Introduction
This blog summarizes the news on training and learning for building AI applications discussed in our 2025-09-05 webinar.
PAI (Partner with AI)
We framed PAI in the context of current markets: with layoffs and fewer new hires, how can professionals adapt? PAI outlines strategies to thrive by partnering with AI.
Labor-Market Signals
- Early-career employment declines in AI-impacted roles.
- Unemployment among workers aged 22–25 in AI-exposed jobs has dropped by ~13%, signaling rapid shifts that require new graduate skills.
Curriculum vs. Tacit Skills
- Current curricula emphasize codified knowledge that AI can replicate.
- Tacit knowledge and AI-augmentation skills protect workers and improve resilience.
- Leaders such as Andrew Ng echo the focus on practical, AI-augmentable skills.
Agentic Frameworks: What’s Leading?
We compared several agentic stacks and when to use them.
- LangGraph (LangChain ecosystem)
- LlamaIndex Agents
- CrewAI
- AutoGen
- LangChain Agents
- OpenAI Swarm (lightweight patterns)
Demo: LangGraph
Live demo of LangGraph: building tool-using, multi-step workflows with explicit state management and control over retries and edges.
Agentic Knowledge Graph Construction (Neo4j)
We revisited constructing knowledge graphs agentically with Neo4j to add relationship intelligence on top of semantic retrieval.
Enjoyed the read?
Subscribe for weekly insights from the Elephant Scale team.