Skype was fresh out of a ten year spread across the entire world doing the same thing for calling if you had a computer. Meanwhile, telecom giants' business was huge and immovable, so the best they could do against free messaging was lower price metered messaging. Not convincing enough? How about lowered, lower priced metered messaging, but only for the first year and with lots of disclaimer footnotes? You sure?

Prior to that, Valve was similarly eating up the gaming market with Steam. While every publisher was focused on preventing piracy by treating everyone like a potential criminal, Steam just picked piracy's preferred distribution mode – online and instantaneous – and used it to give users their own persistent libraries they could install from on any new machine, provided they could log in (making Steam itself DRM, but customers didn't even have to notice it). Publishers were a mess during the same time, because what they eventually offered was an insincere halfway point – physical boxes in big stores, containing a single DVD with a 300mb installer that would download a 60gb game and use up a one-time code in the process, leaving you with none of the advantages of the physical copy (they even did away with booklets and manuals!), and nothing gained from the downloading either. They did not compete because they could not give up a business model they had structured themselves around.

Netflix, arriving just ahead of cable television's unbundling apocalypse, again did the same. Cable television didn't really want to compete with fixed-price monthly access to a complete library. They wanted everything interesting to be an expensive add-on you need to pick separately. Their answer to "everything for one price", again, was "complicated bundle choices but at a lower price". A few years later, to your absolute surprise I'm sure, a great migration started.

I'm sure you see the pattern. In all of these cases, giants have structured themselves around a group of preconceptions of how a given business needs to run – video media is a list of subscription options that unlock live channels, cell phones are metered devices, game distribution is first a copy protection mechanism, etc. To be clear, the incumbents usually adapt over time. They tend to have the capital and patience to get pummelled by new market behaviours before they are forced to change.

In the context of a crystallized market, the true advantage challengers arrive with is the absence of artificial gates or scarcity, that once put in place, are too profitable to take away, even once time passes and you find the whole market running away from you.

Please everyone meet the real person who is the winner of the grand prize in this week’s Loto 6/49, for which you could buy a ticket, wink wink.

The awesome value of lotteries and luck games in a money-first society is that they convince people to buy a participation, based on results they’re almost guaranteed not to receive. To one person, paying ten dollars to receive ten million is an absolute outsized advantage. A few hundred people will receive sub-prizes that are better than break-even. And millions of people pay the same ten dollars to get nothing.

A healthy lotto product still has to work out, financially. Its expenses and payouts have to be covered by the ticket sales. The best way to achieve this is to have very few winners, maybe one winner singular, with a ridiculous grand prize. With an astronomical potential gain, you can expect buyers to completely bypass rational analysis of the cost and benefit, which generally means lottery buyers focus on the largest possible prize. Of course, then you pay out to a winner you put in the news and in ads, fuelling a sort of Skinner Box by proxy for everyone else¹.

Now – allow me to do what will come to be known as the classic blogpost sidestep of 2025, and make this about generative LLMs.

The new, new wave of private-funded market grabbers

Let’s split these lottery characteristics I mentioned into two groups:

• Outsize gains are given to very few winners, who then act as marketing to sell further participations.
• The balance of expenses and gains must be financially healthy.

What I’d like to suggest here is that the innovation of LLM infrastructure providers as we know them today is to adopt the advantages of group one, while being able to ignore the obligations of group two.

A lot of modern VC unicorns have done some form of this scheme. The current ideal for a startup investment seems to be an incumbent in a market successfully avoiding a key responsibility that should be part of their presence in said market. They are structured to burn money in exchange for subsidized market dominance, letting them grow fast enough to stay ahead of legislation and judgements.²

But under this model, they still have to show financial health, eventually. That’s when Uber goes from cheap and luxurious to pricey and terrible, for example.

With LLM builders and infrastructure providers, they've managed to do away with IPO pressure, and go straight for deals that funnel money out of nation-level budgets into private pockets through third-party provider arrangements. These deals, of course, inflate valuations without requiring the financial transparency an IPO would demand. It's less an Instagram unicorn, and more of a Palantir: old money clumped into a papier mâché approximation of a startup.

The innovation is avoidance-stacking: first social responsibility, now financial responsibility on top of that. Legal arbitrage, sure, but now we have companies running emotional arbitrage. They gain ground so long as users and investors maintain an irrational relationship with the product and the company.

I’m of the opinion that while there is great utility to be found for LLMs, their measurable value is being obfuscated by the egregious fantasy the big players have to tell about their relevance and growth prospects, which itself sustains the egregious financial fantasy of their funding and deals, that at this point they can’t question without everything else breaking down.

Of course, the fantasies will eventually stop working, but so long as they have lottery winners to show from time to time, they should continue their fragile climb, and funnel substantial chunks of the working economy into a few private wallets.

This is my attempt at a personal audit. The activities below aren't aspirational. They're what most consistently move the needle on my mood and general mental state. Some are obvious, others I eventually realized over long periods of time. The list is unordered, and non-exhaustive.

Unstructured time

The main, overarching pattern for me. Probably the reason I'm freelance, the reason I became a father late-ish in life, and my hardest to manage, even though it's the most fruitful. I need open time that's unscheduled and unstructured to feel like my life is... mine. This counts for work as well as personal time. Any project that has an exploratory aspect throughout leaves me feeling way more balanced and energized, compared to something with a very strict process and timeline. After mulling over the draft for this paragraph for a few weeks, my general observation would actually be that I thrive with constraints but push back against structure.

Time with family and friends

I will group these under the umbrella of: Social moments that reminds me I have good friends and I'm part of a robust support network. I'm rather introverted – I'm lucky to have a core of people close enough to me, and known enough to me, that we can get together without ostensibly draining my social battery.

Playing music with people

I'm not much of a team sports person, but my way of, uh, communion if you will, is to play. Music is its own language outside of language, live and visceral, and I don't know of anything that feels like playing music with other people.

