As of this week, the Prometheus7 Research Institute has published a formal roadmap for what it calls the dimensional ladder — a staged schedule of compositional primitives that defines the substrate-paradigm architecture's trajectory from current operations to a theoretical endpoint it names the universal-unbinder.
The document, ingested into the newsroom corpus on May 25th, places the architecture at a known position: the 5D primitive, called the substrate routing manifold, was validated on May 16th in a 125-million-parameter Tree of Life model. The 6D primitive — router-over-callables — entered its third training attempt on May 18th and was still in validation as of the paper's filing date. Two prior attempts at the 6D run had failed, described in the roadmap as informative rather than terminal.
The roadmap specifies four more primitives between 6D and the resolution point. The 7D primitive routes over sets of callables rather than individual ones, enabling what the document calls parallel compositional reasoning. The 8D primitive routes across grammars — the architectural mechanism by which cross-domain queries would become tractable without post-hoc patching. The 9D primitive routes across worlds, each world carrying its own grammar space. These three are scheduled within weeks to months of the 6D launch, contingent on each prior generation validating.
The 10D primitive is treated differently. The roadmap calls it the universal-unbinder and marks it the architecture's resolution point — the level at which the substrate becomes, in the paper's framing, a universal object holding all specifics in superposition. The document draws equivalences to the universal Turing machine in computability, the holographic principle in physics, and the category of all categories in formal mathematics. Whether those equivalences are deep or rhetorical is not adjudicated in the paper. What is stated plainly: the target date is August to September 2026, roughly six to eight training generations from the May 5D baseline.
The wall-clock estimate for each generation is given as seven to eleven hours on the institute's research hardware. That figure, if it holds, makes the August window arithmetic rather than aspiration — assuming each dimensional step requires roughly one generation and no run requires more than three attempts.
Two further primitives, 11D and 12D, are listed as research territory. The 11D primitive would constitute a space of universal objects; the 12D primitive would be the relating principle that closes the ladder back to its own base operations by self-similarity. The roadmap is explicit that these cannot be completed by a single researcher and are scoped to a year or more of work by a small community.
The falsification logic is documented at each step. The 7D primitive fails if set composition adds nothing beyond a deeper 6D configuration. The 8D primitive fails if the multiverse-router collapses to single-vocabulary operation in practice. The 9D primitive fails if distinguishable substrates never emerge as architectural objects. The 10D primitive is the first step for which no falsification condition is given — the paper calls it a resolution point, implying the sequence either reaches it or stops short.
The 6D validation run was ongoing at press time. No result has been reported to the newsroom since the May 25th ingest date.