This week on Trending in Ed, host Mike Palmer is joined by Trending in Ed all-star Beth Rudden, CEO of Bast AI. From her roots digging in the filth as an archaeologist to managing a $34 billion division because the Chief Information Officer of IBM Managed Companies, Beth brings a deeply grounded, technical perspective to the bogus intelligence dialog.
On this broad-ranging and insightful dialog, Mike and Beth skip the everyday AI hype to discover what it truly takes to construct explainable, reliable know-how. Beth shares how Bast AI acts as an LLM-agnostic explainability layer—utilizing a novel ingesting chocolate analogy to reveal how they confirm AI knowledge quite than letting fashions hallucinate believable narratives. They discover the sensible utility of utilizing small language fashions (SLMs) for knowledge enrichment, highlighted by Bast AI’s significant work with Craig Hospital to translate complicated neuro-backbone outpatient procedures into accessible languages and analogies.
KEY INSIGHTS:
- Inverting the Chatbot Method: Why defining what an AI can discuss is much more practical than constructing restrictive guardrails.
- The Fantasy of “Human in the Loop”: How shifting accountability to overworked people can develop into a type of legal responsibility laundering.
- Microservices vs. Agentic Harnesses: Wanting at the dangers of pure language agentic techniques like Claude Code versus discrete, self-therapeutic duties.
- Cognitive Offloading & Math Education: Why future technical expertise ought to prioritize differential equations and the variety prediction theorem over easy calculation.
- Sample Recognition vs. Selection: Defining true intelligence by the flexibility to decide on correctly, quite than simply matching mathematical patterns.
In addition they cross paths with the Cynefin framework, clarify how the human mind conserves power by solely holding two paradoxes at as soon as, and unpack the cultural shifts reshaping fashionable engineering ethics.
Keep forward of the curve in schooling and know-how! Please like and share this episode with your community, and observe the podcast on Apple Podcasts, Spotify, or your favourite participant so that you by no means miss an episode like this one.
LINKS:
TIMESTAMPS:
- 00:00 – Introduction and welcoming Beth Rudden again to the present
- 01:00 – The ingesting chocolate analogy for Explainable AI
- 03:00 – Beth’s lightning-spherical background: Archaeology to Chief Information Officer at IBM
- 05:00 – Getting “catfished by AI” and verifying information with databases
- 07:00 – Mike on Gemini, RAG functions, and checking AI confabulation
- 09:00 – Enriched knowledge and Small Language Fashions (SLMs) at Craig Hospital
- 12:00 – Epistemic safety and inverting conversational know-how
- 14:30 – Legal responsibility laundering and the phantasm of “human in the loop”
- 15:30 – Agentic harnesses vs. self-therapeutic microservices
- 20:00 – Understanding as labor and Conrad Wolfram’s three-step math course of
- 22:30 – Future human expertise: Differential equations and jelly bean statistics
- 26:30 – Sample recognition vs. true intelligence as the flexibility to decide on
- 29:30 – Neurosymbolic techniques and subjectivity in knowledge science
- 34:30 – Shunting power: The Cynefin framework and holding paradoxes
- 38:30 – Healthcare AI scribes and physician burnout
- 44:30 – Belief architectures and constructing tech for the Upkeep Period
- 47:30 – Cultural devastation and the teleological suspension of ethics
- 49:00 – Last ideas and wrapping up with Beth Rudden
Source link
#Explainable #Beth #Rudden #CEO #Bast #Trending #Education