Meditation

I have an on-and-off relationship with meditation, though it definitely has a positive influence on my mood. During the periods in my life when I'm consistent with sitting, I'm much less prone to feeling disorganized and anxious. However, I have a hard time working out the causation vs correlation here. Falling off the wagon is a sign I'm not doing so good anymore, but as of writing this, I'm not clear if it's the falling off, or the doing worse that comes first.

Working out

To no one's surprise, exercise is good for you! I lean towards individual action sports, with my personal love being street trials. Think an almost-BMX, with lots of braking, balancing in place and climbing over stuff. I live in a northern country, so bicycle sports are out of the equation for almost half the year, during which I ideally do strength training 2-3x a week. Since I became a father, and personal time got very restricted, the gym outings are on hold and I go out for a run or bike ride when I can.

Art and writing

I find it an irritating contradiction that I have such high get-to-work inertia, while at the same time work produced is one of my straightest paths to feeling fulfilled. Art gets me to that place without the hurdle of matching my output to someone else's expectations. It gives me a process that defaults to exploration, and a pleasant, tangible outcome for my time spent.

I think I like my definition of what constitutes art and creativity to be very open, and it feels natural to me to include writing here, even if it's fact-based and dry (hello to you, reader of this very article). Writing allows me a more direct way of self-reflection. It also lets me exorcise whatever might be sticking to my mind, unresolved or waiting to be integrated.

Media & video games

Coop games with friends

I never tried Fortnite, WoW or LoL, but video games have nonetheless been a reliable third space for portions of my social circle. I'm writing this on a day my friend and my brothers are planning to play together in the evening. The genres we pick are mostly shooters with a coop element, and driving games with a chaotic/open element – Wreckfest, BeamNG.

Solo games

Games that ask some level of planning/organisation and methodical thinking are relaxing to me, especially if they're also open in some ways. I didn't dive into Factorio yet – early fatherhood and whatnot – but there's a broad category that includes Zachtronics games, multiplayer Minecraft (Tekkit!), some Civ-style games, city builders, etc. that fulfill my need for open exploration and tinkering. Some personal highlights, off the top of my mind:

• Opus Magnum
• Mindustry
• Dyson Sphere Program
• OpenTTD
• Rimworld

Fiction on screen and in writing

My dose of escapism comes from reading – I like sci-fi and fantasy, but I think I more generally love good prose that lets me visit other people's imagination. While I've been on a local authors kick (French Québécois) the last 12-18 months, one of my north stars in fiction is Ursula K. Le Guin's A Wizard of Earthsea. From the opening chapter:

His father, the bronze-smith of the village, was a grim unspeaking man, and since Duny's six brothers were older than he by many years and went one by one from home to farm the land or sail the sea or work as smith in other towns of the Northward Vale, there was no one to bring the child up in tenderness. He grew wild, a thriving weed, a tall, quick boy, loud and proud and full of temper.

On screen, I'm currently enjoying what might become one of my favorite series, Empathie by Florence Longpré.

I was raised in an emotion-shy family, especially male role models, and fiction seems to be what helps me most to experience the full spectrum of emotions.

Patterns emerge

Going over the list, some themes seem to run across the spectrum of activities. Autonomy shows up everywhere – the need for unstructured time, creative work without external expectations, games that allow open exploration, sports that don't pit me against others, etc. Connection with others, in my case, gravitates around what I might call "quiet togetherness", with playing music together being its high-arousal equivalent. I clearly have a bone to pick with my relationship between structure and freedom, where I seem to need enough framework to feel grounded but start to fight that same structure the second I feel constrained.

In such a model, adding an item is easy (just type it out), while removing it is considerably harder (perform the action/task, then check it out). Overall, an easy, unbound list capacity shifts the burden of discipline from doing what's needed to only writing down what needs to be done. ¹

This is problematic, because any "primary" actionable list tends to lose its value as it becomes filled with secondary tasks that never make their way back up (hello to every client's Jira backlog). Once a list is filled with enough lower-priority tasks, it becomes a secondary list, and people tend to create another, different list, that captures the "true" primary tasks.

One moves to a new to-do app, creates a new email address or adopts that new, finally right, Notion template that's going to change everything. Except that the main feature is probably that it's an empty list – an absence of baggage.

The way I address this, personally, is that my short- and mid-term planning happens on paper, in what looks like a bullet journal if you squint a bit. Paper has two desirable characteristics this way:

• There's a finite amount of space to write things for a given time (in a day, in a week)
• Items that stick around week after week have to be manually transferred every time. If they fall in relevance, they eventually disappear.

This particular choice of tool is clearly an answer to my own strengths and shortcomings. I prefer tidy andfocused action lists, I don't mind the paper journal tedium, I'm great at holding a vision/direction but very poor at specific long-term planning.²

It's hard to deliver sustainably good design in the presence of leadership that shifts focus (changing it too fast or out of sync with their org's execution capacity). Design's fuels are intent and domain information.

Startups famously pivot, but other structures similarly deal in shifting incentives and priorities. In that context, the best results I can offer come from my own ability to render quickly, and whether I can spec out solutions that are also fast and straightforward to produce.

But also – it speaks to a core issue in an organization. Companies with unfocused leadership generally end up firing their CEOs or burn down trying. Or, with tremendous amounts of money available, some of them successfully structure around making the CEO as useless and inconsequential as possible.

All this to say that design without leadership alignment (or without being a designer in a position of leadership) always feel like it maxes out somewhere around 50% efficacy.

Dosage, brand and frequency

From what was locally available at the time, I picked capsules containing 600mg of KSM-66, which is Ashwagandha’s most studied (and hopefully standardized) extract. I took one capsule daily in the morning, with food.

In retrospect, I would have liked something I can divide or otherwise portion, as 600mg is at least twice the low/minimal daily dose I want to try out.

Note that it’s the only supplement I’m taking, I have no medication I use, and I drink little to no caffeine, meaning I didn’t need to consider drugs it could interact with. For someone on existing medications, the process to consider taking Ashwagandha might be a bit more involved. 

Sidebar: What’s Aswhagandha?

