Had an Idea this morning.
Just jotting it down so I can call prior art when someone try’s to patent it in the future 😛
Data compression comes in two kinds, lossless and lossy, png vs jpeg…
I propose a new kind of compression, Experiential Compression, which is technically a lossy format but perfect at the same time.
Step 1, Expose brain to the stimulus which is to be compressed.
Step 2, Record nerve impulses at some point in the neural stack. (Deeper is probably better for compression, but more difficult to achieve… Also more likely to fail to replay correctly due to brain elasticity)
Step 3, (Optional) Compress nerve impulse recordings with lossless compression.
Decompression requires nerve impulses to be replayed direct into the brain. Kind of a problem right now but maybe not in future.
A more ambitious option would be to simulate the neural stack to determine the most compressible inputs which generate the same nerve signals. Sounds seriously tricky…
Simulating the neural stack is also an alternative to using an actual brain for steps 1 and 2.
I don’t know enough neuroscience to be sure but I am guessing that this compression would be individualised, data recorded this way would only be good for the original brain. But in a neural integration scenario, it could reduce the storage required in neural implants or bandwidth required to talk to external storage for individual playback at a later date.
Okay, crazy idea all typed up, back to work.