Episode 209: Daniel Dobriy
Twenty-five years after the original vision of the Semantic Web, Daniel Dobriy argues we're finally at the tipping point where symbolic AI and large language models can converge into a unified, trustworthy knowledge infrastructure. He contends that LLMs alone are unreliable because they lack verifiable sources—what's missing is a 'neuro-symbolic' layer where natural language is grounded in explicit, structured data like RDF and linked open data. Dobriy's company, Dobry AI, is building tools like Graph Hub—a platform to host knowledge graphs with instant interoperability—and the Web of Data Catalog, a search engine that indexes public data sources to enable AI agents to answer questions with traceable, authoritative evidence. He warns that while big tech has ignored open standards like RDF, the time has come to integrate wikis, enterprise data, and AI into a single, sovereign ecosystem. The future, he believes, isn't just smarter AI—it's AI that can be audited, validated, and trusted because it's built on a shared, open foundation of linked knowledge.
LLMs lack reliability because they can't trace answers to verifiable data—structured knowledge graphs are essential for trustworthy AI.
The Semantic Web’s original vision of AI agents powered by linked data is finally becoming feasible with modern AI and open standards.
Graph Hub by Dobry AI lets users host and query RDF knowledge graphs instantly, with support for Wikidata clones and enterprise data integration.
The Web of Data Catalog (catalog.ai.wu.ac.at) indexes public data sources, enabling AI agents to search and validate answers from authoritative, open data.
Future AI systems should be neuro-symbolic: combining LLMs as natural language interfaces with symbolic representations for logic, inference, and verifiability.
…and 3 more takeaways available in PodZeus
Introducing Daniel Dobriy: From Research to AI Entrepreneur
Jeroen Courant welcomes Daniel Dobriy, founder of Dobry AI and researcher at WU Vienna, who shares his dual role in academia and entrepreneurship, rooted in a long-standing passion for semantic web technologies and knowledge graphs.
The Semantic Web’s Long-Awaited Comeback
“If anything, it is getting validated. If you're getting at that, then we had this initial vision of the semantic web. Right? Maybe you could link the seminal Semantic Web paper on the details of the podcast.”
Why LLMs Alone Aren’t Enough: The Need for Verifiable Knowledge
“That's what essentially is the blocker towards LLMs being applied everywhere in productive environments, in enterprises. This lack of ultimate certainty, right?”
The Web of Data Catalog: A Search Engine for Structured Knowledge
“It's a data search engine that we are also building to showcase all of the link open data ecosystem, but also data that is in the wikis, data that is in the data catalogs and managed by authoritative data producers.”
Graph Hub: Making Knowledge Graphs Accessible to Everyone
“You can set up your own Wikidata endpoint that will be available to you or you can set up another endpoint with your own data and then integrate it with the search engine and make publicly available for everybody.”
“So that's what essentially is the blocker towards LLMs being applied everywhere in productive environments, in enterprises. This lack of ultimate certainty, right?”
“I feel like if anything, it is getting validated. If you're getting at that, then we had this initial vision of the semantic web.”
“So that is a data search engine that we are also building to showcase all of the link open data ecosystem, but also data that is in the wikis, data that is in the data catalogs and managed by authoritative data producers.”
Host
Guest
Daniel Dobriy
person
Dobry AI
organization
Wikidata
organization
MediaWiki
organization
WU Vienna
organization
Web of Data Catalog
organization
Graph Hub
organization
Bilateral AI Cluster of Excellence
organization
Semantic MediaWiki
organization
WikiAPIRI
organization
Beyond the Hype: Where Should SMBs Actually Start With AI?
21m • 6/4/2026
Legacy Systems Are a Liability. Here's What to Do About It.
17m • 6/5/2026
EP1007: Award-winning core banking
22m • 6/11/2026
Start discovering podcast insights today
Start with a 7-day trial and explore a growing catalog of popular podcasts. No credit card required.
No credit card required • 7-day trial • Cancel anytime