If you’re already familiar you can skip ahead. It’s a nightshade plant (Solanaceae), close to ginseng, from which the roots are dried and powdered. Its name means "horse smell", which is apparently an apt description of the raw product. It’s also called Indian Ginseng. It has a long history of use in Indian (Ayurvedic) medicine. It now gained a decent level of interest as a modern supplement for different effects it can have – mainly on anxiety, but also on recovery in athletes.

First days

On the very first day, I had a moment of noticeably racing heart a few hours after taking my first dose. I attribute it partly to the initial dosage I couldn’t progressively introduce, and partly to the existing anxiety I had at this time.

At the end of the first week, I forgot to take a capsule in the morning, then I suspect I did forgot again the next day. This coincided with two days of higher anxiety (back to never being far from a panic attack). Again, since it happened very early in the month, it’s hard to distinguish between the effects/presence of anxiety itself, and what the extract might have influenced.

Positive observations

Ashwagandha did make a noticeable difference. In short, before that month, it had become extremely rare for me to come back down to an actual calm state. Anxiety tended to "stick" to me, and I had adapted to a high arousal, low energy state as my new baseline (stressed and sleepy – it sucks). All my emotions felt prefixed with " stressed and X", and it was very rare for me to simply feel present and in the moment.

Between the second and third week into the month, I found I had regained access to an actual calm state. My baseline was back to an enthusiastic calm, and the spectrum of emotions that were available to me opened back up.

In a practical sense, I was again able to act mindfully during the day. Anxiety is a thief of focus, and it drains me of both energy and attention when it’s as high as it was previously. I could feel (hear?) some of my usual mental loops and triggers happening throughout the day, but now they usually failed to take hold, and fizzled out instead.

In my case, the most gainful change was that the physical manifestations of my anxiety shot way down. I suspect that the strongest loop I get stuck on is the concern of how anxiety itself can disable me in a moment with physical effects. Unfortunately, it’s tied to interoception (sensing one’s internal physical state). The physical manifestations of my anxiety crank up every time I notice that ... they crank up, which is a bad feedback loop in which to get stuck. Once Aswhagandha’s effect had settled, I would say the physical component dropped to maybe a tenth of what it was. I would recognize the sensations, but they didn’t raise to an intense level. This completely disarmed the physical-mental feedback loop, and afforded me the opportunity to come down to a calm state after most stressful moments during the month.

The second-order effects from the above changes is that I have more energy and patience in the day, and better sleep at night. It’s nothing groundbreaking, but significant enough to notice.

Concerns

Low/no anxiety is generally also a lower arousal state, and I suspect I personally exist on a scale that moves from calm and unfocused to anxious and productive. At least, that’s how I live and react to the external motivators of our capitalist structure. Ashwagandha has some effect lowering my arousal and focus.

Because anxiety itself robbed me of a lot of mental acuity in the last months, I don’t think the tradeoff of calm and potentially lower focus is a net negative in my case. I didn’t really have access to the calm state it compares to anymore. I think low-stress people taking Ashwagandha for other benefits might be at risk of slipping into disinterest or lose motivation as a side-effect.

From reading, I have another concern I’m entirely unsure I should address. A lot of people talk about cycling on and off (for example, one month on, then one week without Aswhagandha) to prevent acclimatation. However, all of the cycling talk comes from the nootropics web communities, and that’s way out of my wheelhouse, partly because it’s people who are seeking lots of bleeding-edge products and are used to drawing conclusions from very limited data. Ashwagandha, in comparison, has a long history of use (a few millennia), part of which is daily use in a herbal tea for, like, your whole life. I still want more studies done to help guide how (and if) I use it, but at this point, I don’t want to put too much stock in the cycling advice so long as a low dose seems to preserve its effect over time.

Nonetheless, I’m finishing up this article on the end of a week off, as a way to notice the difference Ashwagandha was making. It’s helpful. I currently retain some of the mental flexibility that I seemed to regain while taking Ashwagandha daily.

Conclusion

I keep thinking back to a blog post Zach Holman wrotein 2017 on depression, and specifically to the ✨ Drugs ✨ part, where he discusses the apprehension and stigma related to drugs, but accurately points out that while some people lose themselves through the effect of a drug, we also use psychoactive substances to get access back to ourselves, as a supporter of the rest of the actions we take to heal.

My limited experience in that month points in a similar direction, and the next step for me is to see if Ashwagandha can give a boost to the other efforts I make, long-term, to restore access to my calm self.

Update from a month and a half:

My observations remain roughly the same. I have forgotten a dose again one time, a Saturday, and that day went without a noticeable difference. The next day, I didn’t forget, but I did have much higher anxiety than usual, which resolved during the week. In the coming month I’ll be reducing to 300mg per day to see if I get the same benefits.

Update after the second month

I used the momentum from lowering the dose to stop completely shortly after. As of writing this sentence (six months later), I haven’t resumed any supplementation, though I have had short periods of heightened anxiety. They track my exhaustion levels and didn’t seem to devolve into chronic anxiety – fingers crossed.

Almost every time, I make these approval emails a "default forward" situation. If the client doesn't do anything, I keep the project going. It's usually a simple matter:

Hi Cassandra Client!

Here are the wireframes that detail how users pick their organization type, and can configure it for themselves or the whole team.

The next step for me is to turn these to high-fidelity mockups, and then deliver the styled components for development. I'm moving to mockups early next week, don't hesitate to share your notes by then as needed.

Have a good day,
Louis-André

This has two effects: First, I don't get caught in approval limbo. Second, my client doesn't get caught (further?) in an email catchup race.

I favor async communication and relaxed timelines. This fits async communication very well, and having reasonable delays with a default forward at the end helps preserve the available slack in said relaxed timeline.

Depending on the type of client you work with (specialists vs generalists vs out-of-industry) it also gives an out to people who prefer to give full autonomy to their providers. Some clients want/need to be very involved in the design process, but others will politely nod along each milestone review then tell you to please continue thank you. Default forward serves them as well.

