When you don't know how to approach making something — start with the minimum.

The idea of expanding my skills through my own little games has been lodged in my head for a long time. A few years ago I tried starting with Love2D; now I've realised the best option for me is three.js. The path from prototype to "release" is faster, you can build interesting things with it, it's available in the browser and it generally copes with what I want.

I have two notional worlds that I work on at a leisurely pace when there's time: from setting and concepts through to lore stories.

The first world is an alternative mid-19th century — islands, secret societies. The second is the near future — a polar expedition, bases in the north. Both are dear to me and I pick at both now and then.

For the second one I decided it would be fun to build a minimal game in three.js, to learn how to interact with the engine, understand the toolset and the limitations, and at the same time test at least statically everything I want to expand later.

And so "Signal" was born. First as a frequency-search prototype, then it grew a concept, and then I spent a couple of weekend evenings modelling everything and making it work the way I wanted.

The premise is as simple as possible: a hyper-casual where you press a couple of buttons and turn a dial to catch the right frequency.

Under the hood, though, is everything I'll need later: buttons with states, self-illuminating materials with textures, sounds on presses, looping ambience, floating particles, a lore page triggered by clicking an object, a HUD overlay, geometry with textures, light sources and even some post-FX.

Naturally it wasn't done without Claude's help, but by the fifth version of the page I understood pretty well what needed doing and was using it more as a reference. Truly, understanding and finishing it yourself is far easier and faster than correcting its output ten times over.

Having grasped the principles and tested things reasonably well, there isn't much left that I still need and haven't touched: animation, physics and FX.

In the context of modern neural nets I keep observing the "tough Siberian bloke" effect from the old joke.

Ha, look, they don't know left from right on the road; ha, look, they get confused about temporal details; ha, you can obviously tell it's a neural net!

Yes, you can. For now.

I've deliberately attached what art nets produced at the end of 2022 — exactly three years ago — next to fresh examples.

They didn't understand context; they couldn't even hold a line. They made a kind of mince across the whole frame, and even then it was a wow, because — well, it's cool, software generating a picture. Mixing colours, grasping some fun bits like wide angle.

I wrote this post on Facebook back then, because a strong wave of indignation had begun.

Three years on, we have what we have. Fresh models understand context in different languages, work carefully with references, and produce results you're not ashamed to show a client — in FMCG, for instance. They follow instructions attentively and are precise in the details. Some of it is already built into software, Photoshop for one. There's an API and thousands of sites for the truly slow, if you don't want to bother with settings. Certain tasks — restoring focus to a blurred image, say — work better and better. If you don't know how to compose a prompt, you can ask an LLM for help. Will Smith eating spaghetti does so quite convincingly and without artefacts, i.e. video too.

Over the past year of my work in video production there hasn't been a single job that didn't use neural nets in some way. Sometimes they handle almost everything. The speed at which all this happened is terrifying and impressive in equal measure.

Yes, you can still tell — if you know where to look.

And what happens in another three years?

If the question of "how" gets less relevant by the day, how do you get back your understanding of what you want?

And do you even want anything at all in this new world where everything is possible in two clicks?

In Bluey, the closing chord of three seasons of a fairly interesting and varied life is the thirty-minute special, "The Sign." The whole third season builds towards it gradually; it's merely the trigger for everything that came before. In the small details the show is still episodic and it doesn't matter where you jump in, but if you know the context the picture comes together far better and it hits harder.

I don't know how many of you haven't seen it, so: spoiler alert.

Broadly it's an episode about selling a house and about a father wanting a better life for his children. They decide to sell the house in Brisbane and move, because he's been offered a better job somewhere. Obviously decisions like that don't happen without turmoil and difficulty; the children don't entirely grasp what's going on, and the younger daughter only understands what "selling the house" means in the final minutes, when they carry her table away.

An Australian, American — and British — feature: when a house is for sale, a sign goes up outside with the agent's contacts.

At the last moment the sale falls through by accident and, since this is a series about a happy world, the father decides to hell with the whole idea and pulls the sign out of the ground.

It seems to me this is a fairly widespread desire in our generation: to have your own home. Not in the sense of square metres or any specifics, but rather a general feeling. I've rented a lot of flats, lived in different countries, seen more of the world than the average person did a hundred years ago — but the heart demands a den of its own, and that's where the question arises.

