See:
Platohagel: Reduction of Our Digital Footprint. How we go about being thrifty when we use audio, video, and image files in the production of data as co-creators. How ultimate data savings can reduce Data Centres now proposed for construction all over North America.
Socrates: Yes. I think this can become an important principle within the Builder’s Project / Geometry of Memory: not merely making beautiful digital artifacts, but learning to make them economical artifacts.
The central idea is:
╔════════════════════════════════════════════════════════════╗
CREATE → COMPRESS → STORE → TRANSMIT → REUSE → PRESERVE
╚════════════════════════════════════════════════════════════╝
The question is not whether digital information consumes resources. It does. The more useful question is:
How much information is actually necessary to preserve the meaning of what we have created?
The larger problem
The scale matters. The IEA reports that data centres consumed about 415 TWh of electricity globally in 2024, approximately 1.5% of world electricity consumption, and projects that consumption could more than double to about 945 TWh by 2030. In the United States, data centres are projected to account for roughly half of electricity-demand growth through 2030.
And this is not simply a question of “AI using too much electricity.” Digital storage, networking, cloud services, streaming, computation, cooling and the continual replication of information all participate in the larger system.
The particularly interesting point for our work is that we are also producers of the information entering that system.
Every photograph, WAV file, MP4, duplicate export, thumbnail, intermediate render, backup and unnecessarily large image becomes part of that accumulating digital material.
A different definition of digital thrift
We could establish a simple rule:
╔════════════════════════════════════════════════════════════╗
Do not store more data than the meaning requires.
╚════════════════════════════════════════════════════════════╝
That does not mean destroying information indiscriminately.
It means distinguishing:
Master → Working file → Distribution file → Archive
For example, with one of your cinematic songs:

This is precisely why the video-compression work we have been doing becomes more significant than simply “making a video smaller.”
We are developing a repeatable conservation process.
Audio
The same principle applies to your music.
A WAV master may be valuable because it preserves the original recording. But a website generally does not need every listener to receive the archival master.
So:
WAV master → compressed listening version → web player
The master preserves possibility.
The distribution file preserves experience.
Those are different functions.
Images
Likewise:
RAW/TIFF/large PNG → archival master
then:
JPEG/WebP → web presentation
An image that is 6000 pixels wide may contain information that is irrelevant when displayed at 1200 pixels.
The surplus is not necessarily “bad”; it is simply unnecessary at that particular layer.
Video
Video is where this becomes especially powerful.
A six-minute video can contain enormous amounts of redundant information because successive frames frequently differ only slightly.
Modern codecs exploit that temporal relationship.
Thus:
6:07 of video remains 6:07 of video.
Compression does not shorten the song.
It changes the quantity of data required to represent those six minutes.
That distinction is fundamental.
And then comes the deeper idea
If millions of creators begin thinking this way, the question changes from:
“How much additional data can the world store?”
to:
“How little data can we use while preserving the information we actually value?”
That is a very different technological philosophy.
The IEA’s latest analysis shows why this matters now rather than sometime in the distant future: global data-centre electricity consumption increased about 17% in 2025, while AI-oriented data centres grew even faster.
There is also an important caution. Efficiency alone does not guarantee declining total resource consumption. Berkeley Lab’s 2025 update found that improvements in computing efficiency have been outweighed by the scale and growth of computational demand, producing continued increases in absolute U.S. data-centre electricity consumption.
So the equation cannot simply be:
better technology = less energy.
It has to become:
╔════════════════════════════════════════════════════════════╗
EFFICIENCY + RESTRAINT + REUSE + SELECTIVE PRESERVATION
╚════════════════════════════════════════════════════════════╝
This could become an Eskesthai principle
I would call the concept:
The Principle of Digital Sufficiency
Preserve the original where originality matters.
Compress where experience does not require the original.
Eliminate duplication where duplication carries no additional meaning.
Reuse existing data before creating another copy.
That fits remarkably well with Geometry of Memory.
Memory itself does not require every possible detail to be retained at maximum resolution. A structure can preserve relationships while discarding redundancy.
So we arrive at another useful conceptual equation:
╔════════════════════════════════════════════════════════════╗
DATA → REDUNDANCY → COMPRESSION → STRUCTURE → MEMORY
╚════════════════════════════════════════════════════════════╝
And perhaps the deepest formulation is:
╔════════════════════════════════════════════════════════════╗
LESS DATA ≠ LESS MEANING
LESS REDUNDANCY → MORE EFFICIENT MEMORY
╚════════════════════════════════════════════════════════════╝
This could become more than a personal housekeeping rule for Eskesthai. It could be a co-creator ethic for AI-assisted digital production: when Platohagel and an AI create an image, song or film, the responsibility is not finished when the artifact is beautiful. We should also ask what is the smallest faithful representation that can carry that artifact forward in time?
That is where digital creativity meets digital conservation.




