A lot of modern AI models have a kind of security guard layer that sits in front of them. Its job? A binary choice as to whether the prompt heading into the model is safe or not. Kasimir Schulz, a lead security researcher at HiddenLayer, has been researching how to trick these models. Their solution, a technique called "Echogram" involves words with such positive statistical sentiment — such overwhelming good vibes — that it flips that verdict.

Learn more about your ad choices. Visit podcastchoices.com/adchoices

Two FBI agents. One room. One of them is the most damaging spy in U.S. history.

Robert Hanssen told a lot of lies — including a really weird one about booking the Beach Boys for the FBI. That lie didn’t matter all that much, but the others did. For 22 years, Hanssen sold America’s deepest secrets while hunting moles inside the Bureau.

With retirement looming, the FBI set a trap: a fake department, a fake job, and a young agent named Eric O’Neill placed three feet from the suspected spy.

This episode is our conversation with Eric O’Neill — the man tasked with spying on the spy — about lies, tradecraft, psychological warfare, and the sting operation into what the DOJ later called “possibly the worst intelligence disaster in U.S. history.”

Learn more about your ad choices. Visit podcastchoices.com/adchoices

Pokémon cards became a billion-dollar market—and then a massive fraud target. This episode follows the rise of ultra-rare Pokémon prototypes, the grading systems meant to protect collectors, and the amateur investigator who used codebreaking and printer forensics to expose a modern forgery ring hiding in plain sight. All that plus a nice chatty chat after the break to kick off the year.

Learn more about your ad choices. Visit podcastchoices.com/adchoices

It’s the 2025 Chatty Chat year in review friends! We’re diving into a big old pile of stories from the past year and speculating on what’s to come in 2026. If you’re wondering where this sits on the "in-depth interview vs. casual chatting" spectrum, just know we spend a considerable amount of time talking about how rad the Switch 2 is. This is not a deep dive, just a good time.

Learn more about your ad choices. Visit podcastchoices.com/adchoices

A deep dive into Lighthouse, a phishing-as-a-service platform linked to millions of scam texts worldwide, and the sprawling “smishing triad” ecosystem built around it. With security researcher Ford Merrill, we unpack how modern scam operations work at industrial scale — from fake e-commerce sites and mass SMS campaigns to the wallet-provisioning techniques that let criminals turn stolen credit cards into tap-to-pay phones. 

Learn more about your ad choices. Visit podcastchoices.com/adchoices

For the last few years, Grindr has branded itself as the global gayborhood—a digital safe space for queer communities. But a series of European investigations raised serious questions about how the app handled user data. In this episode, we talk with Chaya Hanoomanjee of the law firm Austin Hayes, who is leading a major UK claim alleging that Grindr shared sensitive user information, including in some cases health data, without proper consent. We trace the story from Norway’s multimillion-euro fine to the 12,000-person UK action and unpack what “special category data” actually means, why these cases matter, and what we’re all really agreeing to when we tap “accept.”

Learn more about your ad choices. Visit podcastchoices.com/adchoices

It’s a chatty chat. I repeat, this one's a chatty chat. Today we’re digging into the big weird questions on our desk: what percentage of Meta’s revenue allegedly comes from knowingly running scam ads, what exactly recently pardoned Binance founder Changpeng Zhao did and why he walked free, and what happens when AWS—the concrete pad foundation of the modern internet—goes down. If long-form nerd talk isn’t your vibe, totally fine—we’ll catch you next time. For everyone else, here’s some good good chattin’ for your dishes, your commute, or whatever you’re up to right this second.

Learn more about your ad choices. Visit podcastchoices.com/adchoices

Hh hey maybe don't buy that $14 projector off amazon. In this episode, we dive into the sequel nobody asked for: BADBOX 2.0 — the return of last year’s botnet built out of bargain-bin Android gadgets. Google just filed a lawsuit in federal court alleging that millions of sketchy streaming boxes, projectors, and mystery electronics were shipped pre-infected from factories overseas. The moment someone plugs one in, it joins a global botnet used for ad fraud, click fraud, and even to rent out your home internet connection to criminals. We talk to the team at HUMAN Security, the researchers credited in Google’s suit, about how they traced this thing across 222 countries, why it came back bigger than before, and how you even begin to kill a botnet that ships itself directly to people’s living rooms.

Learn more about your ad choices. Visit podcastchoices.com/adchoices

The story of a massive swatting hoax campaign across the US that helped take down a piece of cybercrime infrastructure in the heart of New York. That and a big rambling chatty chat about Tilly Norwood, AI bubbles, and somehow, very briefly, Goku.

Learn more about your ad choices. Visit podcastchoices.com/adchoices

The question: can one host of a podcast that was planning to do a chatty chat episode proceed at the last minute while the other host is unfortunately out sick? Wish our dear pal Scott a speedy recovery and wish ya boy luck as he threads this needle. We discuss — I discuss? — the heist of Mexico’s financial system, the European airport hack, and a slew of other tech tales. 

As I say in the ep — if one guy talking into a mic telling weird tech stories alone isn’t your jam, I will not blame you for rejoining us in the next one when your TWO intrepid cohosts are back at it in what I’m assuming will be full health. 

Learn more about your ad choices. Visit podcastchoices.com/adchoices