sovsrc.llm updates
sovsrc.llm is a specification for an LLM that uses source material with permission, gives credit to the people who made it, and plans rewards for creators.
The project is at specification stage. No application, trained model, or payment system exists. This page gets a new entry for each commit to the repository. A machine writes the summaries from the change, so read the source before you quote a summary.
18 Oct 2026
-
Public reading list now points to the Super Anthropocene article on sovsrc.ai
The public reading list now directs readers to The Super Anthropocene and the Value of Human Contribution on sovsrc.ai.
-
New Super Anthropocene article examines proposed contributor rewards
A new opinion article proposes voluntary contributor rewards that connect AI revenue with continuing knowledge creation. It describes an operating rewards platform as an ambition.
17 Oct 2026
-
New analysis examines Harness2 for proposed sovereign AI development
A new article examines Harness2 as a possible component for proposed sovereign AI workflows. It describes integration possibilities, licensing considerations and evaluation limits.
16 Oct 2026
-
New article examines Claude and OpenAI terms for proposed AI development
A new public article examines how Claude and OpenAI terms could affect proposed competing-model development. It distinguishes ordinary application engineering, competing-model code, and generated training material.
-
Diagram source now uses a Mermaid initialization directive
The diagram source now places its display settings in a Mermaid initialization directive instead of a separate configuration header.
-
New article examines proposed model development with Nemotron 3 Ultra
A new article and diagram describe proposed routes to a competing model using Nemotron 3 Ultra. The article separates model permissions, training origins and proposed contributor rewards.
15 Oct 2026
-
Sovereign Source removes detailed proposal documents from its public collection
The repository removes documents covering proposed implementation requirements, contributor reward standards and business plans. These deletions concern proposals, not an operating model or payment service.
-
New comparison maps proposed requirements to patent claims
A new document compares proposed Sovereign Source requirements with twelve patent documents. It identifies licensed retrieval, contributor rewards and provider reporting as the closest functional overlap.
-
Patent research compares claims relevant to the proposed platform
A new research document compares twelve US patents and published applications relevant to proposed attribution, contributor payments, licensed generation and marketplaces.
-
New draft proposes shared records for LLM rights and contributor awards
The new LLM Rights and Content Awards draft proposes shared records for source permissions, attribution, provenance and contributor benefits. The project remains at specification stage.
14 Oct 2026
-
Documentation gains a searchable library and licensed-content billing research
The repository adds a searchable documentation library and research proposing separate compute and content charges. The library identifies the platform as a proposal.
-
New research compares proposed distillation and training data options
Two new reports assess proposed routes for portable model distillation and permissioned training data. They compare providers, source permissions, deployment control and contributor compensation.
-
Documentation gains GitHub Pages configuration
The documents now specify GitHub Pages publishing for sovsrc.si, with the README as the homepage. The configuration uses Jekyll to build the documentation site.
13 Oct 2026
-
Research proposes a managed community compute experiment
New research proposes evaluating a small, managed compute pool for asynchronous work after permission and accounting flows function. The document reports no implementation or deployment.
-
Documents add a concise business case for proposed AI content rewards
A new article argues that proposed contributor rewards could reduce knowledge acquisition costs and encourage customer demand. It proposes measuring rewards against product improvements and sustainable margins.
-
New article examines the business case for proposed contributor rewards
A new opinion article explains how proposed contributor payments could reduce useful-data costs and support customer growth. The project remains at the specification stage.
-
New article proposes content rewards from AI revenue
A new opinion article proposes payments for authorized training material, documents, workflows, expert evaluations and adopted improvements. It separates contributor rewards from proposed public dividends and worker benefits.
-
Article contrasts creator replacement with proposed content rewards
The article contrasts selling AI imitations that exclude creators with proposed content rewards based on authorized contributions and auditable agreements. The model, marketplace and payment system remain proposals.
-
New opinion article proposes shared gains from superintelligence
A new opinion article proposes contributor rewards, voluntary syndicates and accountable human control as foundations for sharing gains from superintelligence. The model, marketplace and payment system remain proposals.
-
Training research expands to twelve commercial offerings
The research adds Modal and Baseten to a twelve-provider comparison for proposed training from scratch. It proposes different candidates for small experiments, enterprise training and supplier-assisted model development.
-
Research document proposes an open-source stack for an independent LLM
The new research document proposes PyTorch, OLMo-core and Dolma for training an independent model on separately approved material. It reports no package installations, model training or compatibility benchmarks for this research.
-
Research adds ten commercial options for training an LLM from scratch
New procurement research compares ten commercial training offerings and proposes AWS SageMaker AI as the general-purpose managed-training candidate. The shortlist reflects proposed project fit, not measured performance.
12 Oct 2026
-
Business plan proposes research services for markets and community syndicates
A new business plan proposes an AI research service over authorized collections, starting with industry knowledge and later considering financial research. The project remains at specification stage.