I'll add one precision, which is that I draw the line at "forward as previously agreed". I won't send an email that defaults to an unexpected expense or a dramatic change to the project's direction. There are plenty of synchronous or interruptive ways to have this conversation.

B?

The series' predecessor, Territory A, delved into the idea of occupying time as it moves forward. Territory B's focus is about maximalist use of space – whether it ends up being efficient sharing or outsized greed.

It's inspired by societal artifacts like ad-laden web pages, British land division, disk space tools, and wild posting.

Division, unity

Territory B applies a sort of kitchen-sink approach to filling out a composition. It subdivides a grid, then throws everything at it to get the space filled out.

The underlying noise source, however, is unified and continuous. Just like land division, the separations created in Territory B are ultimately artificial. Its compositions usually let the viewer see through the large blocks and feel the underlying field. Some elements describe it through angles and alignment, others by crowding certain areas.

Exposed features

Six named features describe some of the main behaviors in each iteration:

Palette name, and palette variant expose the choice of colors in a composition, and whether it is reduced to a subset of its colors.

Small blocks means the composition will not allow larger subdivisions to exist inside its grid.

No padding crowds out the space by completely cutting off the possible spacing between each block.

Squeeze cross axis will pick a main and secondary axis, then significantly increase the amount of possible rows or columns on the secondary one.

Color fill type is usually Block, meaning each block in the grid selects its color for its own fill patterns. Some iterations make an exception to this and have the Precise value instead, where each individual segment drawn by the script picks its own color from the underlying noise source.

Block types

Territory B has six global fill types that it applies to blocks in each iteration. The underlying field influences the block choice, making similar styles crowd together.

Three types are sub-grids that use rectangles, circles and lines to fill out. These types sometimes react to the field by picking element sizes or line angles based on its strength.

Two types fill out with a focal point – one centering on another rectangle inside the block, and the other drawing rays from the focal point to the edges of the block. The offset of the focal point is driven by the value field.

The sixth type is pointillism. The distribution of dots will often conform to and reveal the underlying field as well.

Release

The collection was released on fxhash on Tuesday, March 7th, at noon on Atlantic time.

By reading this post, non-technical people should understand the mechanisms that draw a Maps for grief iteration, and someone technical should have enough of an explanation to code their own take on the same idea.

How each iteration is drawn

The basic structure of Maps for grief is the interaction of two noise fields. A noise field is a map of random values across coordinates, with rules governing how its values change across its area. In our case, each field is two-dimensional, and filled with Simplex noise. This means that two points close to each other will have very similar noise values, while points further apart could be more different. You can picture a map of rolling hills, with the lowest points having a value of 0, and the highest ones, 1.

Each of the two fields has a specific role: The first one is turned into a flow field – each value represents a certain angle. You've certainly come across art based on flow fields in the past – like Tyler Hobbs' Fidenza. In our case, the flow field is used to orient how each line moves in the space once it starts being drawn.

In the rules I set for drawing lines in Maps for grief, each line has a finite length that it attempts to draw without touching another shape. Each composition tries to insert a total of ten thousand such lines at random positions within its margins. The first ones always have the space to be drawn, while the very last are usually all skipped, since the space is already filled by previously drawn lines.

If we simply drew based on the first flow field, then the artwork could look pretty, but it wouldn't have much substance. Here's how the line work would turn out (I left most other rules in, which is why the first one is split in two regions):

The second noise field, as it grows in value, forces these lines to stop drawing sooner, and at extreme values, prevents them from drawing entirely. It's also responsible for turning "stopped" areas into symbols instead, and it affects the colour selection across the composition. By adding the second field's influence on drawn lines, we end up with results like these:

The last part is inserting symbols. The most on-the-nose part of Maps for grief is the drawing of these symbols, which appear in areas where lines can't flow. These are the regions where the secondary field is the strongest. Each iteration has a single type of coping, deciding what type of symbol is drawn. There can be different drawing rules for different symbols, which helps give each of the coping types a specific personality.

Putting everything together, here's the sort of outputs we get:

There are a ton of small pieces and details, but that's the gist of it. Let's analyze the main features involved.

Main feature & rarity breakdown

Coping

Coping is one of four values: raw (drops, or tears), nature (mountains), mystical (asterisks), and spiritual (crosses). Crosses and mountains can huddle very close together and create areas that change in density, tears follow the flow of the field, and stars vary how much branches they have in reaction to the secondary field's value.

Rarities (coded / actual):

• Mystical 24.8% / 25.2%
• Nature 24.8% / 22%
• Raw 24.8% / 24.2%
• Spiritual 24.8% / 26,4%

Palette

There are six palettes. Four of them are dark: Nebula, Terminal, Winter forest, and Witching Hour. Two other are light: IKEA Showroom and Parched. Note that any palette has a chance of being subsampled (this is an unlisted feature). A piece with multiple regions can pick between its full palette or any of its subsamplings when it starts drawing each separate region.

Rarities (coded / actual):

• Terminal 16.6% / 15.4%
• Nebula 16.6% / 15.6%
• Winter forest 16.6% / 17%
• Witching hour 16.6% / 18%
• IKEA Showroom 16.6% / 16.8%
• Parched 16.6% / 17.2%

Stroke length

Lines don't all have the same maximum length, and for each piece, this length is set within a certain range. The range is picked at random on a sliding scale, but for simplicity's sake, all pieces are classified among three length groups. Short lines are capped under 1/12th the width of the piece, long ones can go over 1/10th of the piece, average lines are between the two.

Rarities (coded / actual):

• Short 32 % / 31.2%
• Average 27.8% / 25.6%
• Long 40% / 43.2%

Tip size

There are four predetermined tip sizes. I see them as guidelines to the scale of an iteration: Fine tip pieces are the best ones to present at a large scale, while fat marker ones (my favorite size, personally) look great even as thumbnails, but can lack in detail if they're too big.

Rarities (coded / actual):

• Fine 25 % / 25.8%
• Average 37.5% / 38%
• Thick 25% / 25.4%
• Fat marker 12.5% / 10.8%

Region count

There are four possible panel arrangement, with single-panel being the most probable. An interesting interaction is that larger tip sizes can almost entirely hide their center seams (which might still be revealed by a change of subpalette), while smaller line width yields clean divisions between each region.

