Stories from an open-ended world

GoS
4 min readJun 6, 2020

--

— Take a wooden stick. Imagine it wider, lying horizontally on top of four other sticks. You now have something resembling a table. Take a few more planks, start laying them horizontally, then vertically.. and after a several more steps you’ll have a house.

A glimmering image of a wooden house was shifting over the projector, reflecting the campfire.

— Make a machine add, remove or change this or that part, even at random, run it long enough, and you’ll get all kinds of things — those that were, that are and that might be.

The image changed into a cloud, expanding outwards, until it was all around them, echoing the starlit sky.

— This is the space of all the possible ideas, and now it’s your time to learn how to navigate it.

Wind was swirling the dust of the barren moon — a perfect view for the Kohr’s new novel. Having already mapped out the main plot lines it was now time to start writing. The first paragraph was shifting in the air as Kohr was exploring its tone — having prearranged the search space using his recent musical inspirations as cues he was now making slight variations to how depressive it felt.

Many authors were content with just using the auto-suggest which filled the whole novel in their usual style, but Kohr aspired to be one the Greats — those acclaimed by the readers for constantly evolving their styles into something new, and that was not possible without the author evolving as well.

Seeing an image of herself move in a dance-like fashion while she moved her fingers was a funny experience. All of the Shai’s engineering peers were working in analytical ways: to find a new kind on engine they started some known ones and iterated on devising the mathematical measures to describe the feasible directions in idea space into which they could explore further. It wasn’t the way Shai’s thoughts worked though.

Exploring the technical ideas, she felt where she wanted to move next, as if the energy flow through the engine was the tension of muscles along her body, and by moving it new ways to normalize both excess tension and excess energy flow could be found.

She searched for a way to move through the idea space the way she moved through an imaginary physical one. She wanted to be like water.

After finishing a book, you may see everything through its lens for some time, before normalizing this new perspective into just one of many you have. When finding an idea in the idea space though, there’s no ending to it — you can dig deeper and deeper into its many developments, till the division between your old worldview and your new one grows too large to be bridged.

This story was all too familiar to Yii — for centuries they’ve observed the exploration of the idea space. There was a large central cluster of activity — different communities exploring different but related things, — and there were slowly orbiting outliers at bleeding edge of avant-garde art or a technological frontier. But from time to time, a small speck of activity would fly off, moving farther and farther over the decades, and then disappear.

Among those specks Yii’ve seen de-evolutionary cults transforming their members into simple and simpler life forms until all that was associated with their names was just a bunch of simple particles, … , . And sometimes they wandered how many other Yiis there are, observing far-away parts of the idea space being explored in ways they could not comprehend.

Notes

The ideas of creating an algorithm that can generate new ideas endlessly — called open-ended evolution — is real. The main question in actually achieving it is how to do it efficiently enough —the universe will end before an algorithm generating bunches of atoms at random will find anything interesting.

In a more realistic scenario many things we know might remain in the world:

  • different companies will run generators for different part of the idea space and take fees for each idea you’ll want to use — as they need to pay for the generating computation somehow
  • you’ll want to get some credit for the ideas you find in the vast generated data, so the some kind of intellectual property rights will need to exist
  • tools that let you navigate the idea space are likely to be produced in both open-source and proprietary forms

Also, in this scenario the bottleneck to the speed of exploration is at the similar place we have it now — at the interface between the digital computation and a brain with organic senses. The solutions with mind uploading or augmentation might work, but there’s a possible solution that bothers me: “who’s better to know what i want then me?” would you ask, and create search algorithm that borders with a White Christmas-style copy of yourself to do the exploration and show you the most interesting results. But then, what’ll be the point of you existing at all?

--

--