This week we feature articles from Juan Sequeda, Dean Allemang & Bryon Jacob, and Steve Jones.
Additional reading comes from Eli Pariser & Deepti Doshi, Tim Bray, Ethan Mollick, and Sara A. Metwalli.
News comes from Contextual AI, Contentstack & Lytics, dbt Labs & SDF Labs, and Foxit.
Next issue arrives January 29.
All previous issues are available at https://gilbane.com/gilbane-advisor-index
Opinion / Analysis
Knowledge Graphs as a source of trust for LLM-powered enterprise question answering
This paper by Juan Sequeda, Dean Allemang & Bryon Jacob, is the result of two years of research and experimentation primarily focused on the relative contributions of LLMs and knowledge graphs to modern knowledge engineering practices. A really useful and forward-looking contribution. Includes link to last week’s updated PDF. (22 min)
“Our definition of trust is based on three main aspects: (1) LLMs hallucinate, therefore we need to ensure accuracy of responses, (2) LLMs have a black box nature, therefore we need to provide an explanation of where answers come from and (3) there is a risk of incorrect information being used, thus we need to ensure governance.”
https://www.sciencedirect.com/science/article/pii/S1570826824000441
When did random code execution become a feature?
The largest security hole, presented as a bonus
Steve Jones on AI Agents… “There are three core times you should allow unverified code to be executed within the enterprise:
- When you want to get fired
- When you’ve decided that working for the company isn’t for you
- When you’d really like to be personally liable for significant damages” (4 min)
https://blog.metamirror.io/when-did-random-code-execution-become-a-feature-e2b78a45ad6f
More Reading
- We need to protect the protocol that runs Bluesky via MIT Technology Review also see…
- Protocol Churn – ActivityPub and AT Protocol via ongoing
- Prophecies of the flood – What to make of the statements of the AI labs? via One Useful Thing
- Good intro and update for a broad audience… The state of quantum computing: Where are we today? And what we need to overcome. via Towards Data Science
Content technology news
Contextual AI launches platform for building specialized RAG agents
Specialized RAG agents orchestrate retrieval and generation based on conversational context for complex knowledge tasks with structured and unstructured data.
https://contextual.ai/blog/contextual-ai-platform-generally-available/
dbt Labs acquires SDF Labs
dbt Labs will integrate SDF’s multi-dialect, dbt-native SQL comprehension capabilities into dbt to improve performance and developer experience.
https://www.getdbt.com/blog/dbt-labs-announces-sdf-labs-acquisition
Foxit launches standalone web-based AI platform
Includes AI chat assistance, document and image analysis, summarization, multilingual translation, and text enhancement for modern document management.
https://www.foxit.com
Contentstack acquires Lytics
The composable DXP adds first-party data, audience insights, content analytics and profile management to elevate brand digital experiences.
https://www.contentstack.com ■ https://www.lytics.com
The Gilbane Advisor is authored by Frank Gilbane and is ad-free, cost-free, and curated for content, computing, web, data, and digital experience technology and information professionals. We publish recommended articles and content technology news most Wednesdays. We do not sell or share personal data.
Subscribe | View online | Editorial policy | Privacy policy | Contact