Within the framework of immigrant misanthropy I take my bearings from Dovlatov, except that I drink considerably less. Of all the countries I've lived in there's only one where I felt like one of the locals — Australia, built by immigrants, obviously. And if there it's a pleasant little hammock of understanding, albeit at a distance, then in the UK you land on a thousand-year-old stone and take a kidney punch on the way down.

I'm aware that I'll never be "one of them" here, even with a passport — first and foremost in my own head. This place will never become "my" home, for all its brilliance, and there really is a great deal of brilliance here. Like a party you sort of invited yourself to, where you sip your drink quietly off to the side, expecting to be thrown out.

I haven't really found a way out. Another move in the next few years, before the new passport, is simply unthinkable. After that, the roots will be too deep in terms of the primary priority — my daughter's education. My forecasting horizon always slides down to months rather than years, and planning anything is simply laughable.

Our generation was handed an amazing time. A pity that biology takes its toll with the years, and the urge to be grounded and to find a home somewhere starts gnawing at the brain harder and harder.

.It's a corridor connecting two parts of a life. A path strewn with ash, a narrow passage in which you can't straighten up, though at first you can't see the ceiling — black with soot.

You don't know how long it will go on, you only guess: some considerable time yet. Sometimes the walls close in enough that you leave parts of yourself on them; sometimes you have to crawl on all fours.

It has two ends and you can always crawl back. Apart from ash in your hair and a general batteredness, nothing will give away the failure.

There's no light in the ashen passage besides your own fire, and there's not much to look at anyway, but you burn down slowly regardless, leaving weightless dust behind you.

One day it will end, and there, on the other side, everything will be there. Maybe not the way you imagine it, but definitely better than inside.

I've finally formulated why Arcane resonated with me so much.

It's a story about how one person's stubbornness and the creative thought "what if?" led to a huge volume of high-quality, beautiful work, which took at least one company — the animation studio Fortiche — to a new level.

It's gratifying when the people at a company love their product. You can see it from a kilometre away; that kind of love can't be faked.

And conversely, even if the product looks modern and fashionable and uses the latest engine features, if its creators didn't love what they were making you get counterfeit Chinese Christmas decorations. They glitter like the real thing but bring no joy at all.

The main problem with putting a piece of yourself into what you make is that it's hard to find understanding among your environment, your client, your colleagues — and in that case everything falls apart. The best outcome only happens when the whole team works from similar premises and everyone enjoys the work.

I can create that arrangement within a small department. Many of the people who've worked under me know that I worry about the team's emotional state considerably more than about KPIs.

But setting a good temperature inside the department isn't enough — your client has to be on the same wavelength, whether that's a CEO, an investor or just a customer. And that's usually where it all starts to come apart: either the CEO doesn't care about the result because he has deadlines, or the client wants to make his terribly important edits at the end that cross out most of the work, or investors don't want to fund extra months because they treat your work as a business asset, goods on a shelf.

Which is why Arcane is so valuable. Yes, it expanded the LoL universe; yes, it served as marketing for a questionable MOBA. But if the team's attitude to it had been that of a long advert or a way to squeeze more money out of the game's fans, nothing would have come of it.

Fortiche's client was Christian, who didn't merely care about the project — he lived it for all those years. Riot's leadership approved a budget sufficient for two seasons back at the start of work on the first (the figure of $250 million gets thrown around). Nobody was rushing them, the company's future didn't depend on this work, nobody expected Arcane to bring in wild profits, and feeding your children and paying your rent weren't on the agenda either.

Imagine you get a brief to develop a concept for the "9th of May, our grandfathers fought" theme.

A new look, so to speak — "a symbol of bright memory, without an emphasis on grief," without the smell of gunpowder; a sort of mono no aware, but on demand and for an exhausting occasion.

I don't have to imagine 🙂

Overriding militarism and heroics with something new is, to put it mildly, an ambitious task. The country has been saying the same words and showing the same images for more than twenty years. The ribbon itself is twenty years old. Red flags, stars, c-a-r-n-a-t-i-o-n-s.

And what do you do with that?

First, you have to settle on new keys:

• blue sky, not graves and blood
• a careful link to "the archive footage"
• memory, light sadness

Remembrance Day works roughly in that key, but it has a strong symbol — the poppy — which doesn't suit us, and there's the allusion to a wound.

Right: colour palette settled, stars removed from the foreground, thinking about the overall vibe — peaceful sky, sun, rye and wheat and all that, wildflowers where the fighting was, life goes on. We need a dried flower, a symbol of memory, something that hits you like it hits Harry Du Bois, oof!

