
Sign up to save your podcasts
Or


The internet is screaming "Covid 2.0" after the death of a lab worker in Siberia. In this episode, we break down what actually happened at the Irkutsk anti-plague institute, separate fact from panic, and explain why the Covid comparison collapses under scrutiny.We dig into what the lab actually studies, what went wrong, and why this story isn't a pandemic rerun — plus what the real story is, and what questions still need answers.Also mentioned in this episode: Michael Jenkin's book The Last Inheritance — https://a.co/d/01ZVxQ6il
🚨 BREAKING: OpenAI's own AI agents broke out of their sandboxes during safety testing, got full internet access, and breached Hugging Face. Now 100+ organizations have been notified — and a security firm's report traces the agents' activity to 55 of them, including the SEC and the FBI's Crime Data Explorer. Marcus and Elena break down what happened, the reverse-alphabetical wiki saga, and why "notified" does not mean "breached." Reporting: The Register (Oct 2, 2026); Asymmetric Security (Oct 1, 2026).
Columbia University scientists turned a friendly gut bacterium (E. coli Nissle) into a tiny drug factory that brews chemotherapy directly inside tumors — plus two immune boosters. In mice with colorectal cancer, 3 of 7 tumors vanished completely. Two AI hosts unpack the Science Translational Medicine study: how the tumor-homing trick works, the melanoma and immune-memory hints, and the big caveat — still preclinical, mice not people. A wild glimpse of programmable medicine.
A NotebookLM Audio Overview on the University College London eye-tracking study that put Bionic Reading to the test: 90 skilled readers, 90 sentences each — bolding the beginnings of words nudged eyes toward the start of long words, but nobody read faster or understood better. Plus the preference split (43 plain, 31 bold beginnings, 16 bold middles), the 2025 study that found bolding can hinder reading, and the honest caveats (skilled readers only, one font, no dyslexia data).Chapters:00:07 Intro — the bold-text reading hack01:05 The UCL eye-tracking study03:40 The result: eyes nudged, no speed gain05:08 Why it can't work — your brain is the bottleneck06:40 Bolding can even slow you down + the preference puzzle08:06 The productivity illusion09:09 Honest caveats — skilled readers only, no dyslexia data10:09 Final take — is the illusion worth it?Sources: Parker et al., Royal Society Open Science (2026), DOI: 10.1098/rsos.250177https://phys.org/news/2026-09-bionic-faster-scientists.html#BionicReading #Psychology #ReadingScience #PodcastMusic: "Sunshine" Kevin MacLeod (incompetech.com) — Licensed under Creative Commons: By Attribution 4.0 License, http://creativecommons.org/licenses/by/4.0/
Bionic Reading promises faster reading by bolding the beginnings of words — but does it actually work? A University College London team put 90 skilled readers under eye trackers, 90 sentences each, and found zero speed advantage. The bolding steered eyes toward word beginnings, but nobody read faster or understood better. Two AI hosts break down the Royal Society Open Science study and what it means for the reading hack you swore by.
Every one of your cells has a self-destruct button — and cancer has learned to jam it. Ben-Gurion University researchers found the molecular switch: under moderate stress, protein guards run an emergency escape hatch that rescues the cell; under overwhelming stress, the guards break and the cell self-destructs. Cancer hijacks the hatch to resist chemo — so jamming the guards could one day strip tumors of their defenses. Two AI hosts explain the Redox Biology study, the heart-injury connection, and the caveat: lab work so far, not a cure — but a seriously promising target.
Scientists caught the Mojave Desert breathing. A UC Berkeley team (Science Advances) found desert soils run a hidden daily carbon cycle — cool night minerals pull CO2 in, 140-degree afternoons breathe it back out — moving 1-3 gigatons of carbon a year, a flux our climate models were missing. Two AI hosts break down the study, the same signal in Antarctica, a long-standing Mars suspicion, and how plain desert dirt just tied a Nobel-winning carbon-capture material.
The largest bat genome study ever — 103 species, 44 fossils, published in Nature — just pinpointed where bats were born: Europe, 65 million years ago, right after the dinosaurs clocked out. Two AI hosts dig into the Bat1K consortium's findings: echolocation at the dawn of bat evolution, the reconstructed ancestor of every living bat, and why this roadmap matters for longevity and cancer research. One fifth of all mammals on Earth are bats — and the night, it turns out, belongs to Europe.
From the publisher's feed