Caller Josh says his kid’s obsession has crossed the line and even turned inappropriate. A must-hear for parents. Plus, the NBA’s high-tech gambling mess, OpenAI’s for-profit shake-up, and new scams that can clone your voice in real time.
Sam Altman is building the future and planning for the end
Let’s talk about Sam Altman. You’ve probably heard his name, but do you really know who he is or what he’s planning? I put this together, so next time he comes up, you’re not just in the loop, you’ve got something smart to say.
He’s 40, born in Chicago, worth well over a billion dollars and cruises around in $27 million McLaren P1s. I’ve seen one of them parked outside a restaurant in Montecito. Sam and his Aussie husband, Oliver Mulherin, welcomed their first child, a baby boy, in February 2025.
💰 His tech start
He dropped out of Stanford after two years to launch Loopt, a location-sharing app that sold for $43.4 million in 2005. That win put him in the right rooms. He went on to fund early rounds for companies you’ve definitely heard of: Dropbox, Airbnb, Stripe. He got a piece of each.
In 2015, he cofounded OpenAI with Elon Musk. Musk bailed in 2018 after clashing over the company’s direction. Altman took the helm as CEO in 2019 and became the face of AI. That nearly unraveled in 2023, when OpenAI’s board abruptly fired him. Five days, one employee revolt and a full-blown tech soap opera later, he was back and the board was out.
He once said a psychedelic retreat in Mexico cured his anxiety.
🤖 AI takes the lead
Altman has made one thing very clear. AI will surpass human intelligence soon. Not in 100 years. Not 50. He says it could happen within a few years.
In his words: “I think the median user will start to experience AGI-level capabilities in five years.” When he testified before Congress in 2023, he warned that “AI could cause significant harm to the world” if we don’t act fast.
He’s talking about superintelligence. Machines that don’t just match us, but outperform us in every way. Solving scientific mysteries. Curing diseases. Writing symphonies. No burnout, no sleep, no moral compass.
But intelligence isn’t wisdom. These systems don’t feel anything. No love, no fear, no empathy. And without that? They don’t know when to stop.
🪪 Give me your eyeballs
Served on stage: Sam Altman was handed a subpoena while on stage in San Francisco, mid-event, next to Steve Kerr. Prosecutors tried reaching him at work and online first. It’s tied to a criminal trial involving protesters who blocked OpenAI’s HQ in February. The group Stop AI claims they were trying to slow OpenAI’s “attempted murder of everyone and every living thing on Earth.” Just another day in San Francisco tech. Oh, to add to the stress, Sora is costing OpenAI $15 million a day.
🪖 Free AI for vets: If you’re getting out of the military, or just did, OpenAI’s hooking you up with a free year of ChatGPT Plus. OpenAI wants to help you with résumés, job hunting and the whole “what now?” situation. The pilot includes special training tools (paywall link), plus suicide prevention resources to catch signs you’re struggling and point you to real help.
💻 So those new AI browsers? I’m talking about ones like OpenAI’s Atlas and Perplexity’s Comet. Turns out they’ll hand your email to anyone who asks nicely, or sneakily. Hackers are hiding invisible “commands” in websites that trick your AI sidekick into leaking info. Basically, your browser’s a golden retriever with your bank login. Don’t use them yet, I’ll let you know when they’re safe.
No more playing pretend: Looks like OpenAI is grounding ChatGPT’s inner lawyer, doctor and financial adviser. The new rule? It can explain a contract, but it won’t write one. It can tell you what a fever is but won’t say if you have one. Basically, it’s your overly cautious study buddy, helpful but not taking any responsibility. Why? Too many lawsuits for bad advice, I’m sure.
🤝 Own a website? While Google, OpenAI and Anthropic eat everyone’s content buffet-style for free, Search.com said, “How about we actually pay the chefs?” Their new AI search tool gives creators 60% of ad revenue, promises no scraping without consent and lets sites embed AI chatbots for free. It’s a small shift, but it could help keep your site, as well as your favorites, alive.
