October 5, 2026 · ITK Podcast · Hosts: Alex and Jordan

Alex: Welcome to the I T K Podcast, Information Technology, the show where we question everything, starting with the news itself. I'm Alex.

Jordan: And I'm Jordan. It's Monday, October fifth, and we've got a packed week, from Google's newest A I brain to a chip factory power struggle to weight-loss drugs that might slow down aging itself.

Alex: Let's start with Google, because they dropped a new frontier model called Gemini four Argon. It can generate up to a million output tokens in a single run, meaning it can think and write for an extremely long time without losing the thread.

Jordan: And the headline use case isn't longer essays. Google says this thing can autonomously find, validate, and patch software vulnerabilities, an A I that reads code like a security researcher and fixes the holes it finds.

Alex: Rolling out first to trusted testers and cyber defenders, with developers, enterprises, and A I Ultra subscribers later. But the L A Times reported internal skepticism about its coding performance, like Google's own people aren't fully sold yet on the benchmarks.

Jordan: That's the question-everything lens right there. A model that finds its own bugs is either a giant leap for cybersecurity or a giant leap for marketing. We'll believe the coding chops when independent developers put it through real projects.

Alex: Speaking of leap or marketing, Elon Musk's Terafab had a wild weekend. Musk confirmed on Saturday that T S M C, the company that makes most of the world's advanced chips, could take part in his Texas chip venture, jointly led by Tesla and SpaceX.

Jordan: This is the fab project Intel first backed back in April with its fourteen A process. The market reacted like you'd expect. Intel fell about four percent this morning, and T S M C rose about three percent in Taiwan trading.

Alex: One analyst floated the idea that T S M C might run a fab serving only Tesla, SpaceX, and x A I. But a Musk post on X is not a signed contract, so we'll see whether this becomes actual silicon or just another headline.

Alex: Over at Apple, October is stacking up. Bloomberg's Mark Gurman reports an event on Tuesday, October thirteenth to unveil a long-rumored smart home hub, a roughly six to seven inch display running something called home O S, plus a HomePod mini two and a new Apple T V four K.

Jordan: And the foldable iPhone, the iPhone Duo, is still slated to launch October twenty-third, with pre-orders on the sixteenth, at one thousand nine hundred ninety-nine dollars. But the supply chain is reportedly struggling with production yields, analyst-speak for they can't build them fast enough without defects. Thin stock at two grand is a fast way to turn excitement into frustration.

Alex: So Jordan, smart home hub from Apple in twenty twenty-six, does this finally make their smart home stuff matter?

Jordan: Only if it talks to everything else. The graveyard of smart home hubs is full of beautiful devices that couldn't play nice. We'll find out on the thirteenth.

Alex: One money story for the I T folks before the break. Multiple outlets report A M D and Intel are both raising chip prices around ten percent, passing through rising T S M C wafer costs.

Jordan: And it's not just processors. Nvidia's upcoming R T X sixty eighty and sixty ninety cards are reportedly getting their launch order reshuffled because G D D R seven memory chips are scarce and pricey, and European trackers show graphics card prices up thirty to forty percent since June.

Alex: Neither A M D nor Intel has confirmed the hikes, so file this under reports say, not settled fact. But if you were planning a P C build or a server refresh, the trend line is not your friend.

Jordan: And now, a word from the people who make sure your computers don't become expensive paperweights.

Alex: Folks, let me tell you about at Your Service Techs. Thirty years. Thirty years of keeping computers, servers, networks, websites, and surveillance systems running while the rest of us would just be crying into our keyboards.

Jordan: When your server decides to take an unscheduled vacation at two in the morning, who you gonna call? at Your Service Techs, baby! In Grand Rapids, Michigan, call six one six, five three five, two zero two two. In Phoenix, Arizona, call six zero two, two three two, six zero six zero.

Alex: Or visit phxfix dot com. at Your Service Techs, because your tech problems don't fix themselves.

Jordan: We now return to your regularly scheduled questioning of everything.

Alex: Health segment time, and this one's a doozy. The question lighting up the internet this week, can weight-loss drugs slow down aging itself?

Jordan: This comes from a Nature study out of U C Berkeley, led by Danica Chen. They gave semaglutide, the drug in Ozempic, to old female mice starting at twenty months of age, roughly the human sixties, and treated them for life. Median lifespan went up about twelve percent, from seven hundred forty-two days to eight hundred thirty-four days.

Alex: Not just living longer either. Better strength, coordination, memory, and glucose tolerance, and even new neuron growth in the hippocampus, the brain's memory center.

Jordan: Now the claim-evidence-unknowns treatment. The mice ate about twenty-four percent less, because that's what the drug does. But a matched food-restricted group didn't get the same cognitive gains. So the researchers say there's something going on beyond just eating less.

Alex: Evidence level check. It's a mouse study, about twenty mice per group, females only, because the males were reportedly too aggressive to house together. Mouse lifespan is not human lifespan, and the authors stress this does not show the drugs extend human life.

Jordan: The human evidence so far is surrogate markers. A U C San Diego trial of one hundred eight adults found thirty-two weeks of semaglutide slowed epigenetic aging clocks by about nine percent. Interesting, but a methylation clock is not a mortality curve.

Alex: And a fascinating companion story. At a diabetes conference this week, Regeneron presented phase two results for trevogrumab, an anti-myostatin drug, which prevented about seventy percent of the lean mass loss that usually comes with G L P one weight loss. So the emerging playbook might be, the weight-loss drug handles the fat, and a companion drug protects the muscle.

Jordan: Still phase two and announced by the company, so treat it as promising early data, not a therapy you can ask your doctor about. But the muscle angle is exactly what Dave Asprey and Gary Brecka have been shouting about for years. If you lose weight and the weight is muscle, you've traded one problem for another.

Alex: Next up, a study that dropped today. Virginia Tech researchers fed fifty-seven adults two sets of meals matched to within one percent on calories, carbs, fat, protein, water, and salt. The only difference was processing, ultraprocessed versus minimally processed.

Jordan: Same calories on paper, different bodies in practice. The ultraprocessed meals triggered a bigger insulin response, higher energy expenditure with less carbohydrate burning, and different activity in the brain's reward centers, the ventral striatum and nucleus accumbens, measured on brain scans.

Alex: Randomized crossover design, published in Nature Metabolism, so every person ate both meal types. Caveats, one set of three-hundred-calorie meals in young healthy subjects, and the researchers can't yet say which processing step or additive drives the effect.

Jordan: Still, it dents calories-in-calories-out absolutism. Dr. Mark Hyman has said for years that food is information, not just fuel, and this is the kind of study that starts to show why.

Alex: Two quick hits. A World Health Organization analysis in The Lancet Oncology estimates infections cause about one in eight cancers worldwide, around two point three million cases in twenty twenty-four, and many are preventable with existing vaccines, screening, and treatment.

Jordan: And a nod to one of our trusted voices. Dr. Daniel Amen is taping a new special called Brain Rescue on October seventh, drawing on his clinics' database of more than three hundred thousand brain scans. It's a P R announcement, not a study, so we'll treat the scan database as practitioner experience, not trial evidence.

Alex: And that's the show. Big week in tech, big questions in health, and as always, we don't have all the answers, just better questions.

Jordan: So here's our sign-off, and we mean it more every week. Question everything.

Alex: Including us, especially us.
Question everything — including us… especially us.