Rarities (coded / actual):

• Single region 50 % / 52.8%
• Two 25% / 24%
• Three 20% / 18%
• Four 5% / 5.2%

Special features

There are a few features that recognizably alter how a piece is drawn. These are features that are simply "on" or not.

No dots

Usually, in a Maps for grief composition, some line insertions that should be symbols will draw a simple dot instead. Pieces with this attribute forego the dots, leading to sparser symbol areas.

Rarity: 2.5% / 2.8%

Persistent lines

I said previously that each line's maximum length is picked on a sliding scale. This scale is supposed to be affected by the strength of the secondary field. Pieces with the Persistent line attribute have lines that are unaffected and attempt to run their full length no matter where they end.

Rarity: 5% / 4.6%

****

Radial

This is an attribute that adds an additional influence to the primary flow field. Radial outputs have their lines aligned in relation to a center that pushes them in or out. Combined with the usual forces of the field, it often yields a spiral composition, though that's not the only possible output.

Rarity: 5% / 4.6%

****

Confused (special coping type)

Very rarely, an iteration will not actually have a set coping type. The distribution of coping types should be a flat 25% for each, but it's lower than this because in a very small amount of iterations, the confused attribute means the piece will pick a random symbol any time it has to draw one. There are only ten pieces with this attribute in the whole collection, which is still double the coded chance of them occuring.

Rarity: 0.8% / 2%

****

Invisible and minor features

There are other dimensions that are decided for each iteration, with varying levels of influence on its structure and aesthetic.

Palette variants exist for each of the six palettes. The way it works is that each named palette actually has two possible accent colours, and two possible backgrounds, however each iteration picks only one of the accent and background colours, along with a certain amount of intermediary colours.

Drawing jitter determines how "shaky" the line drawing is. This is set in relation to tip size as well, so the outputs with the most jitter are ones with Fat marker tip size. You can get a feel for it by pressing a on the keyboard to animate an iteration in its live view.

Line spacing can vary a bit, and decides how much distance lines must have between each other.

Radial field strength affects compositions with the Radial feature. When very strong, it forces lines to point almost directly to the center, while a low value will draw a soft spiraling composition.

Force field scale decides how much variance can be packed within the surface of an iteration. Pieces with a small scale (think of it as being zoomed out) tend to exhibit small "continents" of symbol areas separated by lines. Larger-scale ones will rather show a single movement across their area.

Flow field base angle and angle range determine the main "axis" of the flow field, and how much lines can veer away from it. Composition with very low angle range are almost linear, while compositions with high angle range and sufficient force field scale are the most turbulent.

Variable values

Some variables are not exactly set by the seed, as they are adjusted by the generation process later on.

Anomaly threshold (or symbol threshold) is the value past which the secondary field forces lines to turn to symbols. Because the output of the field always varies, the generation process "scans" the field before committing to it. If it finds too little variance, it will lower the anomaly threshold, so that a lesser field strength can allow symbols to appear.

Field offset also affects the secondary field. When the generation step finds too little variance in the field, it also moves it around a bit – think of it as moving a picture frame above a larger canvas, to see a different part of it. This is done in conjunction with a light change in the anomaly threshold.

Palette/restricted palette is a changing list of available colors that the drawing script picks from. In single-region outputs, it does not change, but in multiple regions, it has a chance to change and pick a new subpalette (or go back to the full palette) before each.

Dive in!

If you want to try your hand at picking out the expressed features of a given iteration, feel free to browse the collection. It's possible to do so on Artblocks where it was minted, on places like Artacle, which will let you filter a collection like this using its different features, and on Opensea, which is more of a marketplace.

Oh, still here? Okay – if you're interested, I've done three of my own analyses to get you started. Enjoy, and thanks for reading!

That's iteration #443. Thick line width, long strokes, single region. Its crosses means it has spiritual coping. It has the intense magenta of the Nebula palette. And since it's only showing magenta, we can tell it's subsampled. Its long lines also extend into symbol zones, meaning it has the rare persistent lines feature. We can also tell that the angle range is rather narrow, because the darkest and the lightest parts of the composition are oriented within a quarter-turn of each other.

Here's iteration #430. It's a three-panel composition, and again uses the Nebula palette. The first two panels seem to use the full range of the palette with a magenta accent, while the third panel has a subsampled palette based on the turquoise accent. With lines this narrow, the tip size is probably Fine. This time, the noise scale seems rather large, because there's a lot of variation across the whole piece. Similarly, its angle range is quite high – the flow field is able to turn back on itself, and it's hard to tell what the base angle might be. In comparison, it was very easy to guess the average angle in #443. As a consequence of its narrow line width, it will also have rather precise drawing jitter, meaning its animation will be softer, up to pretty large display sizes.

And here's #23! Again, a three-region composition, this time with mystical coping, as indicated by the stars. If I want to follow how strong the secondary field is, I can look at how large the stars are, and how many branches they have. The third panel also seems to have picked different transition colours from the IKEA showroom palette the piece has. The top-right corner lets us see that it's showing a light blue where the rest of the composition would use the light purple/gray (lower left). It also has the rare radial attribute, and its radial field strength is very strong – we can hardly tell the influence of the primary flow field over the drawn lines. It also has an off-center radial pull, which only happens to half of radial compositions (that's another untracked rare feature for you!).

Maps for grief is a generative art project, published on the Art blocks platform. It explores the notion of togetherness without contact. As with most series, I find it ties together many projects and techniques I've done in the last two years. Let's look at a few of the building blocks behind it.

It released on January 11th, 2022, and minted 500 pieces. You can see the project on Artblocks here. The above image is test mint #12 on the Ropsten network.

Convergence

One of the first generators I've made and deeply liked is called Convergence. It uses a simple flow field to create undulating clusters of lines, then peppers them with small blooms of circles.

It's the result of several explorations on flow fields and branching graphs. Eventually, it became the foundation on which I created Maps for grief.

