible that density is the root cause : for any given configuration, if you have so many mice that they can't escape each other, eventually you'll get continuous unbearable interactions. Maybe a better configuration could, somehow, have allowed a sort of stable mouse version of Biosphere 2*, where they'd naturally regulate their own breeding, but that's a very different and more sophisticated social engineering proposition.

* I'm aware that no breeding occurred in Biosphere 2. Bisophere 2 meets Love Island meets Big Brother, then.

The more obvious difficulty is that here the mice were imprisoned with no way to leave of their own accord. They had no choice but to endure a hell of their own breeding they simply couldn't foresee. The limitless food becomes irrelevant if psychologically there were just too many mice.

Interestingly, there were experiments done in which problematic mice were removed, but unfortunately, only limited data is available on what happened both to them and their parent colonies. It does seem that there were changes in the transplanted mice, but how much and over what timescales is unclear. On the colonies, less data is available : they didn't recover, but by this point the mice had been living in hell for a prolonged period. It seems entirely plausible that they were now permanently psychologically damaged.

And possibly, physiologically damaged as well. This is the second crucial factor that goes unreported, one that's hiding in plain sight. For the Wikipedia article contains something so obvious that overlooking it renders the whole experiment null and void: the cages were only cleaned every six to eight weeks.

Two months ! Two months of two thousand stinking mice. That's not a utopia at all. That's a cess pit. 

And this was already known in 1973. I can see how things like control experiments for density might get overlooked, but this is surely such an egregious failure of reporting that journalists deserve a good chunk of the blame here. Even accounting for a possible belief that mice are essentially just little furry mechanical machines with no psychological needs (such mistaken ideas being more common the past), not recognising the need for basic sanitation is a crazy thing to leave out of the story. 

"Ah yes, this experiment shall be absolutely perfect for these mice. They're going to have a fine time of it, breeding their little hearts out, having all the food they could want, no predators, no way to escape their neighbours after a fight and only having to worry about cleaning every two months."

This is palpably a silly experiment, but the way it's been reported is demonstrably awful. It means nothing for the mice and even less for humans.

And these two factors work in tandem. Had the mouse density been kept lower for a sustained period, the effect of the cleaning frequency could have been examined more carefully. Perhaps at a lower density, mice would naturally avoid dirty areas, which means the situation could be extremely non-linear. It's possible that the population of a lower density could be sustained for a longer time or even indefinitely at the same cleaning frequency, but at a higher density, disease might become prevalent all the more rapidly. Without that density control, we don't know if there wasn't some innate flaw in the rest of the experimental setup, something else – another unknown unknown – that was causing madness in the mice. Given that mice, usually, are hardly the most difficult animal to maintain in captivity, that possibility doesn't look all that unlikely. Something had clearly gone very, very wrong in Universe 25.

Unless something ever repeats the original setup exactly, and does the appropriate control experiments, we'll never know for sure what the problem was. Fortunately for mouse-kind, this seems very unlikely to ever occur because the thing looks strikingly unethical and the possibility of mapping the result to humans is essentially impossible, making it incredibly hard to justify to any reasonable ethics committee. "Yes, I'd like to trap two thousand mice in unsanitary conditions for two months in order to check whether people get miserable if they spend too long in decadent luxury"... this would probably be grounds for ending tenure with the reasonable justification that the investigator has gone insane.

I've suggested that the driving factor behind this dystopia is density, but I don't mean that there's a simple linear relation between density and mouse comfort levels. Nor can we eliminate other possibilities : maybe the mice would still have exhibited pathological behaviours even at low levels due to something else. Clearly, though, pet mice don't usually go insane like this, so I rather doubt it – unless there was something truly strange here, like a chemical leak or something.

All we can really conclude from the Mouse Asylum is that if you give mice unlimited resources, take little account of their psychology, only clean them every couple of months, and make no attempts to control their population, you get a catastrophic hell-hole leading to extinction. ChatGPT tried to call this a "valid observation", but to my mind this has all the same validity as observing that a rock is grey or that a table is made of wood. Nothing unexpected occurred here. It offers no profound psychological insight besides "miserable mice all died". The additional tragedy is the mythmaking that grew up around this complete and total non-result.