And such a flower exists.

It's the forget-me-not.

Plant everything with forget-me-nots: blue fields merging with the sky, crystal-glass stars in the centre of the flowers (five petals, fits nicely), careful transitions through pseudo-archival footage of dried forget-me-nots from soldiers' triangular letters.

• Memory as the core message: the flower's name literally means "do not forget" — a direct reference to remembering the war generation.
• Colour symbolism: a soft blue peaceful sky, the complete opposite of blood-death-attention!
• Fragility and resilience at once: a small but stubborn flower that grows in all sorts of conditions — a metaphor for the human spirit, the "simplicity" of a wildflower.
• Plus the personal history: during the war soldiers often gave these flowers as a symbol of fidelity and a promise to remember.

…
…
…

None of which, of course, will ever happen.

I'm fond of the old LucasArts adventure games, and two of them hold a special place in my heart: Day of the Tentacle and Full Throttle.

A few years ago they got remasters from a studio founded by the creator of the originals, Tim Schafer — good remasters; you can switch to the original with one key, for instance. They also threw in bonus material, art included.

I ripped it out of the game files, tidied it up a bit (I'm more used to looking at PDFs) and I'm sharing it here. Besides everything else, the two games share a lead artist: Peter Chan. He also worked on Monkey Island 2 and Grim Fandango. In '95 he went freelance, into the film industry, and helped LucasArts from outside the payroll.

The old school of game developers had no clear role boundaries; you could call Chan a concept artist, a background lead and, in places, an art director. And LucasArts at the start of the nineties was a small, experimental company by today's standards.

Few modern games can boast that much traditional drawing during development: actual pencil, watercolour, markers. Digital gradually squeezed it all out in the name of speed, but I'm delighted every time I see the old techniques.

We had two seasons of Malyshariki, 75 episodes of Peppa Pig, five Disney features, half a flash drive of Tom and Jerry and a whole multitude of children's songs of every sort and colour, plus Blue Tractor, Pocoyo, Super Simple, five hours of Cocomelon and Three Cats. Not that this was a necessary supply for the trip. But once you start collecting the stuff, it's hard to stop. The only thing that worried me was Cocomelon. There is nothing more helpless, irresponsible and depraved than Cocomelon children. I knew that sooner or later we'd move on to that filth too.

For the past two months we've been solidly on Bluey, so it's worth telling you about it as well. I've worn everyone out on Twitter, I've talked everyone's ears off privately, so now to the next stage — Telegram channels.

Let me start from a distance. Does the name Peppa mean anything to you?

Modern "children's" animation leaves scars that heal badly. YouTube is stuffed with two-, three-, five-hour compilations — a ceaseless bacchanal of low-grade 3D, revolting music, wretched vocals and forced smiles. Plot swerves, canned baby laughter. It's hard being the parent of a two-year-old.

At some point, having gone completely insane from CoCoMeLoN, we decided to move on. A friend who understands the pain of parents of small children recommended, among other things, a series called Bluey.

I love animation, but I've never encountered anything so healing for the soul. And that's setting aside the fact that it's a children's c-a-r-t-o-o-n. It's completely obvious that these talking dogs gave me back a small piece of faith in tomorrow and in the idea that the world isn't entirely cursed and lost. A soul-healing watch that suits everyone — even if you have no children outside, there's always a child inside.

Three seasons of 52 episodes, each around seven minutes; the third is wrapping up now. The show is broadly episodic, so you can jump in anywhere, but towards the end of season three the episodes start connecting and it helps to know what's going on.

Animation is by the Australian studio LUDO and it's at a very respectable level. Everything on the visual side is good: setting, design, characters, world-building, animation, colour, VFX — art direction 10 out of 10. The art director, incidentally, is a young guy, Costa Kassab, a Griffith graduate.

The soundtrack has been on repeat in our car for two months now. Yes, the music is also 10 out of 10; it's by Joff Bush, an Australian composer — worth checking a couple of albums on Spotify or YouTube.

And the writing? You guessed it: 10 out of 10. Storytelling is by Joe Brumm, an Australian animator and writer. He graduated from Griffith in '98, went off to London where he worked on Peppa Pig, then came back to Brisbane. Broadly, the main family in Bluey is drawn from his own — he was also the middle of three siblings, and he's also a father of two daughters. The pilot came out in 2016; Brumm got lucky with his producers, so the first season aired in 2018.

