
Sign up to save your podcasts
Or


Hydrogen Drops Fast, Agents Get Isolated · 13 Nov 2026
Google ships Gemini 4 Hydrogen straight into Search and the Gemini app without the staged security rollout its Carbon and Argon models got, while AWS pitches hardware-isolated agent runtimes and Databricks pushes a whole lakehouse stack for the agent era. We also dig into preprints on agent reliability and context graphs, and ask what any of this is actually proven to do.
15 stories · 16:06 · voices are AI-generated
Stories in this episode
1. Google rolls out Gemini 4 Hydrogen, finally replacing 3.8 Flash
Ars Technica · New tool
2. AWS launches next generation AgentCore runtime for secure, efficient AI agents
AWS Machine Learning Blog · New tool
3. Google's latest Search upgrade promises better answers to complicated questions
TechCrunch · New tool
4. AWS shows how to build a self-service LLM gateway with Bedrock and LiteLLM
AWS Machine Learning Blog · New learning
5. Databricks announces Genie Migrate for Lakehouse is generally available
Databricks Blog · New tool
6. Google adds Memory summary option to let users edit what Gemini knows about them
ZDNET · New tool
7. Microsoft Research Focus Edition #98 covers AI translation, math breakthrough, and coding agents
Microsoft Research Blog · Scientific discovery
8. AgentSSL studies how MLE agents can leverage unlabeled data for pipeline design
arXiv search: agentic AI · Scientific discovery
9. LiveGeo benchmark tests LLM agents in continuously evolving geographic environments
arXiv search: agentic AI · Scientific discovery
10. Context Graph offers explainable multi-agent decision chains for AI-RAN
arXiv search: knowledge graphs, ontologies and semantics · Scientific discovery
11. How Constellation is building a governed data foundation on Databricks
Databricks Blog · New learning
12. Dispelling LTAP myths
Databricks Blog · New learning
13. LinkedIn article covers context engineering for AI agents as a 3-layer system
LinkedIn public discovery: context graph · New learning
14. Databricks pitches Lakehouse//RT for real-time analytics driven by AI agents
Databricks Blog · New tool · hype watch
15. Databricks features H6.ai's AI-powered financial automation on Lakebase Postgres
Databricks Blog · Event
Agents, Ontologies and the Cost of Context · 12 Nov 2026
This episode surveys a busy week in AI and data: AWS's October updates for Bedrock, AgentCore and Strands, its re:Invent AI track and reusable-asset catalog, Databricks' autonomous Genie with its agent-built ontology, and Anthropic's Claude joining meetings. We also look at semantic layers gaining executive attention, live dashboards from air-gapped factories, AlphaGenome 2, an AI-benefits database, agent behavior and AI hacking fears, plus a McKinsey-ElevenLabs alliance and a call for evidence in US-China AI policy.
15 stories · 21:18 · voices are AI-generated
Stories in this episode
1. AWS recaps October 2026 updates for Bedrock, AgentCore and Strands AI builders
AWS Machine Learning Blog · New tool
2. Claude will soon be able to join team meetings, with some limitations
ZDNET · New tool
3. Databricks announces the new Genie: autonomous AI for complex enterprise work
Databricks Blog · New tool
4. AWS publishes attendee's guide to re:Invent 2026 AI/ML track sessions
AWS Machine Learning Blog · Event
5. AWS offers catalog of 23 reusable assets to accelerate generative AI delivery
AWS Machine Learning Blog · New tool
6. Former Facebook employee builds database documenting AI's measurable benefits
Fast Company Technology · New tool
7. AWS shows event-driven pipeline to keep Agent Registry updated automatically
AWS Machine Learning Blog · New learning
8. AlphaGenome 2: Deciphering how cells read DNA to better understand disease
Google DeepMind · New tool
9. Bright Machines delivers live factory dashboards to customers without opening air-gapped plants
Databricks Blog · New tool
10. LinkedIn post highlights new Google x Strategy semantics layer for agentic workflows
LinkedIn public discovery: semantics · New tool
11. LinkedIn post argues semantic layer is a hot topic finally getting leaders' attention
LinkedIn public discovery: semantics · Opinion and analysis
12. NYT Magazine explores why it is hard to teach AI agents to behave
New York Times Technology · Opinion and analysis
13. Cybersecurity pros argue panic over AI-powered hacks is misplaced
Fast Company Technology · Security incident
14. McKinsey and ElevenLabs announce alliance to scale conversational AI agents
ElevenLabs Blog · Event
15. CSET argues espionage allegations and data-center resistance shape US-China AI competition
Georgetown CSET · Opinion and analysis
AWS says context management is the real bottleneck for production agents, while Snowflake, Databricks and Microsoft all rush to brand their own 'ontology'. We ask what these vendor ontologies actually contain, why semantics professionals cannot agree on their own vocabulary, and whether knowledge graphs can keep agents from believing stories the data never told. A critical tour through the context layer hype.
10 stories · 21:43 · voices are AI-generated
Stories in this episode
1. AWS outlines four layers of context engineering for production AI agents
AWS Machine Learning Blog · New learning
2. Jeff Winter explains knowledge graphs in 90 seconds using a machine-lot-order example
LinkedIn public discovery: ontologies · New learning
3. AI Every Time post asks what an ontology for AI agents actually needs
LinkedIn public discovery: ontologies · Opinion and analysis
4. Kurt Cagle argues the ontologist role will fragment into specializations
LinkedIn public discovery: semantics · Opinion and analysis
5. Snowflake announces Business Ontology, a governed layer for shared business terms, entities and metrics
Snowflake · New tool
6. Chris Brockmann: everyone has an ontology now, almost nobody has an ontology
LinkedIn public discovery: ontologies · Hype · hype watch
7. Willem Koenders notes the irony that semantics professionals cannot agree on their own terms
LinkedIn public discovery: ontologies · Opinion and analysis
8. Shane Gibson asks whether Snowflake's ontology is more of an ontology than Databricks' Genie
LinkedIn public discovery: ontologies · Opinion and analysis
9. LinkedIn post highlights Port.io's pivot toward designing the context layer first for agents
LinkedIn public discovery: context graph · Opinion and analysis
10. LinkedIn post contrasts schemas and ontologies as answering different questions
LinkedIn public discovery: ontologies · Opinion and analysis
From the publisher's feed