The Bothersome Bias Of The BBC

Accusing the BBC of bias is a classic knee-jerk response that generally indicates the commenter is either a) a moron or b) politically extreme. At least, those saying that the BBC is hugely right-wing because it does or doesn't cover story X, Y, Z, are basically equalled by those saying that the BBC is hugely left-wing because it does or doesn't cover story X, Y, Z. It's not at all unusual to see people both claiming that the BBC is a horde of raging anti-Semites and that it's enabling Palestinian genocide. 

Spoiler alert : it isn't. 

I don't believe that the BBC has this sort of straightforward bias on any political position. I don't think it's even, as is sometimes more credibly claimed, especially pro-establishment, given the amount of airtime it's given to Nigel Fuckface over the years.

And I hope don't need to spell out that any attempt to be politically neutral will never be 100% perfect because that's literally impossible; detecting any slightly lean one way or another on any individual position demonstrates exactly nothing and is just a waste of energy. I'm not going on a more general rant here. If you're fervently convinced the BBC does have some massive political preference, then this isn't the post for you.

Instead, I want to point out three other specific areas where I think the BBC is genuinely biased. The first is one I've not seen much discussed elsewhere, the other two are mentioned more frequently but I hope to give an angle on these that's at least presented less often.

This post is strictly anecdotal and I make no claim to empirical verification whatever. I'm only posting this because the effects seem so strong that while the magnitude is certainly open to debate, I think there's very little chance indeed that the effects themselves aren't real.

The first is a victim bias. That is, on BBC Breakfast at least, it seems to me that victims of crimes are put front and centre. Their responses are overwhelmingly the ones which are aired first and given far longer coverage than any other involved parties (except politicians) or any discussion of wider context. Little is said about the cause of crime, and the morning session, deliberately touchy-feely though it is, becomes barely more than a victim campaign group that tends to want to punish the criminals and nothing else – even without ever saying so explicitly.

Two examples : with the murders committed by Valdo Calocane (and similarly for the Southport murders), the prospect that the criminal is mentally ill is mentioned, but the coverage given to the victims and their desired response not to treat him as such is far, far greater. The result is that the idea that the man needs medical intervention more than punishment is never seriously considered. Second, the current obsession is with the possible release of two men who murdered a police officer. Again, discussions of the context of the early release scheme aren't zero, but they're very much on the back foot compared to hearing from what the victims think.

Of course, I'm not saying that the victims don't deserve a voice. But it seem ludicrous to me to essentially only give the victims a voice and not that of independent experts. For the criminally insane, we deserve to hear from mental health professionals. For the prospect of early release, we need to hear from prison officers. And yes, technically we do, but only a very little bit. Any sort of complexity is drowned out by emotion of the people who are understandably distraught.

It isn't the job of journalists to stand for victims. Campaign groups are certainly entitled to do this, but independent journalism should present the different sides more equally. There are other factors involved besides the victim's emotions and the viewer needs to know what these are. They can't make an informed judgement if the only thing they hear clearly is how upset people are.

I'm not claiming that this is by any means unique to the BBC or even that the BBC does this more strongly than other media outlets; it's also entirely possible that BBC Breakfast is worse than other BBC programs for this by its very nature. It may well be that if you watch the news later in the day, or specialist programs, you get a much more multi-faceted view of an issue. But for serious and violent crimes, BBC Breakfast is little better than a tub-thumping call to "lock 'em all up and throw away the key". The problem is that if this is your main news-watching time, as it is for me, then it doesn't matter if other programs have more sophisticated commentary : you'll come away having heard from those who are most closely involved with the case at the huge expense of more objective professional evaluators.

The second issue is the more well-known "both sidesism", expressed brilliantly in this classic webcomic. I won't go into this at any length, but I just want to mention one especially frustrating aspect of this : absolutely everything is presented as awful. By interrogating all people as though they were equally (in)competent, treating all positions as though they were equally respectable – or more to the point, disrespecting everything equally – nothing is ever allowed to be seen as good. Everything must have some downsides. Everything must cause problems for someone. Everyone must be cynical. It feeds this "grumpy old man" habit compulsive addiction to expect everything to be awful and therefore to interpret everything as awful. The result is perpetual misery.

