THMG546 – AI for the Fail!!

THMG546 – AI for the Fail!!

From The HazMat Guys by Bobby Salvesen and Mike Monaco

May 1, 2026 · 29 min · Season 10 · Episode 546

About this episode

The episode discusses common AI failures and how input quality and hidden account context affect AI outputs.

We start by unpacking common AI fails and how the quality of your input directly shapes the output. From there, Bob goes deeper into long-term context that may already be set up in your account without you realizing it, and how that hidden context can influence what an AI assistant generates over time. You’ll hear practical guidance on spotting when account-level memory or context is affecting responses and why that matters for accuracy. Bob also walks through how to disable that long-term context so you can regain control over what the AI is using behind the scenes. If you’ve been frustrated by inconsistent AI results, this conversation helps connect the dots between your settings, your prompts, and what the model gives back. EPISODE SPONSORS Rigaku Analytical Devices Rigaku provides handheld Raman analyzers for first responders facing unknown hazards. Their 1064 nm Raman technology reduces fluorescence interference, enabling chemical identification in under a minute—even in difficult samples. Rigaku’s fourth-generation platform expands accessory support to help teams assess hazards faster and safer. https://www.rigakuanalytical.com Teledyne FLIR Defense CBRNE threats don’t wait…

People in this episode

Hosts: Bobby Salvesen, Mike Monaco

Topics covered

  • AI failures
  • input quality
  • account-level memory
  • AI context
  • accuracy in AI
  • control over AI settings

Keywords

  • AI
  • failures
  • input quality
  • context
  • accuracy
  • settings
  • memory

Sponsors

Rigaku Analytical Devices, Teledyne FLIR Defense

More episodes of The HazMat Guys

Explore listener stats, chart rankings, contacts and more on the The HazMat Guys podcast page.