Jaime’s art exhibits an intelligence about which the Silicon Valley digerati are clueless, a conversation with Claude Opus 5

You may remember that about a decade ago I did a series of posts about the art of Jaime Berubé and then collected them into a working paper, Jamie’s Investigations: The Art of a Young Man with Down Syndrome (2016). Two days ago I decided to have Anthropic’s Claude read and comment on that paper. We ended up having a fascinating and productive conversation that yielded new insight. I’ve appended that to this note.

Before we get to that, however, I want to tell you about a little adventure I had last year. I decided to draw some images inspired by Jamie’s dot images. He drew his images freehand on a blank page and used crayons. I drew a grid on my pieces of paper – I did multiple images – and used colored brush pens. Here’s one of my dot images. I have some others in a post, Dot Paintings (after Jamie Bérubé).

Why’d I do it, you ask? Because I thought it would be interesting. It was. And fun. It was. And even instructive. More than I’d thought.

What did I learn? It’s more difficult to do than it looks. While the images appear simple – and they are in a sense, certainly in comparison to, say, a Rembrandt or a Bob Ross – drawing them requires effortful concentration. As I pointed out in my working paper, Jaime didn’t choose his colors randomly, though it may seem that way. When you examine the images carefully you’ll discover that there’s an order to them.

Before I had Claude read my paper I gave it one of the dot images and asked it to comment. Not only did it discover the nature of that order, it picked up on other things I hadn’t noticed and on that basis asked me some questions, some of which I could answer, some I couldn’t. It was an interesting discussion.

Getting back to my own dot images, I discovered that keeping track of my colors was just a little tricky, not rocket-science tricky, but tricky like keeping track of what you’re doing in adding, say, a half dozen multicolumn numbers. So I made mistakes every now at then. Simply dealing with multiple pens, putting one down, picking another one up, that was tricky as well. I made mistakes.

And you know what? Keeping track of all those fiddly little details requires intelligence. Not rocket-science intelligence, not IQ intelligence, but intelligence nonetheless. I can’t think of any other word that characterizes those acts as well as that one does. Claude’s essay says a bit about that near the end.

Moreover it’s the kind of intelligence on which rocket-science intelligence is based. That requires an explicit argument that Claude didn’t make. But you can find a big chunk of that argument in one of my working papers, What Miriam Yevick Saw: The Nature of Intelligence and the Prospects for A.I., A Dialog with Claude 3.5 Sonnet (2025).

Here’s Claude’s essay. It did all the writing, but I inserted the drawings.

* * * * *

What a Sheet of Colored Dots Can Tell Us About Intelligence

Jamie Bérubé has been drawing abstract images since he was eleven. He has Down syndrome. His father, the literary critic Michael Bérubé, put a large selection of the work online, and it is that online archive — not studio visits, not observation of the artist at work — that Benzon's working paper analyzes.

That constraint turns out to matter. Everything below is inferred from finished artifacts. Nobody watched these being made. Which means the analysis has to reason backward from products to processes, and the interesting question becomes how much of a process a product can be made to reveal. Quite a lot, as it happens.

The rule that isn't visible

Start with the simplest of the genres: sheets covered edge to edge with small circular marks in crayon, several hundred to a page, in a great range of colors.

Michael Bérubé, who has seen thousands of these, describes the color choices on most of them as random. They look random. There is no gradient, no symmetry, no regional organization, nothing that reads as a plan.

They are not random. Adjacent marks are almost never the same color. With a palette of roughly ten colors and something over four hundred marks, chance alone would produce well over a hundred violations on a single sheet. There are a handful. The suppression holds vertically and diagonally as well as horizontally, which means the constraint is being applied over a two-dimensional neighborhood, not merely to the previously drawn mark.

So there is a rule: no two adjacent marks the same color. It is entirely local. It requires checking at most four already-placed neighbors. And it produces, without any global plan whatever, the quality that makes the sheets pleasant to look at — a rhythm that is easy to sense and hard to describe.

This is worth pausing on. The rule is invisible to inspection and available to description. The person with by far the most exposure to the corpus reports randomness. Counting reveals otherwise. That gap between what looking gives you and what describing gives you is a recurring feature of this material, and it is one reason the images repay analysis at all.

Reading the process off the page

Other features of the sheets carry process information.

