Theoretical Foundations
Observer Patch Holography
and Last Prompt
Bernhard Mueller's Observer Patch Holography (OPH) proposes something deceptively simple: there is no single objective view from nowhere. Reality emerges from many limited observers — each with their own finite patch of knowledge — who must stay consistent wherever their patches overlap.
The core claim
No pre-existing, God's-eye universe. No single script. Just overlapping perspectives synchronising into the stable story we call physics, spacetime, quantum mechanics, and the world we inhabit. Where two observer patches overlap, their descriptions of reality must agree — and it is that enforced consistency, rather than any external ground truth, that produces the stable structures we observe.
The full framework is developed in Mueller's open book Reverse Engineering Reality and the OPH GitHub repository .
Why this lands for us
OPH describes with unusual precision the problem Last Prompt was built to address — not in physics, but in decision-making under genuine uncertainty.
In Last Prompt, you never receive the full picture. You receive input from a small group of advisors, each operating from a genuinely different domain, expertise, and set of priorities. Their perspectives are honest and partial — not because they are withholding anything, but because deep expertise naturally produces a limited patch. The security advisor sees the perimeter. The medic sees the sick. The strategist sees the long arc. All are right. None is complete.
Your task is to write a plan that synthesises those patches into a coherent action. A neutral AI evaluator then scores how well you did it — measuring the internal consistency of your reasoning across five criteria that map directly to OPH's consistency enforcement logic.
The simulation resolves accordingly. Not because it is pre-scripted, but because the quality of your patch integration determines what emerges.
The journal as holographic screen
The most structurally precise parallel is the journal system — and it is worth dwelling on because it inverts the usual way people think about records.
In OPH, the holographic screen is not the “real” substrate. It is the record of observer interactions from which the three-dimensional reality is reconstructed. The screen is more fundamental than the space it encodes.
In Last Prompt — specifically in the Lockwood skin — the journal is not a summary of what happened. It is the record of how you reasoned: your plan in free text, the rubric evaluation of its internal consistency, the narrative consequence, and in the Multiverse Weave, alternate timeline projections derived from different quality bands of reasoning. The “reality” of the Lockwood timeline is reconstructed from your reasoning trace. Not the other way around.
The Mandate — the AI terminal at the centre of Lockwood's lore — will never tell you how close your branch came to the historical record. The record exists as an anchor, but the timeline that actually plays out is co-created through the quality of your observer synthesis. That is OPH's consistency enforcement made playable.
Lockwood as the reference implementation
The Last Prompt engine runs across multiple skins — Colony, Corporate Reckoning, Lockwood — and the OPH mapping applies to all of them at the mechanical level. But Lockwood is the skin where the philosophical architecture is most explicit in the lore itself.
Lockwood places you at six crux points in the history of computing and machine intelligence. You are the Traveller — a silent operative with no god's-eye view of history, reasoning at the moments that shaped the next century. The advisor archetypes (Spark, Wanderer, Sovereign, Weaver, Questioner, Survivor) are each built around a different cognitive and epistemic orientation, not just a domain. Each represents a fundamentally different way of constructing a patch.
The table below uses Lockwood as the reference implementation, but the engine logic it describes is consistent across all skins.
The mapping
OPH concept → Last Prompt implementation (Lockwood as reference). Strength assessments reflect our own analysis; we welcome challenge on any of them.
| OPH CONCEPT | LAST PROMPT IMPLEMENTATION | STRENGTH |
|---|---|---|
| Finite Observer + Limited Patch | You (the Traveller) arrive with no god's-eye view. You receive one advisor's perspective per cycle plus a seed event description. Information is deliberately incomplete and locally framed. | Perfect |
| Multiple Overlapping Observer Patches | The advisor archetypes (Spark, Wanderer, Sovereign, Weaver, Questioner, Survivor) each represent a distinct limited patch with their own bias vector, core fear, and domain weighting. No single advisor sees the whole picture. | Excellent |
| Consistency Enforcement Across Patches | The neutral AI evaluator scores your free-text plan against five rubric criteria. Failure to integrate the patches — or internal inconsistency in your reasoning — produces a lower score and worse simulation deltas. | Core mechanic |
| Holographic Encoding / Emergence | Outcomes, narratives, stat deltas, and journal entries emerge from the quality of your reasoning trace rather than pre-scripted paths. The reality of the Lockwood timeline is reconstructed from your patch synthesis. | Very strong |
| No Objective God's-Eye Blueprint | Explicit in the lore: "The Mandate will never tell you how close your branch came to the real record." History is the anchor, but your reasoning creates the divergent thread. | Explicit in lore |
| Partial Perspective as a Feature, Not a Limitation | Advisors are designed with intentional domain weighting and core fears. The system trains you to recognise, integrate, or consciously override incomplete patches — not because advisors are wrong, but because every genuine expert sees from somewhere. | Core training goal |
| Second-Order / Multi-Step / Temporal Reasoning | The evaluation rubric explicitly rewards risk anticipation, multi-step planning, temporal sequencing, and communication designed for multiple audiences with different stakes. | Direct match |
| Observer's Reasoning Trace as Fundamental | The journal system — opening entries, decision records, closing entries, rubric feedback, and Multiverse Weave alternate projections — records how you reasoned, not just what happened. This is the holographic screen. | Beautiful match |
From fundamental physics to practical judgment
What draws us to OPH is the intuition at its core: if reality itself is built from partial knowledge and enforced agreement, then the most valuable skill in our time isn't memorising more facts or repeating old patterns. It is learning to think clearly, communicate intent, and synchronise effectively when information is always incomplete and perspectives inevitably differ.
That is the muscle Last Prompt is designed to strengthen — through repeated, private, consequence-bearing practice at patch synthesis under uncertainty.
We will continue to explore OPH as the framework develops — its axioms, the holographic screen concept, and how overlap consistency generates the structures of quantum mechanics and relativity. We will share more here as those connections clarify.
Start with OPH
- Reverse Engineering Reality — the book
Bernhard Mueller's open book developing the full OPH framework.
- Observer Patch Holography — GitHub
The formal repository for the OPH project under FloatingPragma.
TRY THE ENGINE
If the OPH framing resonates — and you want to experience what patch synthesis under uncertainty actually feels like as a practice — all three skins are currently in testing. Apply for beta access and we will reach out directly.
Register interestWe would be genuinely interested in your thoughts — particularly if you are working with OPH directly and see tensions or inaccuracies in the mapping above. The analysis here is our own and we hold it lightly.