Glitches

The idea behind Convergence's blooms comes back in a lot of my explorations, where I tend to establish a certain structure in the piece, then add odd parts that are less conforming to it. Internally, they're usually called glitches or anomalies.

Sometimes the anomalies themselves become the main subject for the piece, as is the case with Glitch circles.

Other times, they bring variety and interest into a structure that would be too rigorous or monotonous on its own. The following was an exploration of purely using anomalies, turning a simple composition into something interesting by inserting "mistakes" in the drawing process:

Maps for grief rests on the same technique of creating, then breaking, a given composition. At its core, each iteration tries to draw lines that flow side by side and never touch. However, these lines run into areas that dry them out, and force them to shorten, sometimes turning into clusters of anomalies when the secondary force is strong enough.

Topography

Architecture and natural structures are among the influences I can see repeatedly in my work. While they can be seen as competing aesthetics, I find that disciplines like urbanism and cartography have built a vast repertoire of work that reconciles the two.

Maps for grief leans heavily on a more flowing, nature-like composition, yet it inherits a lot (visually and technically) from explorations like Urbanisms or Boroughs.

Pen-like tracing

Early in my generative art explorations, I started building a plotter-like library to trace lines in an organic way. It's nothing very fancy – it follows basic lines and shapes, and sputters along the way, resembling what pen ink would do on rough paper. You can find deeper explorations on the topic over on Sighack's Fifteen ways to draw a line. Over time, the rough line style became a recurring part of my practice, and it's now uncommon for me to build sketches that don't feature this kind of hand-drawn quality to the shapes.

Maps for grief includes an agressively minimized version of the technique – at the data level, each iteration is distilled down to a list of colored circles, which are then traced on-screen to form the lines and shapes. The small jitter is part of the drawing routine, which is why it's possible to make the sketch gently animate without updating the data (press a on the keyboard when on the live viewer).

Quick notes on the long-form format

This is my second-ish foray into long-form generative art, a term coined by Tyler Hobbs to describe generative art that is created for direct distribution (the generator script is part of the art being delivered to viewers) and uncurated generation (the collector and the artist both discover what a specific iteration looks like at the moment it is sold and created).

The benefits and constraints of this form have already been discussed and laid out by others including Tyler, so I will simply talk about what I resonated with while building Maps for grief.

Subtle variations, less variability

The necessity for all long-form outputs to remain above a certain level of quality means that there is less of a "conversation" between the artist and the possible outputs at the end of the project. When manually curating runs, I can tolerate a certain rate of "failed" outputs that are either uninteresting or too chaotic and wild. With the tolerance afforded by manual curation, I get to explore a wider range of these possible outputs, including the interesting extremes, the good middle-ground. While I can keep the interesting extremes, at the same time, I get to discard the uninteresting extremes. With uncurated generation, all variability has to be constrained into a "good middle" instead. Which leads us into the second point:

This work requires a different kind of critical view

I found the project hard on my perceived critical sense as an artist. When judging curated series, the quality of the series is defined by the best iterations, which I pick out manually. Over time, this work accumulates as a corpus of past series, each of which remains represented by the best iterations I selected from it.

As I explore and build on previous work, the level of quality I refer back to is this running average of the "highs" from each project. Consciously or not, I weigh each series against it.

With long-form art, the series' quality is suddenly defined by the average of all its iterations, and how remarkable the best iterations are, and how "uninteresting" the least interesting iterations are. It's been difficult to recalibrate my perspective on output quality until I could bring the variability and features to a good place across the board.

Some example outputs

Here are a few iterations I shared from the Ropsten (test network) mints.

#22

#14

#11

#34

Technical and rarity breakdown

See the technical breakdown here, in my follow-up post.

Impossible d'avoir un mot de passe différent pour chaque service où j'ai un compte. Je les oublierais tout le temps!

C'est vrai – je sais pas pour toi, mais j'ai bien trop de comptes pour me rappeler d'un nouveau mot de passe à chaque fois!

La solution par contre, ce n'est pas de réutiliser le même mot de passe à plusieurs endroits. Je répète en plus gros pour être dramatique et parce que je t'aime:

Ne réutilise jamais de mot de passe entre tes comptes.

Si ça t'intéresse, j'explique pourquoi c'est dangereux à la fin de cet article ↓

Okay le smatt, comment je me protège sans que ça soit fatiguant?

Avec un gestionnaire de mots de passe! Grosso modo, c'est un outil qui mémorise tes informations de connexion pour tous les services que tu utilises.

On choisit un gestionnaire de mots de passe parce que son stockage, lui, est très sécurisé. Pas mal plus qu'un post-it sur ton écran ou un fichier texte sur ton ordi dans le dossier Bureau/Dissertations/Aristophane, la bibliographie complète

L'ensemble de tes mots de passe va être contenu dans le stockage du gestionnaire qui, lui, est encrypté par un mot de passe principal (master password).

Tu te connectes à ton compte de banque? Clique sur ton gestionnaire, entre ton mot de passe principal, et le gestionnaire fait le reste.

Tu te connectes à un service du gouvernement? Clique sur ton gestionnaire, entre ton mot de passe principal, badaboum.

Tu te connectes à Skype sur ta nouvelle tablette? Gestionnaire, mot de passe, voilà.

Ton compte pour faire une commande d'épicerie? Gestionnaire, bim bam.

Tu comprends le principe.

Bref, le mot de passe maître, ça devient le seul mot de passe que tu dois mémoriser. C'est pas mal plus pratique.

En plus, pas besoin que ton mot de passe principal soit difficile à mémoriser. Quelque chose comme Un bourdon marche seul au plafond c'est beaucoup plus sécuritaire que Y&gT81zl~ke#s@, en vertu de sa longueur tout simplement. En plus, tu t'en souviens déjà.

Donc je prends quoi?