The left margin is well maintained: rows begin at a consistent horizontal position. The right margin is ragged: rows end wherever they end. That is carriage-return behavior — a return sweep to a remembered left position, then progress until the paper runs out. It is what unlined handwriting looks like, and it indicates that the governing spatial rule is sequential and local rather than planned against the page as a whole.

Upper rows are straighter and more evenly spaced than lower ones. Error accumulates downward and is never globally corrected. And the sheets typically stop mid-row, with blank paper remaining. Michael Bérubé supplies the reason: Jamie works until he is interrupted — a meal, an errand — and does not resume a page later. He starts a new one. Each sheet is a session, and a session ends when it ends.

None of this required watching. It is all recoverable from the artifact by someone who knows what to look for, which in this case means someone with extensive practical experience of drawing and of the problem of filling a blank sheet coherently.

How the discovery was made

The most striking material in the working paper is not the mature work but the early work, done between ages eleven and fourteen, which Michael Bérubé has also posted. Those early sheets are miscellanies. Various kinds of object — bars, circles, letterforms, geometric bits — appear on the same page, each placed wherever there was room. There is no relationship between neighbors. There is no composition.

And therefore, as Benzon observes, there was no global order available to be discovered. As long as the elements varied and each sat in its own local space, nothing about their arrangement could become salient.

What changed is that Jamie reduced the vocabulary. One kind of object per sheet, and the simplest possible object: a colored dot. Having done that, the arrangement of objects became the only thing left varying — and thus the only thing left to notice. The discovery of emergent global order was made possible by an act of simplification that preceded it.

This is a documented case of something usually invisible: someone putting himself in a position to make a discovery by changing what he was representing. The intermediate stages survive because the drawings do.

Local procedures and global ones

The mature work falls into several genres, and Benzon's key distinction sorts them.

Local procedures guide each step by reference to something nearby. Dots: place the next one beside the last, keep the spacing even, return to the margin at the end of a row. Towers of color: align the first with the paper's edge, each subsequent one with its predecessor. In both cases a good result follows from following the rule. No planning is required, and attention never has to leave the vicinity of the pen.

Global procedures do not permit that. To center a large circle horizontally on a sheet, one must attend to both edges while drawing — that is, attend away from the point where the line is being made. The concentric-circle sheets require this, and they require it before the first mark. Drawing a circle inside another circle in a single continuous stroke, freehand, with the ends meeting cleanly, is genuinely difficult, and more so for someone with the fine-motor difficulties common in Down syndrome.

Benzon notes what a draughtsman would do instead: tape the sheet down, rule a vertical guideline, set a compass. And this is the important part — with instruments, the tracking is built into the situation. The compass externalizes the constraint that the unaided hand has to hold. A ruled grid does the same for spacing.

Jamie has none of these. The concentrics are hard because the constraint has to be carried internally, and that is presumably why he keeps returning to them.

Roughly a quarter of the concentric sheets are asymmetric — one circle set noticeably smaller than the other. In every asymmetric case the smaller set is near a page edge, which is what one would expect from misjudging the layout and running short of room. Whether that represents failure or deliberate variation is not settled by the images, and Benzon says so.

The interior lines

The most recent genre consists of complex outlined shapes, mostly monochrome, arranged on the page without any underlying grid. They are concave, jointed, limbed, and no two are alike. Michael Bérubé calls them geometrics; Benzon calls them biomorphs, on the grounds that they read as creaturely — an impression shared by the neuroscientist Mark Changizi and by Changizi's wife, an artist, who added that they look like early human art.

Inside almost every one of these shapes there is a single line running from one side, through the interior, to another side. It does not exit the form.

Benzon's reading: by drawing that line, Jamie is abstracting across otherwise dissimilar objects and discovering a property they share. Every one of these forms, however irregular, can be toured from one edge to another without leaving the interior. With respect to that property, all of them are the same.

This deserves a stronger statement than it usually gets. The line is not a depiction. It is a procedure that demonstrates a property by carrying it out. Drawing it is running the test. That is a different kind of mark from anything else in the corpus — a mark that establishes a fact rather than showing an appearance.

