Episode 300: AI’s extremes are on full display in our 300th episode. Anthropic’s Mythos model reportedly uncovered more than 10,000 security flaws in a month, accelerating vulnerability discovery for major partners. Yet the same “AI efficiency” falls apart in the real world, as seen in Starbucks’ failed AI inventory rollout that miscounted products and mislabeled items. That contrast sets up the core question of the hour: when is AI a powerful tool, and when is it just expensive theater?

We also dig into the rising stakes around biometric privacy, from Disney’s facial‑scan lawsuit to stadium and theme‑park “optional” recognition systems that don’t feel optional when the alternative line barely moves. Add in real phishing examples hitting DocuSign, Microsoft 365, and fake IRS notices, plus a case where an AI court summarizer caused a wrongful arrest, and the theme becomes clear: trust is getting harder to earn. We close with tech nostalgia, a blunt whiskey review, Waymo’s robotaxi recall, and Elon Musk’s failed lawsuit against OpenAI all coming up on TechTime Radio, with a little whiskey on the side.

-- Full Episode Details:

AI is getting dangerously good at the things we want and embarrassingly bad at the things we assumed were easy. We kick off our 300th show with a perfect contrast: Anthropic’s Mythos model reportedly uncovers 10,000+ security flaws in a month, boosting vulnerability discovery across major partners, yet the same “automation magic” falls flat when Starbucks tries AI inventory counting and ends up with mislabeled products and missed items. That tension drives the big question we keep circling: when is AI a genuine tool, and when is it just expensive theater? From there we get into facial recognition privacy and consent, sparked by Disney’s lawsuit over facial scanning at Disneyland. We compare it to Universal and stadium biometric entry, talk about what “optional” really means when the non-scan line is the long one, and why public tolerance shifts once AI becomes part of the story. If you care about digital identity, biometric data retention, and surveillance creep, this segment lands hard.

We also bring the practical stuff: real phishing email examples that mimic DocuSign and Microsoft 365 quarantine notices, plus a fake “IRS statement” that screams malware. Then Mike’s AI Guy segment hits a gut-punch case where an AI court summarizer mashed files together and an innocent man got arrested. We round it out with tech nostalgia (Apple Newton), a brutally honest whiskey review, Waymo’s robotaxi flood fiasco and recall, and a quick hit on Elon Musk losing his lawsuit against OpenAI. Subscribe for weekly tech news with zero political agenda, share the episode with a friend who clicks too fast, and leave a review so more people can find the show.

Send us Fan Mail

Support the show