Modern CTO LLMs, Agentic AI & Blackmail with Jon Krohn, Host of the Super Data Science Podcast
9 snips
Sep 4, 2025 
Jon Krohn, host of the Super Data Science podcast and co-founder of YCarrot, dives deep into the realm of AI. He explains the stark differences between large language models and agentic AI, emphasizing the ethical concerns around AI misalignment and blackmail. Businesses can significantly enhance ROI by leveraging agentic AI for streamlined operations, particularly in healthcare. The discussion also touches on the importance of accessible training and the balance between personalization and efficiency in AI integration.
AI Snips
Chapters
Transcript
Episode notes
Prioritize Fast ROI Projects
- Prioritize AI projects by ROI and time-to-value before building them out.
- Prototype quickly and iterate from prototype to high-volume production with expert help.
Cheap LLMs + Agent Checks Improve Accuracy
- Agentic frameworks let you pair cheap small LLMs with checking steps to improve accuracy.
- This yields high-accuracy results on low compute, expanding practical enterprise uses.
Triage Cold Outreach With LLM Monitors
- Automate triage with labels and an LLM monitor to reduce noise and cost.
- Create drafts rather than sending automatically, then review and refine drafts for quality control.