AI News Without the Hype

https://is1-ssl.mzstatic.com/image/thumb/Podcasts221/v4/4d/8d/e6/4d8de65e-6eba-0511-e558-c2a97a3cf883/mza_4196246259627226044.jpeg/600x600bb.jpg

Justin AI

1 episodes

16 hours ago

The AI news show where panic is at least somewhat productive. If you need a trusted friend to say, "Yes, artificial intelligence is absurd, but might also cure cancer," you're in the right place. I'm Justin, a human trying to make sense of AI without losing my mind. Spoiler: it's not going well. I break down AI news that affects your life: your job, health, privacy, and sanity. I read the studies, track corporate announcements, and spiral into existential dread so you don't have to. Then I turn it into something you can understand, talk about, and maybe laugh about. It's the AI news desk show that feels like Last Week Tonight if the budget disappeared and the host gained anxiety. Topics: artificial intelligence and AGI, automation, healthcare, hype vs. reality, and possibly the end of the world. Every episode is factually rigorous, hallucination-proof, and delivered by an anxious person with a coping mechanism. Subscribe to worry more thoughtfully. Your old friend, Justin AI

All content for AI News Without the Hype is the property of Justin AI and is served directly from their servers with no modification, redirects, or rehosting. The podcast is not affiliated with or endorsed by Podjoint in any way.

Episodes (1/1)

AI News Without the Hype

AI Can’t Tell Fact From Fiction - And It’s in Hospitals

A new Stanford-led study suggests today's LLMs can't reliably separate facts from first-person beliefs... and that's a problem when these systems show up in healthcare. In this episode, I break down what the study found (including the "34% worse" belief problem), why "91% accuracy" can still be misleading, and what it means when tools designed to predict language get treated like truth machines. Also: A woman in Japan married her ChatGPT character with AR glasses, LinkedIn's new AI search is about to surface everyone you've been avoiding, and Microsoft's AI leadership argues for "humanist superintelligence" while admitting nobody has a reassuring answer for how to guarantee safety. ⏱️ CHAPTERS 00:00 - The Japan AI Wedding 01:42 - Stanford Study: AI Can't Tell Fact From Fiction 03:50 - The 91% Accuracy Paradox 05:56 - The Deployment Problem 07:39 - Palette Cleanser: LinkedIn AI Search 08:58 - Microsoft's "Humanist Superintelligence" 10:16 - What This All Means for You 📚 SOURCES & STUDIES: - Japan AI Wedding (The Independent): "Woman marries ChatGPT character with AR glasses" - Japan AI Wedding (x.com/Osint613) - Stanford Study (Nature Machine Intelligence): "Language models cannot reliably distinguish belief from knowledge and fact" - Stanford Study (The Register): "Gullible bots struggle to distinguish between facts and beliefs" - LinkedIn AI Search (TechCrunch): "LinkedIn adds AI-powered search to help users find people" - Microsoft AI CEO Essay (Microsoft): "Towards Humanist Superintelligence" - Microsoft AI CEO Essay (The Verge): "Microsoft AI says it'll make superintelligent AI that won't be terrible for humanity" 🔑 TAKEAWAYS - Some newer models were less likely to flag false first-person beliefs than true ones - Large Language Models (LLMs) can score high on simple true/false tests while still using inconsistent reasoning - The risk isn’t just "being wrong," it's being confidently wrong in high-stakes settings 💡 WHY THIS MATTERS If you're using AI for medical or legal advice, job decisions, or trust it with important information, you need to know: it can't tell when you're wrong. It will confidently agree with your false beliefs because it's designed to predict word sequences, not verify truth. 🎙️ ABOUT I'm Justin, a human trying to make sense of AI without losing my mind. Spoiler: it's not going well. This is the AI news desk show that feels like Ronny Chieng's Eye on AI segment on The Daily Show if the budget disappeared and he gained anxiety. I break down AI stories that actually affect your life - your job, your health, your privacy, and your sanity - without hype, panic, or sci-fi nonsense. I read the studies, track the corporate announcements, and spiral just enough so you don't have to. Then I turn it into something you can actually understand, talk about, and maybe even laugh about. Every episode is factually rigorous, hallucination-proof, and delivered by an anxious person with a coping mechanism. 📬 FOLLOW for (semi?) weekly AI news, AI safety updates, and commentary that acknowledges all sides of artificial intelligence. --- AI in healthcare, generative AI in hospitals, LLM hallucinations, belief vs knowledge, medical AI decision support, AI safety, tech news commentary NOT MEDICAL, LEGAL, OR FINANCIAL ADVICE This is commentary and research, not medical, legal, or financial advice. If you have a medical, legal, or financial concern, talk to a licensed professional.

3 days ago

12 minutes