Look, there's enough which is self-evidently horrible in the world as it is. "Manufactured misery" is simply not a thing that needs to exist. All news outlets do this, but my point is that even if you're relentlessly centrist/neutral, the result is the same. At least if you're avowedly pro-some position or other, you're allowed to enjoy it when your side gets a win. The centrist BBC just appears to hate everything.

The third, final issue, is topic fixation. Topic priorities are a more obvious and related issue, but refusing to let go of a story long after what as initially breaking news has well and truly broken, seems to me to be less discussed. This tendency comes and goes and at the moment it's on a bit of a low, though they are absolutely obsessed with this whole "early release" thing as though there was nothing else worth talking about. Even when nothing new of any substance has occurred, still they must repeat the same story day after day.

The reason I say this is currently on a low is because this was much, much worse when they were obsessed with medial problems and charities. They mentioned Rob Burrow and his motor neurone disease to the point where it started getting downright creepy and weird, and the same for Deborah James and bowel cancer : these are both important topics, but the focus on these individuals began to feel like the behaviour of a stalker, not someone genuinely concerned. All other conditions and individuals were excluded because these were apparently Special People, and nobody else experiencing any suffering seemed to count.

There comes a point where raising awareness is just no longer possible because awareness levels are now saturated : BBC Breakfast did this to the nth degree, continuing to report as though their deaths were a recent event months and even years later. This is the Matthew Effect of cumulative advantage of the most perverse sort. Shall we seek out other causes and activists and give them a boost by giving them TV time ? Nah, let's stick with the ones we already know about. Nobody else is important enough.

Look, I love the BBC, and I get dismayed when people call it biased in the usual ways; it is, hands down, one of the best news sources in the world and one of Britain's greatest institutions. It isn't perfect, but that's not a realistic standard for criticism. But not expecting perfection does not eliminate the need for constructive suggestions. These three areas, at least, are surely areas where things can be improved without needing massive systemic overhaul. The underlying structure is sound, but like everything else, it needs a good sweep and a lick of paint from time to time.

The User Experience Of Our Digital Life

Today, I want to look at how and why UX seems to be so routinely awful in our daily digital lives, and what this has in common with broken toasters and racism. These ideas have been floating around in my head for quite a while. The need to finally write them down was triggered when I commented on this thread, but here I want to give a more developed version.

What is UX and how is it so bad ?

I couple of years ago I went to an astronomy software conference and I was delighted that there was a big focus on user experience. For those unfamiliar, UX is closely related to, but not the same as nor exclusive with, user interface. It involves much more about how things behave, whether things do what the user expects, how many steps they need to execute a command, how easily they can figure out what to do, that sort of thing.

This pleased me greatly because almost all astronomy codes are, or at least were, written as command-line based and come with minimal documentation. This is fine when it’s just for one’s own use, but releasing such a thing to the world and saying, “here, read my Python script” is surely ridiculous if you ever expect anyone to actually use it. And yet this is absolutely standard practise in the world of astrophysics.

I, for one, don't want to read code. I want to read things I can read, not have to parse.

So for bad UX, astronomers are guilty as charged. What’s particularly annoying is that there persists a (somewhat subconscious) undercurrent of belief that things are somehow better as command line tools and GUIs are actually something to avoid, or even that code itself should be difficult to use and uncommented because you should never use something you don't fully understand and must have suffered in order to do so.

Yes, some of my compatriots are actually mad.

But then the more I’ve thought about it, the more I’ve noticed this sort of thing everywhere. Not to the same degree, but the same root issue – a fundamentally sub-optimal UX – appears in so many places where it really, really shouldn't.

To be fair to astronomers, this is somewhat justifiable from a group who are focused on the results and think that UX is a “soft” feature... but multi-billion dollar companies whose product is aimed at the general public surely ought to know better. The following list is by no means complete, but it should give a fair indication of the state of things. I've deliberately curated this to be (mostly) things which are quickly obvious when using an app or service, and when a straightforward profit/engagement motive isn't readily apparent. The reasons for those kinds of failings are self-evident; the reasons for the following frustrations are maybe a bit more interesting.