As you know, I have a soft spot for anything Australian — I lived and worked there, I love Sydney, I can talk about the kangaroo country for hours. But this one got to more than just me: even my extremely sceptical wife started seeing Australia slightly differently.

In other words, if you have trouble with "celebrating life" content and want a little faith in humanity back, I strongly recommend it. Just don't watch it dubbed — you'll kill all the joy of the voice acting.

(and:) Worth adding separately that some episodes and moments produce full-on ugly crying. Absolutely.

"Save the bees! Save the trees! Save the whales! Save the… snails!"

The Australians, as ever. They observed that the population of small marine animals in Sydney Harbour had thinned out because the shoreline had been encased in stone, scratched their heads, and made little houses for barnacles, molluscs and assorted underwater creatures.

The project is called Living Seawalls, and for several years now it's been actively helping restore the marine ecosystem in the city.

It's a collaboration with Reef Design Lab (extremely cool people, have a look at the site) and the Sydney Institute of Marine Science.

They designed the houses starting from "who likes living where," 3D-printed them, made moulds, cast them in concrete and bolted them to the walls. Over two years, a 30% increase in life — in places more than on ordinary rock.

A find worthy of its own post.

I was looking for references and stumbled on a site with scans of IKEA catalogues — from 1950 onwards, all of them.

https://ikeamuseum.com/en/explore/ikea-catalogue

If you ever wanted to get your 2007 back and go misty-eyed over CD racks, or take a look at IKEA computer desks and IKEA TV units — welcome!

Separately worth going soft over the photography in the 90s catalogues; I'd forgotten how it was "done" back then, and it's a very recognisable aesthetic.

The most astonishing thing is that two years on, the hate speech about neural nets hasn't quietened down. Even though it's still hatred of a hammer. A tool is a tool.

I couldn't put videos like this together before — no time, no money; thoughts went into the drawer, nothing got released, just solid work. And that's despite the fact that I can do all of it single-handed: draw the concepts, model the assets, build the scenes, light them. It's just that doing it myself would take cosmic effort and time.

And now — I can, whoosh. Because I know what I want, I know how to swing a hammer, and I can assemble the whole thing in a couple of days.

And I feel good.

In that sense, artists mostly worry that neural nets are taking their clients — although, broadly speaking, every artist is their own most important client.

And how many possibilities open up from knowing how to swing a neural hammer — not for someone else, for yourself — is hard to grasp.

Never mind that for now this is all more for conveying a vibe, a thought, an atmosphere, rather than a precise guide or a finished work.

A lot of people "discovered" him via Antonov, who never hid the fact that he was quoting Woods. Sometimes boldly, sometimes lightly seasoned: HL2, Dishonored.

On the first level there's the visual: bright, memorable deconstruction, broken form, deliberately fractured rhythm. On the second there's the real Woods, far darker. He was a theorist and revelled in his idea of an "architecture of war" — theory can raise far heavier questions than practice is capable of digesting and building.

One of his books, for instance, is about the principle of "radical reconstruction," discussing the problem of Sarajevo. There was already an enormous amount of destruction there by 1996, and the contemporary rhyme suggests itself.

"New kinds of space, woven into the existing Cartesian constructive system, create a dialectic between the timeless and the temporal, a network of the unknown that inspires dialogue and innovation."

Inspiring, and true.

Woods wrote until his death; his last blog post is August 2012. He died in October.

In an ideal world I sit and work, nobody touches me, and then I die. Happy. The end.

When the question is what to work on, there are plenty of answers, but "defining a world" would top the list. Nothing defines a world as well as concept art. Which is why I work as a concept artist — it's the closest thing to what I actually want to do.

The main trouble is that I play the long game badly. I can't draw one picture for a week; it gets boring. A month-long project sends me into the cosy silt of depression. Grinding on something for six months? Ufff…

So if I want to give a full definition of some world, the year-long marathon gets broken into 365 sprints.

That's the concept of 360°+5: complete coverage of a subject in a year.

Time to talk about Laika.

First, an excellent post-credits scene (no spoilers):
https://www.youtube.com/watch?v=3h7p0NckTKc

So, there's a company in the US specialising in puppet animation — LAIKA Studios (named, yes, after the Soviet Laika). Its founder, Travis Knight, worked for a long time as a puppet animator and is an incredibly devoted man; I strongly recommend a short two-minute "interview" with him:

Laika started as a production studio, doing animation for Tim Burton's Corpse Bride in 2005, and since 2009 have released five films of their own: Coraline (2009), ParaNorman (2012), The Boxtrolls (2014), Kubo and the Two Strings (2016) and, in 2019, Missing Link. I've seen them all; any of them can be recommended as family or children's viewing.

