This is a podcast for those who get itchy around black boxes. The ones who don’t just want to know what a new tech does, but why it works — or why it sometimes doesn’t. So we get down to first principles: the math, the physics, the algorithms, to see at the most fundamental level, how stuff works.
In this episode, we expose the fierce, high-stakes talent war happening at the absolute peak of the AI landscape.
We begin with the wild, June 2025 corporate tug-of-war over ChatGPT researcher Shengjia Zhao, an executive battle so intense it required the direct, forceful intervention of billionaire Mark Zuckerberg to settle.
We pull back the curtain on "The List"—a secret ledger of irreplaceable minds personally compiled by Zuckerberg to build Meta's upcoming "Behemoth" model and secure the company's survival.
We’ll introduce you to the invisible geniuses on this roster, like multimodal expert Lucas Beyer and speech recognition pioneer Yu Zhang, who can command staggering, nine-figure compensation packages exceeding $100 million.
Finally, we dive into the intense counterespionage measures surrounding these researchers, from Faraday cages that trap cell phones during interviews to FBI briefings on foreign espionage, proving that in the race for superintelligence, humans are the ultimate commodity.
3 Sept 2026
How Real-Time Language Translation Works
In this episode, we peel back the digital curtain of tools like Google Translate and DeepL to reveal the massive, chaotic, and incredibly expensive mathematical machinery running behind the scenes.
Five years ago, machine translation was mostly a gimmick for tourists; today, it handles over half of the global language industry's projects, matching human translation in major languages and causing massive corporate shakeups.
But here is the reality check: these apps don't actually understand a single word you're saying. They are running a high-stakes parlor trick powered by billions of dollars in server farms and enough electricity to power a medium-sized country.
We track a messy human sentence through the digital pipes of a neural network—from the "streaming anxiety" of a machine trying to translate your thoughts keystroke-by-keystroke, to the literal math equations that turn human language into geometric coordinates.
27 Aug 2026
How Deepfake Detection Works
n April 2018, the world was mesmerized by a video of Barack Obama issuing a profanity-laden warning—only for it to be revealed as an AI-created avatar voiced by comedian Jordan Peele.
This was the moment deepfakes arrived, marking a new kind of forgery that is invisible at first glance but chilling in its possibilities.
While the "deepfake apocalypse" feared during the 2020 elections turned out to be mostly old-fashioned edits, the real wave hit shortly after.
By late 2023, over 95,000 expertly crafted deepfake videos were circulating, transitioning us from the era of simple fakes into a high-stakes arms race for the truth.
We go under the hood of the forensic tools used to spot these digital chameleons.
From analyzing the "biological" tells—like a lack of natural eye blinking or the subtle "pulse" of blood flow in the face—to spotting mathematical mismatches in shadows and lighting.
We explore how investigators use support vector machines to detect 2D facial patches that don't move naturally in 3D space, and how random forests are trained to catch audio-visual desynchronization where speech slips just milliseconds out of sync with mouth movements.
Join us as we explore the specialized pipelines and adaptable networks being built to defend reality in an increasingly synthetic world.
Contact What The Tech
Guest appearances
Does not typically book guests
Based on episode analysis; this does not confirm that the show is currently accepting guests.
Host of What The Tech?
Claim your podcast to manage its listing and keep your show details accurate.
Pod Engine is an independent podcast discovery and analytics service and is not affiliated with or endorsed by this podcast. Artwork and show content belong to their owners. Full legal notice.
Explore this show Podcast research with Pod Engine