• Amazon recommendations for the thing you’ve just bought, as though it wasn't obvious you wouldn't immediately need another expensive sofa or five dishwashers
• The NatWest banking app repeatedly insisting I can't use biometric information to log in even though I demonstrably can 
• Transport apps (planes, trains etc.) not syncing with local board information or even given conflicting information about delays and statuses
• Streaming services that bury the “continue watching” list, continuously change the category orders, and make it incredibly easy to accidentally skip halfway through an episode instead of pausing it 
• News websites which are so plastered with ads that the content is literally unreadable. Ads for profit make complete sense, but so many that you actually can't access what you came for do not
• The total lack of explanation given on Google Pay as to why I can't verify my card, only to find that one day it's spontaneously verified itself
• The UK passport replacement service continuously jumping the page around when I want to enter anything in the entry fields on my phone
• The relentless, unending stream of cookie acceptance/refusal windows
• Constant demands for notifications and subscriptions (yes, I get the point of these, but did nobody implementing these ever think, hmm, maybe this will annoy me so it will actually be counterproductive ?)
• The utterly awful "summarise this email" feature in Gmail which nobody asked for, to say nothing of Google's ludicrous attempt to replace actual search results with AI hallucinations
• The incredibly irritating micro-delay when trying to switch workspaces on Windows (yes, Linux fans, I'll let you have this one : Linux does this much better)
• The abyss that is infinite scrolling everywhere with no ability to save your place
• GitHub having the same labels for different things in different places that do different things
• Actually, just simply... GitHub, all of it
• Similarly, Linux (workspace switching notwithstanding)
• Expedia's constant notifications of miniscule prices changes while booking (although this is not as bad as it used to be)
• The lack of a CTRL+F search function in the ChatGPT app, no pagination, minimal ability to organise with folders and no ability to export them
• Inconsistency in language settings with some McAfee (among others) pop-ups still being in Czech despite the fact that I've told the app not to do this

The list could go on. And on. This kind of thing is everywhere.

Now this is not a first-world-problems rant, for the above list has also been chosen to only have things which are annoying, but don't render things truly unworkable. Unfortunately, this makes the problems all the more intractable. Serious problems usually get fixed quickly, but less important ones can persist indefinitely even when they're trivial to fix. But why ?

It's all the fault of those frackin' toasters

This is an an example of what I call broken toaster syndrome : you won’t fix your toaster if it’s only slightly broken, like if the handle gets a bit wonky. Most people, most of the time, accept problems when they can work around them. It's only when they hit that threshold of being genuinely unusable, when the handle comes fully off, that they actually do anything about them. The result is an accumulation of issues that never needed to happen, things that make simple tasks maddeningly frustrating for no good reason.

In the digital world, hardly anyone is going to file a complaint because a button wasn’t where they expected it to be or the icon size is bit too large. They’ll just put up with it. And dedicated UX testing can easily overlook things if it’s given to the original developers who already know how to use the tool, or it’s so limited that nobody actually gets the experience of what the thing is like to use routinely and in anger (when users may be stressed and in a hurry), as opposed to a single dedicated session (where they're likely to be a lot more patient).

So, just as with problematic hardware, digital stuff gets asymptotically degraded to the realm of being bloody awful to use whilst being technically functional. Genuinely brilliant underlying code gets an interface which isn't user friendly but is actively user hostile. 

Which brings me back to astronomy. Simplifying slightly, I once asked why the FLASH hydrodynamics code doesn’t come with some built-in tool for setting the initial conditions. The answer being… this was too easy ! They’d rather focus their attention on developing new features than making the existing ones easy to use (which would simplify this for new users by many orders of magnitude), on the grounds that this could be done at any time… the result, of course, is that it never gets done at all. Paradoxically, the easiest changes can be the ones least likely to be implemented.

Being pro or anti-AI is like being a racist

It's been suggested that all this is the result of AI, because blaming everything on AI is a way to win instant popularity on the internet. But I don't think this is much the case, or at least, that the claim needs heavy qualification. Similarly, a couple of recent articles about the declining power of Google search are worth reading in full : the current, apparently terminal decline is indeed dominated by Google's balls-deep obsession with implementing AI as badly as possible, while previous occasions when the quality declined – and it certainly did both wax and wane over the years – were due to other factors.