-
SkyRL research adds a proposed training workflow and evaluation criteria
The research document now explains a proposed SkyRL training loop using admitted documents, approved tools and reviewed scoring criteria. It separates learning scores from proposed creator payment eligibility.
-
Research compares national sovereign AI projects and commercial providers
A new research document compares sovereign AI projects on ownership, compute, model control, source rights and contributor rewards. It proposes supplying permissioned knowledge collections, attribution evidence and contractual content rewards to sovereign AI operators.
-
Documents propose self-hosted OpenCode 2 for development and experiments
The documents propose evaluating self-hosted OpenCode 2 as an engineering assistant and experiment interface with separately approved models. The assessment reports no installation or runtime validation.
-
Documents add a proposed commercial plan for research publishers
The new commercial plan proposes US specialist technology research publishers and analyst firms as initial customers. The project remains at specification stage, and proposed prices and revenue targets require customer validation.
-
Production proposal adds a licensed research assistant and model approval conditions
The new production proposal prioritizes a research assistant for a bounded, licensed technical collection, with source evidence and agreed contributor payments. The project remains at specification stage.
-
Meta Muse assessment adds proposed evaluation and teacher candidates
The research proposes Wonder and Spark as evaluation baselines and self-hosted Glimmer as a conditional model or teacher candidate. It grants none approval under the project's strict provenance and model lineage requirements.
-
Research summary proposes foundation model and teacher choices
The new research summary proposes KL3M for a source-provenance pilot and OLMo-core for training newly initialized weights on admitted sources. The project remains at specification stage and has not benchmarked or approved any model for production.
-
SkyRL assessment adds a proposed post-training option
The new assessment proposes SkyRL as an optional, self-hosted post-training framework for the research assistant. The work remains research and a proposed pilot, with no installation, GPU training or deployment performed.
-
Perplexity research adds conditions for a proposed integration
The new research proposes studying Perplexity’s publisher rewards while requiring negotiated permission for any search partnership. It records no established account, API integration, partnership or training permission.
11 Oct 2026
-
KL3M research adds a proposed provenance and rewards pilot
New research proposes evaluating KL3M and its separate Data Project in a bounded provenance pilot. KL3M remains a research candidate alongside the selected Nemotron 3 Ultra target.
-
Hermes Agent research proposes an optional interface for Sovereign Source
New research proposes evaluating a self-hosted Hermes Agent interface for permission and attribution services. The proposed integration has no implementation or benchmarks.
10 Oct 2026
-
Documents add proposed routes for independent model development
A new pathway proposes independently authorized development or explicit contractual permission for ChatGPT-assisted model development. The specification adds a proposed requirement to record authorization before using externally generated material.
-
Gemini research proposes permission checks for model training
A new report proposes excluding Gemini-generated training material from Nemotron adaptation unless a governing agreement expressly permits the workflow. It compares access routes, exceptions and commercial exposure.
-
New report examines OpenAI restrictions on competing model development
A new research report examines OpenAI output-use restrictions and proposes permission checks for Sovereign Source model development. It distinguishes hosted customization from independently distributed models.
-
Research documents assess Nemotron ancestry, evaluation methods and data suppliers
New research records ancestry conflicts for the proposed Nemotron 3 Ultra target and outlines performance evaluation and supplier review proposals. The performance review reports no model downloads or benchmark runs.
-
Specifications select Nemotron 3 Ultra and propose commercial integrations
The proposed architecture selects NVIDIA Nemotron 3 Ultra while retaining an unresolved conflict with the earlier model ancestry requirement. The specification also proposes commercial software, licensed content supply and payout candidates without establishing vendor agreements or live integrations.
-
New comparison index proposes training from scratch and holding Nemotron
The new index proposes Ai2 or Marin training tools for a model trained from scratch on permission-reviewed material. It proposes holding the Nemotron checkpoint because its documented Qwen ancestry fails the proposed selection screen.
-
Model selection proposal puts Nemotron training target on hold
The documents put the proposed Nemotron 3 Nano training target on hold because its official card discloses Qwen use. No replacement model has approval.
-
New draft audit compares model provider terms and proposes creator reward controls
A new draft audit compares provider terms on ownership, training permission, attribution and compensation. It proposes separate rights records and creator reward controls.
-
Specifications add proposed creator rewards, licensing and marketplace terms
The documents introduce SIP Creator Rewards and draft licensing and marketplace terms for agreed contribution use. The project remains at specification stage.
-
Specification adds proposed Super Intellectual Property framework and voluntary syndicates
The revised specification proposes Super Intellectual Property to recognize works, ideas, contributions and influences through evidence and agreements. The project remains at specification stage, with no application, trained model or production payment system.
-
Documents propose a research assistant pilot and contributor royalty terms
New documents propose a research assistant that checks source permissions, provides attribution receipts and records eligible contributor use. The project remains at specification stage; no application, trained model or production payment system exists.
-
Project specification document added
The repository adds a project specification document. The diff contains no text to describe its proposals.