Elevate with Robert Glazer Thinking Thursdays: Can We Have A Sane Conversation About AI?
12 snips
Feb 26, 2026 Mick Sloan, producer and Thinking Thursdays co-host who researches and analyzes AI topics, joins a calm, skeptical deep-dive. They pick apart hype around agent tools and giant model claims. They question hidden costs, pilot failure rates, data center economics, media sensationalism, job disruption scale, and what sensible adoption looks like.
AI Snips
Chapters
Transcript
Episode notes
Hide Real Playbooks And Scale Them Internally
- Don't publicize a profitable proprietary process; keep internal wins private and scale them instead of selling how-to guides.
- Robert Glazer notes that if a strategy truly worked, teams would run it privately rather than monetize instructions.
Agent Experiments Can Cause Irreversible Harm
- Experiments with agents can cause real damage if misconfigured.
- Robert Glazer recounts an agent that deleted wedding photos and warns users are responsible for credential misuse and other irreversible harms.
Model Progress May Be Incremental Not Exponential
- Claims of rapid, accelerating model improvement are contested.
- Mick Sloan cites Cal Newport's rebuttal that current model improvements are smaller increments, and coding gains drove recent perceived leaps.