Skip to content
Episode from the podcast InsideAnalysis
How Do Large Language Models Work? artwork

Episode

How Do Large Language Models Work?

Released Monday, May 11th 2026 52 min

Show notes

About this episode

Large Language Models are rapidly reshaping the way people search for information, write software, analyze documents, and interact with machines. But beneath the polished interfaces and conversational fluency lies a complex architecture built on probabilities, patterns, transformers, and attention mechanisms. Understanding how these systems actually work is becoming increasingly important for business leaders, technologists, and everyday users alike.

More from InsideAnalysis

All episodes

Agentic ERP: The Future of Enterprise Software

44 min

What happens when enterprise software moves beyond traditional workflows and starts working proactively through AI agents? Sam Yen of Lensing.ai joins InsideAnalysis to explore how Agentic ERP is transforming the way businesses operate, enabling greater agility, automation, and innovation. Learn how secure, virtual sandboxes powered by real enterprise data allow organizations to experiment, test AI agents, and redesign business processes without risking production systems. Hear more about the evolving role of traditional ERP, the importance of trust and governance, and how AI is dramatically reducing the distance between a business idea and its real-world implementation.

AI-Driven Business and the Battle for Trust

48 min

AI is reshaping how businesses operate, how services are delivered, and how entrepreneurs build value. Join this episode of InsideAnalysis as host Eric Kavanagh discusses with Aater Suleman of Vixul how AI is transforming technology services, creating new market opportunities, and challenging traditional business models around hourly pricing, productivity, and incentives. In the second part, Nick Loui, co-founder and CEO of PeakMetrics, explores the growing importance of narrative intelligence, showing how organizations can distinguish genuine public sentiment from bot-driven activity, track misinformation, and understand how AI is influencing brand perception. Listen to this episode of InsideAnalysis to find out how businesses can adapt, compete, and build trust in an increasingly AI-driven world.

Personalization at Scale: How AI Is Transforming B2B Marketing

44 min

For decades, personalized marketing has been a promise that businesses have struggled to deliver. Now, AI is making it possible to move beyond generic campaigns and create experiences tailored to individual buyers. Join this episode of InsideAnalysis as Eric Kavanagh speaks with Matthew Langie, Co-Founder and CMO of Personize, about how the company combines CRM and external data to build detailed customer profiles and deliver more relevant marketing across ads, email campaigns, and landing pages. They explore how AI is changing B2B marketing and helping companies make every customer interaction count.

Exploring the Future of AI and Enterprise Innovation

45 min

How is AI transforming the way organizations work, build, and innovate? Join this episode of InsideAnalysis as host Eric Kavanagh interviews Les Ottolenghi of FuzeBox AI and Poonacha Kongetira of Classie AI. Hear more about the evolving AI landscape, emerging opportunities, and the technologies shaping the future of business.

Grounding LLMs in Data: Quantitative & Causal Reasoning

44 min

LLMs are remarkably powerful at language and open-world knowledge, but that doesn't necessarily make them reliable for quantitative reasoning or analysis of enterprise data. Join this episode of InsideAnalysis as Eric Kavanagh speaks with Pradeep Ravikumar of WoodWide.ai about bridging that gap. Learn more about how WoodWide.ai is working to automate the journey from data to intelligence while keeping data teams in the loop. Hear more about the fundamental differences between language models and structured enterprise data, the importance of quantitative and causal reasoning for use cases such as customer churn and fraud detection, and how organizations can leverage the strengths of LLMs without sacrificing the rigor of their existing data models.

This podcast, its content and its artwork are not owned by, affiliated with, or endorsed by Castquire.

Up next

  1. Nothing queued. Use + Queue on any episode to line it up.

0:00 0:00