The fastest way to lose your privacy in 2026 is to assume the default settings are on your side. We start with a big question that’s about to hit schools and workplaces hard: can AI watermarking actually prove a piece of writing was generated by a model? Anthropic says yes, using a subtle “fingerprint” inside the text itself, built through token steering and patterned word choices. We walk through how it works, why it’s different from metadata labels, and where it can go wrong, especially when false positives turn into real consequences.

Next, we get practical and blunt about online safety. The FBI is warning that criminals are breaking into accounts to steal private photos and then using them for sextortion, harassment, and resale. We talk through what the attackers look for, why public visibility makes student athletes a target, and what actually reduces your risk: strong passwords, multi-factor authentication, and keeping sensitive content out of accounts that can be compromised.

Then we bring back our listener-favorite letters segment and read real scam emails, because nothing sharpens your instincts like seeing the tricks up close. We break down “support ticket” bait designed to push you into granting remote access, a surprisingly smart security-themed email that hides a malware download behind a “read more” link, and a fake Facebook lookalike domain that’s almost impossible to notice at a glance. Mike the AI guy caps it off with a rant you’ll want to hear about Twitch and Amazon training generative AI on streams and chat logs unless you find the buried opt-out.

If you like smart tech news, scam breakdowns, and privacy talk with a whiskey pairing, hit subscribe, share the show with a friend who clicks too fast, and leave us a review so more people can find TechTime Radio.

Send us Fan Mail

Support the show

A judge calls Meta a “public nuisance” and compares social media harm to air pollution. That line hits hard because it forces the real question: if the damage is baked into engagement and ad targeting, what kind of fix is even possible? We dig into the record child safety fine, the idea of “mitigation funds,” and why accountability gets messy when platforms are both the marketplace and the referee.

Then we jump to a story that’s funny right up until it isn’t: Google’s AI Overview repeats a meme as if it were fact, claiming license plate reader cameras are packed with gold and copper. We unpack what AI hallucinations mean when search becomes “answers” instead of links, how misinformation can drive real-world behavior, and why verifying sources matters more now than ever.

On the security front, we break down phishing-as-a-service aimed at Microsoft 365 accounts, including spoofed RingCentral messages, email authentication gaps (SPF, DMARC, DKIM), and the simple habits that keep a bad click from turning into a full compromise. We also take a breather with Gwen’s gadget pick: a Kickstarter breakfast robot that cracks and cooks eggs and syncs a toaster so it won’t “scare” you, plus our whiskey tasting and a data-driven teardown of the Tesla Cybertruck’s flop era.

If you like smart tech news with humor and practical takeaways, subscribe, share the episode with a friend, and leave us a review. What’s the last “AI answer” you trusted that turned out to be wrong?

Send us Fan Mail

Support the show

The “robot future” is not coming quietly, it is showing up as bans, backlash, and battles over who controls the machines that will run warehouses, factories, and daily life. We kick things off with a big move from the FCC to block future imports of foreign-made humanoid robots and quadruped robots, citing national security risks like surveillance, data collection, and remote control. Then we ask the uncomfortable question: is this about safety, or about buying time in a trillion-dollar AI robotics race where China already leads production?

From there, the story gets weirdly personal. A school district tries to bring a humanoid robot into the classroom as a teaching assistant, and the community reaction is instant and loud. We talk cost, optics, student privacy, and why “it will not connect to the internet” is not the same as “it cannot.” If education needs help, what kind of tech actually supports teachers instead of replacing trust with a gadget?

We also dig into policy that looks strong on paper but collapses in real life, with Australia’s under-16 social media ban failing under weak age checks and easy workarounds. And we highlight a professor’s clever AI cheating trap that exposes a bigger issue than plagiarism: students skipping the habit of reading and verifying their own work. We close with a sobering cybersecurity fail involving attacks on water facilities, plus our whiskey tasting and the most absurd $57,000 “AI luxury pillow” you will ever hear about.

If you like sharp tech news with humor and plain-English analysis, subscribe, share the episode, and leave us a review so more listeners can find the show.

Send us Fan Mail

Support the show