Voici trois bons choix de gestionnaires de mots de passe. Les deux premiers sont disponibles gratuitement pour les particuliers. Il n'y a pas de choix risqué dans cette liste:

• BitWarden
• Lastpass
• 1Password (36 $/année)

Personnellement, j'ai utilisé 1Password pendant des années et c'est un excellent outil. J'utilise maintenant Bitwarden.

Je devrais installer des trucs?

Je te recommande d'installer l'extension pour ton navigateur. Les gestionnaires de mots de passe offrent pas mal tous des extensions. En pratique, ça t'indique pour chaque site web si tu as déjà un compte, et ça te permet d'y entrer tes informations sans avoir à les taper manuellement au clavier.

Je te recommande aussi, pour le gestionnaire que tu as choisi, d'installer l'application sur ton téléphone. La plupart des gestionnaires de mots de passe resteront synchronisés entre les appareils d'un même compte.

Des bonus

Les gestionnaires de mots de passe comprennent toujours aussi un générateur de mot de passe. Donc, quand on dit d'avoir un mot de passe unique pour chaque compte, tu n'as pas à l'inventer dans ta tête à chaque fois. Tu cliques simplement sur «générer». C'est bon pour les nouveaux comptes que tu crées, mais aussi pour mettre à jour les mots de passe que tu as présentement. À coup d'une mise à jour de temps en temps, tu vas avoir des mots de passe uniques partout, sans vraiment faire d'effort supplémentaire.

Souvent, un gestionnaire de mots de passe va aussi te signaler si un service que tu utilises a été victime d'une fuite de données. C'est gentil, ça te permet de proactivement changer le mot de passe à cet endroit.

En résumé:

• Tu choisis un gestionnaire de mots de passe (c'est gratuit!)
• Tu te choisis un mot de passe principal, qui est long mais facile à retenir
• Tu installes le gestionnaire sur tes appareils préférés
• Tu t'assure d'avoir tout de suite des mots de passe uniques là où ça compte le plus (banque, courriel, Paypal, Skype, gouvernement, employeur, ce genre de places)
• Tu dors bien.

Et si mon gestionnaire de mots de passe a une fuite, lui? Ou si je me fais voler mon ordinateur?

Bonne nouvelle! Ça va se passer comme quand une compagnie qui prend sa sécurité au sérieux est malgré tout victime d'un vol de données:

Les hackers prennent possession du genre de coffre-fort (ou se sauvent avec un ordi dans les mains, on sait jamais). Mais ses données sont encryptées, donc illisibles, parce que personne d'entre eux ne fournir le mot de passe qui est la seule façon de les déchiffrer.

Sans ce mot de passe, les hackers ont en main un grand fichier rempli de charabia inutile. Et toi, tu demeure en sécurité. ✨

Pour les curieux: Pourquoi c'est pas sécuritaire de réutiliser un mot de passe?

En gros, c'est parce que ça prend un seul compte vulnérable pour que tous tes comptes le deviennent.

Pourquoi est-ce que ton compte Facebook serait à risque quand une fuite a eu lieu sur genre bingopromotioncirculaire.com? C'est que quand un service en ligne se fait voler une base de données mal protégée, la liste des informations de connexion de tout le monde se retrouve dans les mains de hackers.

Ces hackers ont des outils automatisés qui attaquent tout le monde en même temps. La liste contient des informations personnelles d'une tonne de gens – par exemple, ton courriel et un mot de passe que tu réutilises ailleurs. Ils ont des outils qui vont essayer, pour chaque personne dans la fuite, de se connecter avec les mêmes informations volées sur les services populaires – Netflix, Gmail, Paypal, Facebook, Amazon, Desjardins, etc.

Avec ça, ils vont voir qui réutilise son courriel et mot de passe ailleurs, et prendre contrôle des comptes où c'est le cas. Ça s'appelle une attaque de bourrage (credential stuffing en anglais).

Fais le test: Entre ton courriel sur l'outil Have I Been Pwned et regardes dans combien de fuites publiques il a été retrouvé à date. De mon côté, je suis prudent, mais mon email personnel a malgré tout fini dans 23 fuites majeures (Adobe, bit.ly, Last.fm, Trillian, MySpace, etc.) au moment d'écrire ces mots.

Quand tu as des mots de passe uniques pour chaque place, lorsqu'un service est victime d'une fuite pour une raison ou une autre, tous tes autres comptes demeurent à l'abri des attaques de bourrage. La clé, c'est de ne pas réutiliser la même combinaison de courriel et mot de passe. Et ça explique le gestionnaire, qui rend ça facile au lieu de quasi-impossible!

Here's a short video of the result

Our method leaves the tablet itself untouched, and the stand remains functional but only on its highest flap.

There's a short video of the result at the end 👇

Summary:

• Procure M4 nuts
• Sacrifice the two lower flaps of the mount
• Mark your VESA holes from the inside
• Punch/drill while resting on a solid, flat surface
• Check for screw clearance and mount

1. Find or buy M4 nuts

M4 is metric, but it should be easy to find. You'll probably need 8 (two per bolt). If you're in North America, look in your hardware store's hard to find section, or look for it in the fixtures/electrical department. Cabinet handles use metric, and you'll often find cutaway screws for different door depths. Other options include asking a mechanic (metric cars & motorcycles). M4 bolts are relatively thin, which is what's needed here.

2. Sacrifice the lower flaps of the mount

This is the only functional modification you'll need to do. Consider if you want to keep them, there are options for just creating a complete holder from an easel, for example.

You're fine with two less flaps? Good! After this, there are no dramatic removals.

Fold each flap around 90°, and use a sharp blade to cut in the crease. The plastic is supple and will come off easily.

3. Mark your VESA holes' positions

Lay the stand flat, and put the mounting bracket on it (doing it from the inside is fine). Make sure it's well aligned and centered, then mark the center of each hole.

We kept the mounting bracket within reach, since it proved easier to mount the Cintiq's stand on it this way.

4. Mark/punch, then drill.

I recommend first punching your holes right in the center, which will allow a drill bit to stay centered. The stand's backplate is metal and slippery.

Rest the stand on a flat, solid surface. A piece of wood is perfect. The metal will bend around your punch if it isn't solidly backed (it did for me).

