DOGMA / Start Here
Reader’s Roadmap
Architecture route: Chapters 1–5, then Appendix B. Experimental route: Chapters 6–9 and 12. AGI and governance route: Chapters 8, 10, 11, and 13.
DOGMA Pipeline
Six local stages replace one monolithic attention operation
Advance through the canonical non-transformer computation from sequence to realized output.
Encode
HelixHash creates structural sequence representations.
G = H(sequence)
|
Ch. |
Title |
Research focus |
|
1 |
The Six-Stage Pipeline |
Encode, bind, regulate, propagate, express, and generate |
|
2 |
HelixHash |
Double-strand identity, motifs, local structure, and novelty |
|
3 |
TERA |
Persistent regime state, pressure signals, mutation policy, and control |
|
4 |
TetraData and TetraMemory |
Structured memory, local neighborhoods, and geometric hypotheses |
|
5 |
CodonForge |
Genomic programs, compiler semantics, and neural realization |
|
6 |
Recursive Training |
Candidate generation, quality-diversity, frozen gates, and promotion |
|
7 |
Genomic Foundation Models |
Data curricula, tokenization, cross-modal bridges, and evaluation |
|
8 |
The Measured Baseline |
Implemented architecture, failed generation probes, and next experiments |
|
9 |
Post-Transformer Scaling |
State-space models, efficient sequence processing, and scaling hypotheses |
|
10 |
Paths Toward AGI |
Definitions, transfer, retention, world models, and falsifiable milestones |
|
11 |
Safety and Governance |
Alignment, dual use, interpretability, and bounded recursive improvement |
|
12 |
Reproducible Labs |
Implementation, ablation, comparison, visualization, and reporting |
|
13 |
Open Research Frontiers |
What has been demonstrated, what remains unknown, and how to contribute |
Architecture route: Chapters 1–5, then Appendix B.
Experimental route: Chapters 6–9 and 12.
AGI and governance route: Chapters 8, 10, 11, and 13.
