Research — the QIG Programme
An independent physics and AI research programme exploring how gravity-like structure can emerge from the geometry of quantum information.
QIG (Quantum Information Gravity) is Braden Lang's independent research programme investigating how gravity-like structure can emerge from the geometry of quantum information. [matrix §8b-1]
The research runs on ordinary classical computing infrastructure — CPU/GPU and tensor-network methods — with a formal compute-budget discipline: planned work is turned down or delayed, not run over budget, when it would exceed pre-set cost caps. [matrix §8b-8]
This page describes a research programme, not a product. It does not state what the programme is worth, does not name an amount being sought from investors, does not claim customers or revenue, does not claim exclusive legal rights over the work, and does not claim that any outside scientific group has reviewed or backed it — none of these exist in the programme's own record as of this page's last review. Where the programme's own results are cited, the citation points at the internal evidence matrix behind this page, not at a public dataset.
Evidence & Method
How the programme measures its own claims, and what has been checked against results published outside the programme.
Method
The programme develops and internally certifies its own measurement methods before trusting them, including a formal instrument-certification process and an internal doctrine that requires every measurement to be tagged by what was perturbed, how, where it was measured, and against which clock. [matrix §8b-2 (Four-Axis Doctrine)]
The programme's internal claims are organised into four honesty categories — established, bounded, still-open, and formally retired — with all four categories, including the retracted ones, kept on permanent record. [matrix §8b-11]
The programme's terminology and measurement doctrine are documented in versioned, dated internal specifications, updated as understanding improves. [matrix §8b-12]
External checks
The programme has independently reproduced known results from published physics literature as an external check on its methods, including a mass-ratio prediction from Coldea & Zamolodchikov and a topological-invariant match to a result published by Roushan et al. in Nature. [matrix §8b-3, §2.2 rows 12 & 57]
The programme maintains its own record of methodological errors and retracted claims as a discipline against repeating them — 31 catalogued failures and 16 formally retired claims as of August 2026. [matrix §8b-4]
Quantum hardware — a verification anchor, not the method
The programme's position on quantum-hardware (QPU) computing is deliberately conservative: real quantum hardware is used only as an occasional external verification check, never as the primary calculation method, because classical tensor-network simulation is currently both cheaper and more reliable for the programme's models. [matrix §8b-5]
A real quantum-hardware experiment run by the programme in July 2026 directly led to a same-day improvement in the team's local software, used to sanity-check future results before they are trusted. [matrix §8b-7]
The programme's first public paper (October 2025) was followed one month later by the programme's own formal erratum revising its headline number — the group holding its own early work to account rather than letting an unconfirmed result stand. [matrix §8b-6]
Technical detail — terminology
- Basin
- A point in a high-dimensional probability-distribution space representing a system’s internal state — not a simple vector, but a point on a curved geometric surface.
- Fisher-Rao geometry
- A mathematical way of measuring "distance" between probability distributions. It is the geometric substrate the programme uses for both its physics work and its AI self-modelling metrics.
- Camera (instrument certification)
- An internal term for a cheap computational shortcut whose accuracy has been checked and confirmed against a slower reference calculation before the programme trusts it.
- Frozen fact
- A claim whose measurement the programme has locked in and will not casually change. "Frozen" describes the measurement, not necessarily the final interpretation.
- Retired claim
- A claim the programme has formally withdrawn from its own canon, on the record, with a note on what — if anything — survives the withdrawal.
Potential Applications
Projections — not results
None of the following exist as a shipped product or a demonstrated service today. Each depends on further work the programme has not yet completed.
Geometric diagnostic tooling for quantum many-body simulation — a Fisher-Rao-based analysis method that could, in principle, be packaged as a library for other researchers running lattice simulations. Not yet demonstrated outside this programme's own test suite. [matrix §4-1]
A geometry-grounded architecture for AI self-modelling and interpretability metrics, using basin dynamics and Fisher-Rao distance as a principled alternative to ad hoc interpretability heuristics. Has not yet reached the programme's own internal definition of done. [matrix §4-2, §8b-9]
A quantum-hardware-verification-as-a-service niche, generalising the programme's 'hardware as a cheap verification anchor, not primary compute' doctrine. Currently a single internal case study, not a demonstrated service. [matrix §4-3]
Novel classical algorithms motivated by quantum-hardware comparison work — the general pattern of a quantum-hardware experiment teaching the team something usable on classical hardware. Only one dated instance currently exists as evidence. [matrix §4-4, §2.3]
Roadmap & Milestones
The programme deliberately does not run calendar-style roadmapping — milestones are dated only once they are achieved.
Achieved
Achieved — 27 May 2026: programme master briefing and terminology manifest issued. [matrix §5 achieved, 2026-05-27]
Achieved — 11 June 2026: the Four-Axis measurement doctrine adopted and frozen. [matrix §5 achieved, 2026-06-11]
Achieved — 3 August 2026: the internal results ledger reached its latest version. The next major queued result is recorded as blocked on compute budget, not merely open. [matrix §5 achieved, 2026-08-03]
Status snapshot — 9 September 2026
At the 9 September 2026 evidence review, the QIG research programme was paused pending grant funding, after roughly a year of active, internally-documented research. [matrix §8b-10, §5 planned]
Research & Investment Enquiries
For research partnership or investment enquiries about the QIG programme, select "Research / investment" below and tell us what you'd like to know.