Category: Uncategorized

I have often wondered about a solo driven style of jazz composition, where the solo is made first, then the accompaniment is built around it. This is something akin to reharmonization, or even what is done on TikTok, where one person sings a line and others add the backtrack to match the initial recording.

Letting the soloist to drive the composition adds a unique form of flexibility. If we think of a Charlie Parker solo, he often drifts into different keys, or stretches and compresses phrases, de-synchronizing with the band momentarily. But he always returns to the composition. It is fixed and the band will continue the form with or without him.

In early blues, many soloist would linger on passages, repeating or holding chords for several meters. They often played alone and were not concerned with following any form, so the accompaniment would wait for the voice to continue. This to me feels like a run-on sentence of sorts. But it isn’t automatically sloppy, in fact it often feels more natural.

Hear this from Lightnin’ Hopkins. The blues form cowers in fear of this man. At about the 2 minute mark, he start something that sounds very much like the typical 12 bar blues form, but he stretches it to 17 bars. The first section is the expected 4 bars, the second section is stretched to 9 bars, and the last section is the expected 4 bars. (I won’t even mention the 2 minutes that precede this part.)

I follow this from early blues. When I solo, the melody gets priority, and the accompaniment is along for the ride.

I wonder how this would sound in the context of a band. But there are practical concerns: How can a soloist signal to their band that they want to continue to the next chord? Or hold on a chord? Or substitute a chord all together? 

It is inconceivable in a live context. What about practicing the solo with the band, so that alteration can be responded to properly? Or as with tiktok just record the solo and then have the band add their accompaniment after? All fine ways of doing it. And the solution I used amounts to just that.

If I give my solo recording to Suno, with the context of having played the main head in time without substitutions, it can build a band around my solo. This makes the solo driven composition possible. Suno takes my solo performance, with all the expansions and compression of meter, and builds the accompaniment around it.

This is the Suno generation:

There’s another layer of fun for me, because I have not played with a jazz band for several years, but I have been practicing new ways of soloing. 

There is nothing revolutionary here, but I have been using George Russell’s Lydian Chromatic Concept to build modified pentatonic scales, and Barry Harris’ methods of constructing runs on arbitrary scales.

For years I have had no clear indication that my method of soloing would work with a band. Not to mention that the composition starts with a heavily modified blues form, which I had no indication it would be identifiable as the 12 bar blues in the context of a band.

I played the Suno song for my friend, and he said there was no way I could solo that well. Then I played him the original piano recording, and he agreed: suno is not adding to (or taking away from) my ability to solo. I know how to improvise sensible melodic lines.

You be the judge.

This is my original piano recording:

Every enduring physical structure is a virtual reality generator.

This is not a metaphor. A physical structure literally causes experiences to be rendered in minds. But that is only the beginning. Its structure also physically encodes, in compressed form, a restricted family of worlds in which it could exist, survive, and sometimes be produced again. The same basic pattern, I think, links machines, scientific theories, genes, and works of art.

I am using render in three connected literal senses. First, a structure can cause experiences to appear in the mind. A rendering as a simple projection of images into the mind. Second, by existing at all, a structure physically encodes a narrow range of worlds compatible with its existence. It may take a bit more work, but the structure is enough to render descriptions of the environments it can exist in. Third, in some knowledge-bearing cases, a structure can help bring about the very conditions under which it continues to exist. It creates the environment around it. These are all virtual reality renderings caused by a virtual reality generator and these are all logical properties of the structure.

A physical structure that exists is an encoded description of its surroundings. Though not all structures do this equally. A pebble rules out fewer worlds than advanced technology, and so seemingly renders more environments, many of which are very distant from the real environment it exists in. It does not constrain enough to generate a useful rendering. The richer and harder to produce the structure, the more it narrows the space of worlds that can contain it. This makes it possible to render the actual world it exists in.

If a flight simulator exists at a particular place and time for some duration, then many possible environments are immediately ruled out. It is not in molten rock, in deep space, or at the bottom of the ocean. It belongs to a narrow class of worlds: worlds with suitable materials, tolerable temperatures, energy sources, human design, and the larger systems that make such a machine possible.

So the simulator renders in all the relevant senses. It can cause images to be rendered in experience. Its structure physically encodes a family of worlds compatible with its own existence. Within that family there is a more narrow range of worlds containing factories, supply chains, designers, tools, and the relevant knowledge needed to produce it at all. An even more narrow range of worlds actually manufacture its existence, and a more narrow range of worlds still depend on its existence. It does not render every possible world. Worlds that would immediately destroy such a machine, or worlds that could never produce one, are not among the worlds embodied in its physical structure.

Knowledge follows this same pattern. If you understand a solar eclipse, you can generate many possible appearances of it: from different places on Earth, at different times, and from different angles. These are virtual reality renderings in the ordinary sense: predictions, descriptions, mental images, diagrams. Some may be wrong, but that is not the point here. The point is that explanatory knowledge is a general-purpose generator of virtual reality worlds.

