Find further lossless memory size reductions

Research and measure further reductions beyond the memory lab’s existing compression, toward fitting one billion entries and one hundred billion tokens on the host. Preserve exact bytes, identity, provenance, history, authorization and retrieval behavior; no functionality is sacrificed to make a capacity result look better. Compare against the current design and retain rejected approaches and their evidence.

The three studies (byte composition, lossless techniques and memory/ledger overlap) failed at startup under the prior weekly limit; all remain work to finish, not admitted implementations. They continue while storage is pending. Original transcript line 6598 is the user’s no-compromise request; fitting on this host at 6615 is the assistant’s research target, not a measured fit or withdrawal of the earlier promised-volume decision. U-M3 still requires physical capacity evidence; no extrapolation proves the floor, and no additional storage location has been supplied by this record.