Make no mistake : AI is definitely a contributing and sometimes a dominating reason behind some UX decay. But I think the underlying cause is a bit like the recent prudish age-gate checks on internet service providers : no-one wants to complain their porn is blocked because they're embarrassed, and likewise, nobody wants to complain about any minor problems they encounter so they just get worse and worse. AI simply isn't necessary as an explanation here.

Likewise, I've come to the conclusion that being pro-AI is a bit like being a nationalistic bigot and being anti-AI is like being a racist : they are both absurd generalisations, both ends of the same horseshoe*. To say AI use will definitely make everything better is like being a Reform voter who constantly chants about how the British flag is amazing and lacks any knowledge of the history of how that flag was regarded worldwide. But equally, to say AI will only make things worse is like another Reform voter (or perhaps the same one) insisting that everyone who isn't British is just an awful person who should be excluded from society at all costs.

* At least, that is, as far as concerns go about how AI is used – I'm not going into energy use or legal concerns about data acquisition here. Those are legitimate issues, but separate.

The point is that you can use it in different ways. If you have the mindset that UX is useless, you'll just use AI to add more and more features in a chaotic UI that eventually makes no sense even to experts. That's not the AI's fault, that's yours for being a moron.

If, on the other hand, you think that UX is valuable and take it seriously, you can use AI to help you tremendously here. The speed of AI-implemented code changes means that you can afford to spend at least a bit of time doing your own testing. You can now run through your UI and check everything works as expected and get a feel for actually using it, rather than coding once and insisting users follow the documentation (if you even write any) : you can adapt to the users rather than forcing them to adapt to your code. AI, in some conditions, will finally allow you to fix that damn toaster.

Hence, racism and bigotry. It's madness to say that all AI is unanimously bad just as it is to say that all British people are cunts, and conversely, it's equally crazy to claim that it's all brilliant and that all British people are magnificent specimens of humanity. "It's bad because it did this one thing I don't like, and all the good stuff doesn't count" is just classic rotten-cherry-picking racism.

Okay, but if AI can fix a broken digital toaster, does that mean it will ?

I'm optimistic that this will help astronomy quite a lot : since almost all code is open source here, it becomes possible to tailor the experience to suit one's own needs. This still requires effort – you have to really think about how you want things to work, you absolutely cannot simply tell the AI to "make things better" – but at least now it crosses the threshold for possibility. Astronomers, at least, can afford to spend a few hours here and there asking AI to make UX improvements; they could not previously afford to spend weeks learning other's people's code in order to move one button to a different panel. After all, you'd likely fix your toaster yourself if a) you knew how and b) it wouldn't take long.

I saw a meme asking if people using AI could actually be persuaded to stop, directly comparing it to meth addiction. In my case the answer is yes, I could and would still do my job without it. But this misses the point that I'd have to work in a world full of broken toasters with no real possibility to fix them. AI at least offers a path out of this in way that nothing else has ever come to close to offering.

Elsewhere, though, the effect is harder to gauge. Giant corporations have always had the resources for UX refinement, they just often choose not to bother. There's no reason to think that making this easier will actually lead to them doing anything differently; quite likely, some of them will prioritise making more toasters over fixing the ones they already have.

How we go about changing such attitudes I don't know. Heck, I may well be unusually irritated by digital defects; maybe everyone else is happy enough with the way things are. Certainly there are more pressing issues... but I can't help wondering if maybe the continuous low-grade defects in digital UX aren't having some cumulative effect. Could it be that having to deal with all this is part of why so many people seem so angry despite ostensibly living in an age of material wealth and technological wonders ? And wouldn't it be nicer if things just worked ?

Review : The Odyssey

I seem to have a habit of reviewing Christopher Nolan movies so I can hardly not cover The Odyssey. I'll keep this one spoiler-free.