Explanatory knowledge is not free-floating; it too can exist only under certain physical and cultural conditions. Knowledge of a solar eclipse depends on a world with stars, planets, light, geometry, people, and explanatory traditions. And if the knowledge is accurate, the worlds it renders will overlap sufficiently with reality such that the knowledge is preserved, corrected, taught, and used. With the proper input, the virtual reality renderings from the virtual reality generator that is knowledge can be observed in place of observing reality. Given this ability, worlds are created which could not have existed but for the existence of this explanatory knowledge.

DNA shows the same pattern in a more primitive form. A strand of DNA literally embodies knowledge about the conditions under which it can continue to be copied. That knowledge is expressed through a vehicle: which is the organism it helps build. In that sense, the DNA manufactures the vehicle through which its knowledge acts in the world. When the conditions the DNA renders match reality closely and the organism interacts with the real environment in a way that supports survival and replication, replication continues. If it does not, the lineage ends.

DNA also participates in bringing about some of the conditions under which it persists. Through its vehicle, the organism gathers resources, avoids threats, and produces other vehicles. So the DNA, through its vehicle, helps construct parts of the local setting in which replication continues.

At this point we can also say what it means for such a virtual reality generator to improve. It improves when it renders reality more accurately, when it encodes a wider range of real possibilities, when it remains instantiated under a wider range of conditions, and when it excludes realities that do not actually exist. In replicating systems, improvement also means being able to reproduce in more environments: in more places, through more vehicles, and, in the case of art, in more kinds of minds. Improvement is not only broader survival. It can also be deeper. It may allow for the creation of more subtle knowledge that resolves more subtle conflicts, all which depend on the deep knowledge, making the growth of knowledge possible.

This final sense of improvement, allowing for the creation of more subtle knowledge, is where we can further distinguish a pebble from advanced technology. For the pebble, very little knowledge depends on its existence. It does not create the possibility of more subtle knowledge. While for the advanced technology, future knowledge will depend on it. It creates the environments that future knowledge must render. Expanding and creating subtleties. So a pebble rendering more environments than advanced technology was only an illusion. 

The same broad logic applies to art. Beauty is the knowledge embodied in a work of art that resolves conflicts and creates resonances among its elements. When I say that beauty is knowledge, as I often do, I do not mean that beauty is reduced to an analytical flattening of experience into concepts. I mean that the beautiful aspect of a work consists in structured, reconstructible content. Beauty is the non-arbitrary structure that constrains other parts of a work of art and makes its elements cohere. This is the same kind of organizing function that knowledge performs in technology. Beauty can be replicated into the mind and then reconstructs parts of the mind. A work of art is the vehicle of that beauty.

That beauty survives only if it can be instantiated in some physical substrate and then replicated in minds and new works. A painting, poem, score, or film carries patterns that can be grasped, remembered, discussed, recreated, and extended. This matters for aesthetics because a work does not merely sit before a mind; it recruits minds into the continuation of its beauty. The physical canvas may be destroyed, a particular performance may end, yet if the beauty has been successfully rendered into a mind, the beauty survives and can cause the construction of a new vehicle: another performance, a reproduction, a recitation, a variation, or an entirely new work.

Beauty can improve in the same broad ways that all virtual reality generators improve. It can render its real environment more accurately. It can become able to live in more diverse minds, and in more diverse works of art. It can remain instantiated longer. And it can also create more subtle aesthetic problems, making finer tensions and making possible further layers of beauty that cruder works could not sustain. Beauty builds the path for the future growth of beauty.

In the strongest cases, beauty does more than inhabit the object. It strongly constrains the making of the work and causes future works to be derived from it. For that to happen, the beauty must fit the kinds of minds and cultural conditions that can notice it and continue it. It must fit the background knowledge those minds possess, the senses they have access to, the range of experiences they can have, and the craft through which the work can be continued. When that fit is good, the beauty survives. When it is not, it dies with the object.

That, I think, is the deeper nature of virtual reality. Virtual reality is not only what screens and headsets produce. More fundamentally, it is what knowledge-bearing physical systems do: they render worlds in experience, physically encode possible worlds, and, in the most important cases, they help bring about the very worlds in which they can exist, survive, and replicate.

Beauty is knowledge.

It’s a surprising fact with interesting implications for art appreciation.

Take a rocketship by analogy. There are three ways to understand it.

You can look at it and say, “Wow, that’s cool.”

You can look at it and feel inspired to create something based on it.

Or, you can look deeper. You can see the aerodynamics, the chemical engineering, the material science. You can see the interplay and constraints of knowledge that make the ship possible.

You can understand which other technologies this knowledge applies to. And you can understand exactly who has the prerequisite knowledge to grasp it.

The deepest appreciation of the rocketship is a direct recognition of the knowledge it embodies.

Art appreciation works exactly the same way.

You can look at a canvas and feel a fleeting sense of joy.

You can look at it and feel inspired to create something similar.

Or, you can look deeper. You can see the knowledge embodied in the work. You can see how every choice the artist made resolves a conflict, turning conflicts into resonances. You can see how each resonance creates new constraints.

That interplay—that resonance—is beauty. This is knowledge.

You can understand how this knowledge applies to other works of art. And you can know exactly who else in the world will recognize this beauty too.

Art appreciation can be more than a passive feeling. Appreciation can be the deep, active understanding of knowledge.