Context Graphs & Personal Intelligence at ASIMOV DevLabs #10

Presented July 9, 2026 at Frontier Tower by Arto Bendiken

https://luma.com/asimov-devlabs-10

Transcript

1. Context Graphs & Personal Intelligence Presented July 9, 2026 at Frontier Tower by Arto Bendiken (ar.to) ASIMOV DevLabs #10
2. Table of Contents 1. Table of Contents 2. Your Presenter 3. The Era of Data Silos 4. Not a New Problem… 5. The Old Paradigm 6. The New Paradigm 7. The Paradigm Shift 8. The Algebra of Knowledge 9. AI’s Trillion-Dollar Opportunity 10. ASIMOV Platform 11. ASIMOV Protocol 12. Thank You!
3. Your Presenter @artob on GitHub, @bendiken on X Coding since 1993, full time since 1997 Prolific open-source author for 25+ years (Rust, Ruby, Python, Scheme, Drupal, etc) Built the first graph database-as-a-service (GDBaaS), years prior to Neo4j’s Created the Unlicense, a GitHub default license used by 3%+ of all repositories Built OSINT systems for the US Navy and data warehouses for S&P 500 and ESA Led the EVM team at NEAR Protocol Featured in exhibit 270 in the Silk Road trial Raised $15M+ total so far in four startups
5. The Era of Data Silos You probably have hundreds of apps installed, most of which are effectively data silos: you can check out any time you like, but you can never leave You probably have 4-10 distinct messengers, none of which interoperate with each other Your personal AI is going to have a hard time offering true utility with this mess… If you’ll have separate personal and work AIs, will that help or will it make it worse?
6. Not a New Problem… Enterprises and governments have invested heavily into data integration for decades Large enterprises often have hundreds or even 1,000+ distinct information systems No modern enterprise could possibly be competitive without a large budget for data integration and data warehouses Enterprise data integration (EDI) is a massive, fantastically profitable market Enterprise data warehouses are best built using graph technology (cf. Palantir)
7. JSON, CSV, & Roman Numerals Today’s data ecosystem suffers a structural problem analogous to the late Roman Empire’s mathematical stagnation Roman numerals were an absolute bottleneck that held back science & technology in Europe for centuries By continuing to rely on ad-hoc lexical data structures, we force software to act as a physical "abacus", manually translating and mapping data at every step All the aforementioned is based on "dumb" implicit procedural knowledge
8. Graphs & the Hindu- Arabic Numeral System The Hindu-Arabic system was one of the most crucial technological leaps in history It was popularized in Europe as late as the 13th century by Fibonacci who learned it from Arab traders Its widespread dissemination laid the groundwork for the Scientific Revolution and everything that followed The new paradigm is based on embedded explicit declarative knowledge (the notation’s place value, URIs, ontologies)
9. The Paradigm Shift (1/3) Feature Syntactic Files (JSON, XML, CSV, etc) ("Roman Numerals") Semantic Graphs (RDF, OWL) ("Hindu-Arabic System") Core Nature Syntactic & lexical: meaning is tied to the physical shape, position, and local context of the file Semantic & symbolic: meaning is universal, self-describing, and independent of any particular serialization format (!) The Concept of "Zero" Absent: Missing or null data breaks schemas; there is no native representation of universal "non-existence" or open-world assumptions Present (the open-world assumption): Absence of data simply means it is unstated or unknown, allowing graphs to grow dynamically without breaking Abacus arithmetic with syntactic files versus symbolic algebra with semantic graphs
10. The Paradigm Shift (2/3) Feature Syntactic Files (JSON, XML, CSV, etc) ("Roman Numerals") Semantic Graphs (RDF, OWL) ("Hindu-Arabic System") Integration Method The manual abacus: custom ETL (extract, transform, load) pipelines, hardcoded APIs, and manual schema mapping Mathematical synthesis: automatic merging via graph union (A ∪ B). Universal identifiers (URIs) align data automatically Scalability Linear friction: every new data source requires O(N^2) custom connectors to talk to existing sources Network effect: every new data point natively connects to the existing web of knowledge in an O(1) integration Abacus arithmetic with syntactic files versus symbolic algebra with semantic graphs
11. The Paradigm Shift (3/3) Feature Syntactic Files (JSON, XML, CSV, etc) ("Roman Numerals") Semantic Graphs (RDF, OWL) ("Hindu-Arabic System") Source of Meaning Centralized/authoritarian: a central authority dictates the schema and distributes documentation for others to parse the strings Decentralized/democratic: anyone can coin a URI and publish an ontology; meaning is defined globally and collaboratively across the web without requiring permission Evolution of Meaning Destructive migrations: changing a schema breaks downstream APIs; must rewrite tables and update all consuming application code Non-destructive augmentation: schemas are just more graph data; the data instantly adapts via inference without breaking existing structures and consumers Abacus arithmetic with syntactic files versus symbolic algebra with semantic graphs
14. “That’s the context graph, and that will be the single most valuable asset for companies in the era of AI.” — Jaya Gupta & Ashu Gang, Foundation Capital, AI’s trillion-dollar opportunity: Context graphs (Dec 2025)
15. ASIMOV Platform & ASIMOV Protocol We are building the ASIMOV Platform: an open-source polyglot development platform for trustworthy, neurosymbolic AI based on flow-based programming (FBP) and semantic knowledge graphs (RDF) Neurosymbolic AI marries neural networks (think LLMs) with symbolic knowledge representation & reasoning (KR&R) to create AI systems that can understand and reason about structured knowledge beyond mere plain text The ASIMOV Platform is built as a modular Rust framework delivered as a command-line interface (CLI) and software development kits (SDKs) for Dart, Python, Ruby, Rust, and TypeScript The ASIMOV command-line interface (CLI) allows extracting and elevating any URL or file format into a knowledge graph so long as we have a module that supports it Currently, we have modules e.g. for extracting your chats & contacts from Signal and Telegram, your emails from any IMAP server, your bookmarks from any browser, and some 40+ other data sources Follow us at x.com/ASIMOV_Platform and github.com/asimov-platform!
16. ASIMOV Platform & ASIMOV Protocol We are also building the ASIMOV Protocol: the knowledge layer for trustworthy personal AI, built as a decentralized P2P protocol for sharing cryptographically-verified knowledge graphs The ASIMOV Protocol is built on top of Iroh, enabling end-to-end encrypted QUIC connections between devices based on cryptographic identity without regard for location or network topology Sync your personal knowledge base between your devices (iPhone, iPad, MacBook, etc) in an end-to-end secure, P2P manner without needing any centralized cloud infrastructure Share knowledge—the gold standard of information—directly with your family, friends, and colleagues, with all knowledge preserving provenance and traceability (where did it come from, who asserts or underwrites it, etc.) Incorporate large external structured knowledge sources (e.g., Wikidata, OpenStreetMaps, IMDB, etc) that your agents can access through efficient P2P graph queries Follow us at x.com/ASIMOV_Protocol and github.com/asimov-protocol!
17. Thank You! x.com/bendiken github.com/artob ar.to Find us at: x.com/ASIMOV_Protocol github.com/asimov-protocol github.com/asimov-platform github.com/asimov-modules Find me at: