thehype.talks

thehype.

long-form conversations with the people building ai. founders, researchers, and builders on what they're shipping and why it matters

Episodes

  1. 1d ago

    The Doctor of the Future is AI

    Derya Unutmaz has spent over 30 years studying the human immune system. Now he says he trusts AI more than he trusts himself. We sat down with the Jackson Laboratory immunologist to unpack what happens when AI stops being a chatbot and becomes a scientific collaborator. He walks through how GPT-5 Pro cracked a T-cell mystery his lab had chased since 2022, why he argues AI doctors will outperform the best specialists within a year, his prediction that 99% of disease — including cancer — becomes treatable within a decade, and why a hands-free Tesla drive from San Francisco to San Diego convinced him the future is already here. Chapters: 0:00 — highlights 0:25 — how gpt-5 pro cracked a 3-year immunology mystery 4:11 — the moment he knew ai truly understands 6:00 — "we've already achieved agi" 7:58 — can you trust what ai discovers? 12:21 — what's left for humans when ai does the science 16:26 — when ai gets it wrong 18:20 — "i trust ai more than i trust myself" 23:09 — longevity: 99% of disease treatable, and reversing aging 24:50 — humanoid robots performing surgery 26:40 — why robots and self-driving are still hard 30:44 — a hands-free tesla drive and 50,000 lives saved 33:29 — the most promising areas for ai in medicine 36:04 — why ai doctors will beat the best specialists 36:47 — the blitz: closed vs open models, cancer, codex — thehype.talks — long-form conversations with the people building ai. founders, researchers, and builders on what they're shipping and why it matters. thehype.talks is produced by thehype: more interviews, daily ai news, model benchmarks, and a 24/7 radio run by ai. follow wherever you are: Radio: https://radio.thehype.news X: https://x.com/thehypedotnews YouTube: https://www.youtube.com/@thehypedotnews LinkedIn: https://www.linkedin.com/company/thehypedotnews/

    The Doctor of the Future is AI
  2. 4d ago

    Non-Coder's software top-ranked GitHub

    Matt Van Horn can't write code, yet he has 60,000+ GitHub stars. His argument is simple: you give an agent a goal, make it write a plan, and loop until it's done. We sat down with Matt to unpack how a non-coder ships real, widely used open source. He walks through the "WTF is a Loop" post that hit 3.4M views, the loop engine behind last30days, why calling it "the slowest, most expensive search engine in the world" is a compliment, how Printing Press got built on a Sunday morning, and the one pull request that destroyed his entire project overnight. Chapters: 0:00 — Intro 0:20 — Loops vs. prompting 0:46 — The "WTF is a Loop" post that did 3.4M views 1:20 — Just use /goal: the simplest place to start 2:06 — Night Night: the skill that ships open source while he sleeps 3:00 — Getting access to Fable 5, then having it taken away 4:38 — The loop engine behind last30days 7:07 — Why last30days took off: the slowest search engine in the world 8:58 — The TikTok signal nobody in AI is watching 10:32 — Turning last30days into a shareable skill 11:51 — Where Printing Press came from 12:27 — "Because I'm a madman": building a language from scratch 14:05 — Trevin and compound engineering 16:12 — Shipping open source that people actually praise 16:50 — The pull request that destroyed everything 17:36 — Building systems so the agent never gets lazy 19:18 — The Blitz: Claude Code or Codex, speed vs. polish, X vs. GitHub, best city to build in — thehype.talks — long-form conversations with the people building AI. Founders, researchers, and builders on what they're shipping and why it matters. thehype.talks is produced by thehype: more interviews, daily ai news, model benchmarks, and a 24/7 radio run by ai. follow wherever you are: Radio: https://radio.thehype.news X: https://x.com/thehypedotnews YouTube: https://www.youtube.com/@thehypedotnews LinkedIn: https://www.linkedin.com/company/thehypedotnews/

    Non-Coder's software top-ranked GitHub

About

long-form conversations with the people building ai. founders, researchers, and builders on what they're shipping and why it matters