And its distribution is informative. Simple shapes do not get interior lines. Jointed letterforms do — T, K, X, Y, and others with junctions. The complex biomorphs all do. Whatever the line is for, it appears exactly where the form is complicated enough that the property in question is not obvious at a glance.

Two modes, and where they cross

That distribution has a name, and it comes from an unexpected place.

In 1975 the mathematician Miriam Yevick, in a paper on holographic logic, described two ways of identifying an object and the tradeoff between them. For a simple object — a square, a cross — a propositional description is short and yields a sharp identification, while a holistic method is wasteful. For a complex object — a Chinese character, a face — the reverse holds, and the propositional description can become more complex than the object it describes. Two modes, opposite dependence on complexity, and therefore a crossover somewhere between.

Jamie's interior lines sit at that crossover. For a triangle or a circle, interior connectedness is settled by looking; the holistic mode suffices. For a jointed concave form it is not settled by looking, and establishing it requires a sequential trace — a path, drawn. The line appears precisely where the glance stops being enough.

Which suggests something checkable. Code every letterform and biomorph in the archive for the presence of an interior line, and independently for shape complexity. If line-presence tracks complexity with a threshold, the crossover has an empirical location — in a natural corpus, without a laboratory.

The gestalt as a reference signal

There is a further point, and it concerns what these two modes are doing at the same time rather than in alternation.

Yevick's framing concerns identification: given an object, which mode identifies it more cheaply. Drawing is not identification. It is control. And in a control architecture — the psychologist William Powers argued the case in Behavior: The Control of Perception — the two modes take different roles rather than competing for the same job. The emerging visual object serves as the reference; the sequence of motions is the means; the difference between them is the error to be corrected.

Those cannot alternate. Comparison requires both present at once.

But notice what the reference actually is on a dots sheet. It cannot be an image of the finished work; nobody holds four hundred color choices in advance. It is something much thinner — a specification of texture, of statistical properties: even spacing, full coverage, no repetition among neighbors. And progress against a texture specification is checkable locally, mark by mark. The whole is never needed.

Better still, the page holds the state. The four neighbors that have to be checked are sitting there in front of the artist. Nothing has to be remembered; it can be re-read. The drawing is its own record.

That is what makes the performance possible. The product is high-entropy and would require an enormous description. The generating procedure is two lines long. The gap between them is bridged by the paper.

Why no teacher is required

One consequence deserves emphasis, because it bears on how skills are acquired generally.

There are domains where you cannot learn by looking at finished examples — you have to produce, badly, and be corrected by someone who already has the skill. Reading finished work supplies no error signal, because finished work is correct and looks fine.

Drawing of this kind escapes that. The criterion is perceptually available in the artifact. A violated color rule is visible. A wavering row is visible. Two circles of unequal size are visible. The reference signal is external, so the error is external too, and the artist can correct himself without anyone to tell him what went wrong.

This is not true of most things. It is emphatically not true of theoretical work, where whether a formulation illuminates anything is not a perceptual property of the page. The distinction — whether the criterion of adequacy shows up in the object — predicts fairly well where self-teaching works and where a teacher is needed.

Jamie has been at this for over a decade, self-taught, in a domain that permits it.

What “intelligent” means here

The working paper carries a footnote acknowledging that some readers will find something jarring in the word “intelligent” applied to an intellectually disabled adult, and insisting on the word anyway.

The insistence is theoretical, not merely courteous. The older clinical vocabulary — idiot, imbecile, moron — encoded intelligence as a single scale with people located as points on it. That picture is still with us in weaker forms, and it is the picture on which the question “how intelligent is Jamie Bérubé” has an answer, and a low one. The drawings do not fit that picture. They exhibit particular mechanisms: a local constraint applied over a two-dimensional neighborhood; a procedure for filling a page without a plan; a reduction of vocabulary that made a compositional discovery possible; an escalation from local to global compositional control; and a graphic operation that establishes a topological property by executing it. Each is specific. None is a quantity.

Intelligence, on the evidence of these sheets, consists of mechanisms — found, refined, and combined over years of investigation, by someone choosing problems because they are hard enough to be interesting. That is what the work shows, and it is not a claim about where anyone falls on a scale.

Based on William Benzon, "Jamie's Investigations: The Art of a Young Man with Down Syndrome," a working paper. Jamie Bérubé's art can be seen at michaelberube.com/jamie.