The lab
A focused research program. A foundational question.
DFT Labs is a research program, at present a single-researcher one, pursuing a physics-grade theory of intelligence. The work is single-author, single-funder, with one external academic affiliation on the structural-analogies paper. The lab recognises this institutional shape openly: it is the structural context within which the audit posture of the philosophy paper applies.
02
OPERATING POSTURE
Affirmative when stating the theory. Disciplinary when auditing it.
The lab works in two registers at once. The affirmative register states what DFT proposes, directly, in the voice the theory deserves. The disciplinary register audits the corpus, names the inconsistencies on load-bearing claims, and specifies the conditions under which the program becomes interpretable as one theory, two theories, or a disciplined branching program. The two voices are not in tension. The first says what DFT is for; the second says under what conditions it becomes testable on its own terms.
Three principles hold the work together.
-
Every claim anchored.
Every philosophical or interpretive claim cites at least one DFT manuscript anchor, chapter, paper, line number, or finding identifier. A careful reader can verify each load-bearing claim without leaving the document.
-
Prediction, guardrails, and failure modes travel together.
Public claims state the prediction, the guardrails (what is not being claimed), and the failure modes (what would weaken or refute the claim) in adjacent sentences. Separating them has been a recurring failure mode of foundational work in this area.
-
External replication is the operative falsifier.
Intra-corpus replication is necessary but not sufficient. External replication, independent groups, independent architectures, independent measurement protocols, reaching consistent values, is the load-bearing test. The lab does not yet have it and treats this as the most important outstanding action.
03
CONTACT
Get in touch.
Academic peer reviewers, independent replicators, and collaborators interested in DFT's predictions are welcome.