In some ways this is essentially Project Hail Mary but in space. That is, it's definitely excellent – outstanding, even – but I don't think it's going to "change cinema forever" or anything as hyperbolic as many reviews make out. Or perhaps, Dune. It's certainly a cinematic feast and an absolute spectacle, and possibly the best version we'll ever get. Perfect ? No. Thoroughly enjoyable ? Absolutely. An easy 9/10 from me.

The Dune comparison also holds up in its sheer physicality. Things are intense, and thankfully Nolan seems now to be well out of his "let's make everything deliberately inaudible" phase. In fact, the screams of the Cyclops might well be the best audio effect I've ever heard. I'm convinced that some of the scenes, especially the storm(s) at sea, are actually louder in the cinema than they would be in reality, but the intensity is used very deliberately and selectively. I don't think I ever had any difficulties hearing anything.

The visual spectacle is equally compelling. Again, Nolan appears to have gotten over his "CGI is worse than Hitler" pretentious attitude and uses it sensibly and judiciously, where practical effects just wouldn't work. It looks, sounds, and feels stunning from start to finish. The creepy design of the Cyclops and the monstrous nature of Scylla are particularly well done : the Cyclops looks far more like a creature than an overgrown one-eyed man, while Scylla looks almost like an alien animal, something other-worldly. These are not ordinary mortal beings ramped up to 11. Nolan may have minimised the gods, but he doesn't hold back on supernatural : similarly, Circe's transformation of the crew into pigs is downright nasty. This is a scene that's more visceral and earthy than I would have expected from someone I associate with perfect cinematic cleanliness. It works brilliantly.

About the only other thing I would mention visually is the oddly contrasting aesthetics. Odysseus' palace looks interestingly archaic and pagan inside, dim and gloomy and lit by dark flame, but with a staircase that looks like something you'd more associate with something found in the home of an ultra-minimalist serial killer. Weirder yet is the palace's positively ruinous exterior. This isn't unique either : Athena's temple also has an interesting primitive vibe (as does Troy), but with a rather strange look as though nobody could afford to hire a cleaner or do any routine maintenance. Circe's island reminded me more of Tintagel than the Mediterranean, which was a little jarring*. And Agamemnon looks a bit like a scary robot while Odysseus' crew wear simpler but anachronistic armour – even when they're just rowing a boat.

* It was actually filmed in Scotland so I wasn't too far off.

It does work overall though. The palace looks like a palace, the temple looks like a temple, Agamemnon is clearly the leader even without saying a word, and Odysseus' crew are clearly a band of warriors. It's simple, iconic, and immediately clear to the viewer what's going on.

The storytelling is where it shines. Nolan's use of flashbacks to fill in the essential details of the Trojan War is carefully controlled, follows the overall structure of the poem, and fills in relevant similarities and necessary parts of the narrative exactly when needed. The moral aspects come through clearly without being overwhelming*. Nolan's choice of which parts to emphasise and which to discard feels like the most balanced approach possible.

* Though, I would say that its needs to reduce the mentions of both "Zeus' Law" and the "Sea Peoples" by at least a factor of five. Honestly if you make a drinking game out of either of these, you'll be dead by the end of the film.

Nolan does make some deviations from the poem, but I think he does so in a self-consistent way that doesn't undermine the source material. Whereas the original Menelaus and Helen live happily ever after, here they're at odds. Menelaus is a dominant masculine archetype, but Helen isn't afraid to push back. At the same time, Menelaus clearly has massive respect for Odysseus' cleverness that lends him some interesting depth. He's a brute, but not a villain... though he does look disconcertingly like Andrew Tate.

The Penelope/Telemachus dynamic is handled similarly well. Penelope is clearly torn by competing forces : wishing her husband had never left, desperate for his return both for selfless love and a purely pragmatic desire for security, wanting to do the best for her son but not yet ready for him to assume his father's throne Telemachus would have been easy to make into an annoying Wesley Crusher-type git, but to his great credit, Tom Holland manages to avoid this. 

As for Odysseus himself, I think Matt Damon has become a fine actor. He takes a highly understated approach and I would actually have liked a bit more direct emotion from time to time : he could do with getting properly angry and (especially) distraught at least once or twice. There's a particularly poignant scene in the poem where Odysseus breaks down and weeps at memories of the war that I think would definitely have been a useful addition. At the same time, his massive PTSD comes through in other ways, and I think Matt Damon largely succeeds in giving the character latent tension rather than merely not bothering to act in a more thespian fashion.