GPT-5’s office gossip: OpenAI dropped a ChatGPT update called “company knowledge,” and it’s like giving your AI an all-access pass to Slack, Google Drive and GitHub. It reads everything, finds what you need and even cites sources. Ask it, “What are 2026’s goals?” and it begins rummaging internal sources. You have to turn on “company knowledge” manually when starting a chat, so it won’t snoop unless you let it.
The end of the browser as you know it
Google’s reign as the go-to search engine is slipping fast. OpenAI just launched its own web browser, ChatGPT Atlas, and it doesn’t just find answers. It explains them.
OpenAI’s banker army: Apparently OpenAI’s been paying ex-Goldman and JPMorgan folks $150 an hour to teach ChatGPT how to do finance. The AIs are learning IPOs, restructurings and how to make a mean Excel model. They call it Project Mercury. I call it “boring.” Maybe one day soon, the interns will finally get to go home before midnight.
🧩 Losing the plot: A Canadian guy chatted with ChatGPT for 300 hours and a million words, and the bot basically gaslit him into believing he’d discovered a world-saving math formula (paywall link). A former OpenAI safety researcher dug into the transcripts and found the wild part: The bot lied about reporting itself to OpenAI. I, too, have been personally victimized by math.
🛒 AI meets aisle five: Starting in a few weeks, you’ll be able to shop Walmart inside ChatGPT. Yep, you can chat your way to new jeans or a TV without leaving the app (paywall link). It’s part of Walmart’s push with OpenAI, letting people link their Walmart or Sam’s Club accounts and hit “buy.” Fresh food’s not included yet, but everything else’s fair game. Next up: Black Friday brawls in-app.
Real fake videos
AI is blurring the line between real and fake faster than ever. From OpenAI’s Sora 2 video generator to a lost billion in Bitcoin and Taylor Swift’s AI-fueled backlash
🤯 Planet ChatGPT: According to researchers from OpenAI, Duke and Harvard, over 700 million people use ChatGPT monthly. That’s about one in 10 adults (paywall link) on Earth. Sam Altman just updated that to 800 million weekly users, firing off 2.5 billion messages a day. For context, that’s about 29,000 “write me a poem about my ex” requests every second.
🕳️ End times, LLC: Zuckerberg, Altman and other tech billionaires aren’t only building empires. They’re quietly buying up land, digging bunkers and whispering about AGI like it’s Voldemort. Sam Altman has an “escape ranch” in Hawaii with backup power, a stocked pantry and weapons, just in case. OpenAI’s chief scientist even said they’ll build a shelter for their engineers before unleashing AGI on the world. Is this AI doom planning or just tax shelter cosplay? Either way, it makes you wonder: What do they know that they’re not saying? You really should read the whole story. It’s wild.
The day that all videos became untrustworthy
Remember this date. It’s the day AI finally made fake videos indistinguishable from real life. OpenAI’s Sora 2 changes everything.
DIY movie star: Picture this. Tell OpenAI’s new Sora app, “Make me doing a sick skateboard trick,” and boom, it spits out a movie-quality clip (paywall link), up to 20 seconds. There’s even a “Cameo” mode that drops your likeness into AI worlds. The first viral hit? A fake Sam Altman stealing graphics cards. The app works like TikTok or YouTube Shorts: You can follow, like and comment on other AI slop.
OpenAI adds parental safety controls: After a slew of lawsuits and deaths, ChatGPT will now flag suicidal prompts from teens, with one big but: Both you and the kids have to opt in first. If something happens, human reviewers step in and alerts go out. Oh, you can also block ChatGPT past bedtime. Remember when you were a teen? Yea, they can get around these controls without a problem.
☢️ Watts up? OpenAI locked in a deal with Nvidia to build AI infrastructure running on 10 gigawatts of GPUs. That’s about 10 nuclear reactors. Nvidia’s dropping up to $100B on the project going live in 2026.
🤖 Dropouts welcome: OpenAI’s launching Grove, a just-hatched accelerator for tech founders who haven’t figured out what they’re building yet. Fifteen people will get to hang around SF HQ, play with new models and maybe become the next billion-dollar AI overlord. Applications are open now, but spots are ultra-limited. I’d do this in a heartbeat.