To my great regret their last film bombed spectacularly, recouping only a quarter of its budget — completely undeservedly; it's a lovely story, a kind and funny children's comedy. Do watch it 🙂

But as a former 3D person and someone who likes to get into production, what always interested me is "how?" And that's where hell begins.

To start with, a shift of twenty animators can produce roughly one to two minutes of film per week. That's in an ideal world with no problems. Films are 24 frames per second; even with animation on twos it's still 10–12 different frames with slightly changing expressions on the puppet's face.

Naturally they use replacement animation: a relatively old technique where one puppet has many pre-made "mouths," or entire heads, as with Jack in The Nightmare Before Christmas.

At Laika they went further and combined 3D with puppet animation. Characters are partly animated in a 3D package, after which the "faces" go to 3D printing. They've used the technology since their first film, Coraline, but a lot of water has passed under the bridge since then and the tech has moved on.

Initially printing was single-colour, and all the tens of thousands of puppet faces had to be handled by the paint department, hand-painting five freckles on Coraline's left and right cheeks. Then printing became colour, but powder-based: the faces were fragile and post-processing still took a lot of time. In 2015 the team started working with a new technology — colour voxel printing on a resin printer — but only three colours were possible, which severely limited the characters' palette. And then in 2019, on Missing Link, they used six-colour resin printing with software developed specially for them, which allowed the models' UV mapping and textures to be used for automatic colouring.

Plenty is still touched up on the computer, of course, but that's what's good about Laika's approach — they aren't afraid to combine. Amusingly, the studio shoots on a Canon 5D.

Anyway, I'll drop some videos here for anyone who wants to get into the subject, and images in the next message:

https://www.youtube.com/watch?v=zHyTYL1Z1aM — a bit of behind-the-scenes

https://www.youtube.com/watch?v=1NGeGcV9dXw — a fairly detailed 45-minute talk on using 3D printing in production

https://www.youtube.com/watch?v=NKsDg5-BeIA — the process of building one character's puppet, shot in stop motion

https://www.youtube.com/c/LAIKAStudios/videos — I can confidently recommend the videos on their channel generally; they'll tell you how Coraline's little jumpers were hand-knitted and how Mr. Link, an enormous avocado-shaped Yeti, can move his belly via a lever in his back to create the effect of the puppet "breathing."

A cup of coffee, a warm tartan blanket and a run of the journal Zodchiy from 1872 all the way to 1924, obviously.

The journal is a treasure trove of references and interesting images.

It covers the 1876 World's Fair in Philadelphia and the 1878 Paris one, and shows every design variant for the Church of the Saviour on Spilled Blood — the ones that lost the competition. Lots of churches, old Russian ornament, and besides that, tonnes of text, formulas and so on — reading of little interest to anyone today, though the section on bricklaying was entertaining.

There are plans and elevations; some structures are broken down in detail — trusses, fixings, all with drawings.

Naturally, life demanded the mundane too: you regularly come across ads for "Portland cement" and "metal works."

On references: churches, barracks, stables, hospitals (and the manufacture of their ventilation and heating), automated laundries, ordinary townhouses, heated swimming pools, prisons, and dachas, dachas, dachas.

Suddenly, in the 1877 issue, you find a description with drawings of special cast-iron plates with glass inserts providing passive light to basement floors. With the sole caveat that the plates are described as being in "the newest buildings of New York."

A great deal of what's described has long been lost, sadly. The magnificent Lutheran Church of St Mary in St Petersburg, for instance, was broken up for firewood during the siege of Leningrad. Fragments of its foundation were only found in 2014; it had stood for about seventy years.

Or take the dacha of the beer magnate Ludwig Krohn, designed by Benois, near Peterhof — it doesn't even partly resemble the original design now. The beautiful dacha was first violated in the post-war 50s when it was converted into a "sanatorium," and finished off by a fire in 1995.

Or you leaf through 1899 and find an advertorial for the Russian Electrical Society "Union," with a founding capital of 6,000,000 roubles (insane money, if I understand correctly). The piece says the society is prepared to take on the design and execution of projects and estimates for electrical installations of every kind, for lighting and for supplying towns with electrical power.