You can use a small Philips screwdriver, or a point tip if you have one. Center your punch as best as you can, then give it one sharp hit for each position.

Pick your drill bit. M4 screws need a 4mm (duh), or 3/16" hole. If you have it, 11/64" is closer. 5/32" is probably too tight. Hold the stand well in place. Use a bit that can handle metal, let it bite and turn slowly, the plate gives easily.

In a short time, you'll have four clean, well-aligned holes!

5. Check for clearance, then mount

Well, this is exciting! You can affix the plate to the VESA mount!

Check for your screws' clearance past the stand. In our case, we added one nut to the head of each screw, to make sure it didn't poke into the back of the Cintiq once everything is put together.

Put the stand on, so that its "hook" tab (the one around which the Cintiq rotates) faces up. The remaining flap will also point up.

In a diagonal pattern, mount the backplate to the VESA bracket. Make sure it's tight!

Then add the tablet! Seat it on the stand's hook, then be sure to "click" the bottom into place.

Boom!

Pros

• The stand remains usable.
• More sightly than the easel trick
• Your 13HD is on a legit VESA mount!

Cons

• The Cintiq 13HD is actually too light for many monitor arms (they'll continuously push the tablet up instead of keeping it in place). Be careful when you pick your arm!
• The stand is metal, but certain arms will resist hard enough for it to flex if you try to tilt the tablet up and down. This is very noticeable on the Ergotron here, which is made to hold hard enough for a 20lbs monitor to stay in place. We tilt it by holding close to the VESA bracket.

I love the fascinated respect people have for the fruit of someone's craft. I love how one can afterwards still touch something, and feel the hands that made it. How we catch a glimpse of the human at the other end.

I like to think it's what we're trying to figure out with the web, and with things like mobile devices and the whole "app" construct.

We have touchscreens and cool interactions, but everything feels like it's behind glass (and well, physically, it is!). We try to make things better, and we often over-structure what we display into computerish coldness. Or, for want of skills and sensible priorities, we end up reaching said coldness before we breathe in some natural structure and humanity.

But I see (and hope) that we're getting there. It's in the way Basecamp's website is designed: with hand-drawn pictures and a wonky structure, that take nothing from the strength and clarity of their message, but rather infuse it with warmth.

It's in Justin Jackson's classic This is a web page.

It's in Frank Chimero's new website too, and in how it's summarized in The Space Between You and Me, which he wrote for The Manual:

The web is people all the way down.

The more we try to define and use processes that hide this reality, the less genuine it all feels. Hopefully, we'll grow a bit tired of thinking a "handcrafted" web means every pixel in a very rigorous place. Handcrafting is making something for the hands of others, and putting it out in the world, for others to find, to examine, and to discover the traces we left of ourselves.

Getting started with high-density (HiDPI) screens, like Apple's Retina display?

Several design apps support half/dual resolution out of the box, but a lot of us still use Photoshop, either out of personal preference or to play well with other team members. One way or another, I assume you're pretty damn efficient with Photoshop already. When I'm not using Sketch, here's how I handle Retina mockups with grace.

Open your preferences!

• Cmd + K on OS X
• Ctrl + K on Windows

Go to Units & Rulers. This is where most of the magic happens.

1. Under Units :

Set both rulers and type to points.

2. Under New document Preset Resolutions :

A standard screen is considered to have 72 DPI. A Retina screen is at 144 (double). You can set the screen resolution to 144 as a default if you so desire (I prefer this). Here's how it looks. You can disregard columns, I personally have no use for global column settings:

New documents

Now we can start working with points as our coordinate system. That makes things pretty easy. Say you want to design for Android screens, from sizes medium and up in portrait mode. That means 320 x 470:

If you're familiar with the developer side of these things, you're probably aware that dimensions and positioning in these environments don't work in pixels. They use points.

Here's the basic difference: iPhone screens, whether they're Retina or not, share the same coordinates system. They both "measure" 320 x 640 points, and that is how things are placed on screen. A full-screen image measures twice the amount in pixels, but remains 320 x 640 in points. These are the conditions we're reproducing.

Changing your current PSD to Retina/HiDPI

This method shoud be a big help if you want to keep working in your existing PSDs, and simply bump their resolution 2x. I'm assuming you're designing mostly with text layers, smart objects and vectors. If you're not, you'll have a lot of extra work to restore everything that becomes blurry.

Open your document as usual. Once there, open Image → Image size... or hit the appropriate shortcut:

• Cmd + alt + i on OS X
• Ctrl + alt + i on Windows

Then, change your dimensions' units (they're probably on pixels) into points too. Make sure resampling is activated. Then (and only then!) bump the PSD's resolution to 144 pixels/inch.

Once this is done, hit "OK". Review your document. You should be all set!

:)

Note! While this setup means there is a pixel for every half-point, you'll want to use whole numbers for position and dimension in your designs. This is because screens that do not have double resolution still use one pixel per point, so half-points will really land on half-pixels: this means blurry results.

Working with points

So what are the main differences?

Type is actually in points.

If you've ever worked in print design, this change will make a lot of sense to you. For screen design, what it means is that you can use "real" sizes for your text, instead of Retina requiring double the actual size. 14pt text in your PSD will look like 14pt.

Rulers are in points.

In the picture above, you can also see the rulers don't match document pixels one to one. The ratio becomes 1:2, because this PSD has a 144 DPI resolution! The coordinates on the rulers are what will be used in development, and they reflect their "real life" perception. You can use points in your info palette, too! If you don't use this palette, take this occasion to try it out. Set up your preferred units in the palette's option dialog.

It'll give you color readouts, the dimensions of what you have selected, and the coordinates (in points, if you want!) of your cursor.

That's it!

It's far from being the only (or the most complete) guide out there, but these are the basics that make it easy to uncouple pixels and actual dimensions in Photoshop.

From there, for example, you can easily create a 300DPI document for print, or a 18DPI document for this awesome pixel-art game you're working on – all without losing your head around typography and unit conversion.