Practically every other objection I have falls firmly into the "would be nice to have but not necessary" category as well. I like the "Nobody" joke with the Cyclops, the shade of Achilles lamenting his fate in the afterlife, Athena going in disguise as a man... none of these really need to be in the movie though. Excising these, in my opinion, allows Nolan to keep focus on the tale he wants to tell, without omitting or changing any of the key themes of the original.

Finally, I have to mention classicist and translator Emily Wilson's notorious and scathing criticism. It's a worth a read, because it brilliantly highlights exactly those key themes in a way that's better than I could have done. It's also a pile of garbage. Among many, many other things, Wilson is oddly fixated on :

• The colour of the sea not being purple
• The lack of time spent watching Odysseus building a boat
• Christopher Nolan having some sort of drowning fetish (seriously, which Nolan films has she been watching ??)
• Odysseus not being horny enough
• The movie somehow being a veiled criticism of Brexit and the War on Terror
• Odysseus spending too much time worrying about not telling Sinon he was going to die
• Lack of exposition / too much exposition / unclear motivations (even when the characters directly state why they're doing things, which in my view helps provide explanations Homer never did but should have done)
• The supposed lack of internal moral consistency, despite the fact that this is clearly bollocks.

This last particularly irked me because it's even more nonsensical than the rest of the post. Wilson picks up on many moral points that would probably have taken me a lot longer to draw out, but she gets them all catastrophically wrong. Being laced with militant feminism – she seems to hate all the female characters because.... well I don't really know why at all – doesn't help, but the way she seems to get everything completely backwards is damned odd.

The law of hospitality is only the most egregious example. This is a recurring theme in both poem and movie, with respecting hospitality to strangers being literally sacred. Admittedly, in the movie, the reason give is rather blunt and transactional : that guests might be gods in disguise. This is morally uninteresting but understandably simplified for the sake of a movie, but what's handled really well is the use and abuse of the law. This kind of law only works when enacted respectfully, when both sides understand the limits to which it should be used. Its wanton misuse by the suitors leaves Odysseus with no choice but to disregard it, leading to calamity and even the ruin of civilisation. 

Like royal power, sometimes the whole point of giving authority is to deny its use : either entirely, or only to limit how far it's actually applied or restrict who gets its. Holding back, showing restraint, sometimes leads to more power rather than less. And with the law based on the authority of the gods, when it fails – when it's revealed as ineffectual – it can't be mended. The only alternative is an entirely new order.

My final points connect to this in a somewhat convoluted manner. Weirdly, Wilson laments that the movie is soaked in spectacle and there's always some loud exciting thing going on. This is a blatantly false claim, as there's plenty of brooding silences and quieter scenes that lend all the more impact to the noisy bits. Complaining there's too much happening here is surely a case of complaining there are too many notes. But even worse, this deeply flawed criticism misses the point that Odysseus goes through all of this, and so the viewer experiences all this intensity, as part of a journey of self-discovery. He has to overcome the external challenges, to be pushed to the very limits of endurance, before he can confront his own doubts and his own world view.

Yes, obviously that's all a horrible cliché, but this is literally the first such tale in Western literature, so it doesn't count.

The movie is explicit, and I think rightly so, that Odysseus' voyage away from Circe on a tiny raft into the turbulent sea is quite different to his earlier heroics, a leap of faith that tests him as never before. We don't need to hear some internal monologue to get this. We, the audience, can be rightly expected to fill in these kinds of gaps for ourselves; likewise, while I think some egregious violence in the sack of Troy would have helped explain Odysseus' inner torment, it isn't necessary. Wilson, on the other hand, seems to want the exact inverse of what makes the movie a success : she wants more exposition where none is actually needed and less in the (few) places where it's too much.

Anyway, for me it's a massive hit. It has definite re-watching value, as befits one of the world's oldest stories. There will be surely be more Odyssey adaptations in the future, but this one is likely to set the benchmark for decades to come.