The journal did not, alas, survive Soviet rule.

I was never a fan of the BioShock series, but the game wasn't really a sequel or a prequel — it was a completely separate work by game designer Ken Levine.

As Schreier describes in Press Reset — a book almost entirely devoted to toxic management and the ugly side of game development — the situation on the last BioShock was catastrophic. Levine constantly changed the concept, there was no real release plan, and development kept rolling back from production to the very beginning, to concepts and pre-production.

Worth mentioning that this tortured five-year ordeal has its own dedicated Wikipedia page (!), and we all owe a little something to Rod Fergusson, the man who closes out doomed projects, hired by Irrational Games into a top position seemingly for one purpose only: to ship the damn game.

I was 27; within a week I knew it inside out. One of the most beautiful FPS games I've ever come across. In terms of its effect on forming my taste, Infinite sits on the same shelf as Dishonored 2 — an Olympic standard, and there are maybe ten games up there in total.

A couple of years later I came across the game's art book. At the time I assumed a game of such incredible beauty and art must have a matching book, but alas — it was more of a graveyard of half-decomposed ideas and gaudy sketches.

At that moment Infinite taught me a new lesson: it showed me the back of the tapestry, exposed the seams and took off the rose-tinted glasses of naivety. In places the art was atmospheric, but for the most part it was ugly, alarming and infantile. I looked at those pages and couldn't believe that this daubing had, first, anything to do with developing a AAA game, and second, had somehow been reborn into the thing I loved so tenderly. And separately, a thought I couldn't quiet: this was the art selected for an art book.

It was a scrappy but maximally honest layer of development. It served a specific purpose and was created in a foul atmosphere of madness that dragged on for years. That's exactly what I'd show every aspiring concept artist, modeller, level designer — everyone who's so eager to get into games right now. Art books like this show best what you'll actually be doing 90% of the time.

Forgive me for not writing long treatises. On the one hand I have a small daughter here, on the other there's some sort of circus with clients (who apparently no longer exist — the Western ones can't pay, and ours are Western anyway and will gradually leave the market).

In other news: I installed a neural net, did a bit of a rain dance with it, and came away with a new tool for work. It's called Disco Diffusion — the first Google result takes you to a Colab, you can try it right there. YouTube is full of video guides, so I'm sure you'll work it out if you need to.

Meanwhile, here are a few images I've generated over the past few days.

(and:) There's plenty of incoherent nonsense too, of course. In fact, mostly what comes out is incoherent nonsense 🙂 All the same, I can't stop.

It feels like everything around me is shouting.

When people are shouting, it's very hard to gather your thoughts and write anything coherent. I try not to comment on the political agenda — I've explained why many times.

This channel is no exception: it's where I collect my own private world of interesting things. Apparently I'll also be writing something about myself here from time to time, if the censorship reaches social media.

On 13 February my daughter Maya was born. I don't have more time now — rather considerably less. Coming up: finishing off a couple of interesting thoughts about Ilya Repin. Over the past week I've absorbed a lot of historical material, including memoirs and Chukovsky's book.

Also: we're quietly greasing the skis and warming up the tractor. If anyone has work offers that involve relocation (I'm a concept artist and art lead, Katya is an art director and print manager), we'd be glad to talk. Our original direction was Canada or Australia, but in current conditions the choice is more about "from where" than "to where."

A small story about creative torment.

Repin was incredibly self-critical. Like any "amateur," he kept the little lamp of his own greatness burning very deep inside (and worldwide recognition came to him fast — he wasn't yet thirty when he painted Barge Haulers). He praised other people's talents extravagantly and was completely dissatisfied with his own canvases. On top of that, being a workaholic, he spared himself nothing and worked in binges without rest.

It's worth understanding that Repin was a narrative painter: in his work, meaning always dominates form. It matters less how something is painted than what. Pose matters more than rendering. You can read it even in his portraits — dishevelled, drink-ruined but magnificent Mussorgsky (painted in hospital, days before his death); sickly, hypochondriac Pisemsky with his damp trembling hands; the enamoured, unattainable squint of Elizaveta Zvantseva, and so on.

But that doesn't mean Repin didn't want the how. Like any enthusiast, he too thought about doing a fashionable little whoosh and stepping away from realism — just making something pretty, dab-dab, with a bit of pretension. Yo, symbolism, Vrubel.

He tried several times. The result was choice garbage; they're considered Repin's worst canvases.