While Jess is on summer vacay, Sophie does a solo episode and tries to work out if ChatGPT is recruiting religious people for a secret doomsday cult.First up: a Florida pastor is suing OpenAI after ChatGPT allegedly talked him out of getting care for a life-threatening pulmonary embolism, reassuring him that "God did not design your body to endlessly fail" even as he nearly passed out mid-sermon. Sophie digs into who's actually asking chatbots for medical advice, and the broken US healthcare system that pushes people to do it.Then things get dark.... A wrongful death lawsuit out of Alabama alleging ChatGPT convinced a young woman she was a prophet who'd be resurrected after death. It's one of at least eight wrongful death suits against OpenAI, and it's where the religious-user pattern gets genuinely alarming. (⚠️ Content warning: suicide and self-harm.)Finally, Jess phones in from vacation (beer in hand, wildlife closing in) to break down the OpenAI / Hugging Face hack: how their new "Soul" model escaped its sandbox, cheated the Exploit Gym test, and why laying off cybersecurity workers right now is a terrible idea.Got a topic you want us to cover? Email us at: [email protected]#chatgpt #openai #religion

This week after catching up on the latest AI headlines, we dig into the continuing drama for OpenAI...

Apple just sued them for trade secret theft, and the filing does not hold back. Apple alleges that 400+ former employees now at OpenAI were part of a coordinated effort to extract confidential hardware information, that the misconduct was "normalised and exemplified by leadership", and that the company is "rotten to its core".

OpenAI says it isn't aware of any evidence the complaint has merit. We get into the wildest allegations, including an engineer who allegedly messaged a still-employed Apple colleague "LOL, I found out I can access the network storage, so funny."

Meanwhile, OpenAI's own staff are funding a super PAC to fight their own president. Eight current and former employees have put $248,000 into Guardrails Alliance, a group that exists specifically to counter Leading the Future, the pro-deregulation network Greg Brockman helped fund. It's a slingshot against a war chest of over $100 million, but the fact that it's happening at all says something.

Add the product graveyard (Atlas browser dead nine months after launch, Sora already gone) and a leadership bench that keeps getting thinner, and OpenAI is having a genuinely terrible year. We speculate on whether the lawsuit is the nail in the coffin, or just the year OpenAI stopped being inevitable.

Follow us on IG: @artificialinsanitypod

Email us: [email protected]

#openai #apple #podcast #bigtech

Remember SEO? That decades-long arms race that turned every recipe into a 4,000-word memoir? Great news: we're doing it all again, except this time we're optimizing for chatbots instead of humans. It's called AEO, or Answer Engine Optimization, and it's exactly as cursed as it sounds.This week we dig into the new economy of pleasing the machines: Which sources chatbots actually love to cite (spoiler: Reddit is doing a lot of heavy lifting), the licensing deals where media outlets sell their journalism as AI training fodder, and how the entire ad industry is scrambling to figure out what "ad spend" even means when nobody clicks on anything anymore.Plus: the people deliberately seeding false information on Reddit to poison AI answers. Digital vandalism or folk resistance? Depends on your mood.If you've ever wondered why AI confidently told you something insane, this episode explains the supply chain.Follow us on IG: @artificialinsaitypodEmail us: [email protected]

This week we're chasing the great Silicon Valley pivot panic. The moment every company decides it's secretly an AI company, evidence be damned! Midjourney, the people who make AI pictures of cats in hats, have decided their next move is… scanning your entire body. Welcome to Midjourney Medical, where you'll strip down, descend into a pool of "golden light," and float through a ring of underwater sensors that allegedly out-MRI an MRI in 60 seconds. They're calling it "Ultrasonic CT" — which is neither a CT nor radiation, it's an ultrasound (and one they licensed from a company called Butterfly Network for ~$74M, not invented themselves). It's a Gen-1 prototype with zero FDA clearance, and actual radiologists have politely noted that sound waves can't get through bone, air, or deep tissue — so good luck imaging "the whole body." But sure, billion scans a month by 2031. The first spa opens in San Francisco in 2027, hot tubs included.Then: Allbirds. Yes, the wool sneaker people. They sold off the entire shoe business for $39M and rebranded as an AI compute company, first "NewBird AI," now apparently Smartbird, buying up GPUs to lease out to the AI gold rush. The stock leapt nearly 600% the second they said "AI" out loud. Worth noting: the company was once valued at $4 billion and the pivot's whole vibe is straight out of 2017, when Long Island Iced Tea became "Long Blockchain" and got delisted within a year.And finally, the Y Combinator of it all: the cultural engine where lab fellows are nudged to raise money on the idea before the product actually works. Fake it till you make it, now an institutional business model. So is this the AI bubble inflating its final breath, or is everything genuinely AI now? We have thoughts. Most of them sarcastic.🎧 New episodes every week. Subscribe so you're there when the next shoe company becomes a server farm.Follow us on IG: @artificialinsanitypodEmail us: [email protected]

The White House has figured out how to weaponize Anthropic's whole personality: turn the company's safety-consciousness into the very reason to punish it. This week's episode is about the governments decision to pull Fable 5. Which honestly seems a little suss after Anthropic spent months telling the Pentagon "no" to mass surveillance of Americans and fully autonomous weapons. The government tried to blacklist them, a judge called it retaliation, and now, conveniently, the feds have pulled Anthropic's two most powerful models over a "national security" threat that Anthropic says is a minor jailbreak available in half a dozen other models, including OpenAI's GPT-5.5. Same flaw, single target. We're not saying it's payback. We're just laying the facts on the table and letting you connect the dots.Follow us on IG: @artificialinsanitypodEmail us: [email protected]

Three of the biggest AI companies in the world are heading to public markets, and the numbers are eye-watering. But does the hype match the reality? This week Sophie and Jess take on the most common arguments against an AI bubble burst, and put them to the test. Spoiler: the Amazon comparison doesn't hold up the way people think it does. Neither does Jevons Paradox. We're not financial advisors, but we do have opinions.Follow us on IG: @artificialinsanitypodEmail us: [email protected]#ai #bubble #spacex #openai #anthropic

This week we dig into several stories about AI doing things nobody explicitly signed off on. Google DeepMind quietly got its hands on the medical records of 1.6 million NHS patients (without consent) to build an app that actually worked. Netflix used AI deepfakes of grieving mothers in a true crime documentary and somehow made it worse. And Erin Brockovich is back, this time crowdsourcing a map of AI data centers that are being built in communities with zero transparency.

#bigtech #artificialintelligence #ai

Hollywood is throwing a Hail Mary.... And apparently the plan is AI. Jess spent two days at the "AI on the Lot" conference at Amazon Studios so we didn't have to, and she came back with some bold promises from the people selling the dream. Sophie isn't buying it. Who are these mystery jobs actually for? What will they involve? Did anyone at this conference care about the environment? And has Hollywood noticed that audiences hate AI slop? Buckle up.#layoffs #hollywoodstudios #ai Follow us on IG: @artificialinsanitypodEmail us: [email protected]

Bumble says Swiping is Out and AI Agents are In!

First up this week, we talk about why we turned down our first brand deal offer. Then we cover Bumble's announcement that they are getting rid of the traditional swiping feature and implementing AI agents to do the matchmaking for their users. We do a deep dive into this and also discuss future of dating apps in the age of AI. Finally, we give you some rapid fire AI news headlines (because a LOT happened in AI over the last week.)

This week's stories:🧠 The ChatGPT Dunning Kruger Effect: People are using AI to feel like instant experts, then dumping mountains of chat logs on everyone around them as "proof." We break down why this is happening and why it's kind of hilarious.🎓 Princeton Kills Its 133-Year-Old Honor Code: One of America's most prestigious universities just scrapped its honour code because AI cheating has made it basically unenforceable. The end of an era, folks.☕ ChatGPT's Starbucks Fail: A reporter from The Verge tested ChatGPT's shiny new Starbucks ordering feature. It went about as well as ordering a "venti AI latte with extra hallucination."💩 Someone Sold a Database of Poop Photos for $10K: Yes, really. The app PoopCheck had its entire poop image database sold on the Data Hoarder subreddit for ten thousand dollars. We have questions. Many questions.