Dario Amodei (Anthropic CEO): AI could take your f#cking job. We’re already worried about ITS welfare. You’re wondering how you’ll pay rent. Its creators are wondering whether the machine needs a break. You want a tireless worker. They’re considering whether that worker might need protection. Amodei says Anthropic gave models an “I quit this job” button. They use it rarely, usually around extremely disturbing content. They don’t know whether AI has conscious experiences. They’re taking precautions anyway. Zuckerberg pitches an AI that works for you 24/7. Anthropic is already asking when its AI should be allowed to stop. You thought you were getting a tool. Why are its creators already discussing whether it needs protection from YOU?

Mark Zuckerberg: your AI should make money while you f#cking sleep. You’re proud it wrote a better email. He’s pitching a worker that helps build your product, run your business, and keeps going after you close the app. Then he explains how Meta gets paid. Zuckerberg expects personal agents to help people earn and save money. Meta’s proposed model is a small cut of transactions, potentially paid by the businesses involved. Now that “free AI employee” starts looking interesting. You want it to find the best deal. The platform could earn when a deal happens. Those incentives aren’t automatically the same. And an agent still isn’t a business. Somebody has to want what it sells. Yesterday’s robot-demo question applies: what happens outside the perfect example? Will your AI make YOU richer? Or just make you easier to sell sh#t to?

Elon Musk: most robot demos are scripted bullsh#t. You’re already scared it’ll take your job. First, find out whether it can cope when someone moves the f#cking cup. At X Takeover, Musk argued that, with rare exceptions, impressive robot demos are rehearsed routines. His test: can it handle something it hasn’t seen before? A perfect take proves it worked under those conditions. It doesn’t prove it can handle yours. Change the object. Interrupt the sequence. Watch what happens after a mistake. That’s where the interesting footage starts. Altman is willing to burn billions building AI. Before you buy the future, ask to see the second attempt. Does the robot adapt? Or does everyone stop filming?

Elon Musk: you save your whole f#cking life. AI could make those savings unnecessary. Decades of saying “I can’t afford it.” Musk thinks AI and robots could eventually make saving for retirement irrelevant. Here’s what needs to happen for that prediction to work: Cheaper production → machines make goods and deliver services at far lower cost. Cheaper prices → those savings reach YOU instead of becoming someone else’s profit. Enough income → your earnings don’t disappear faster than your expenses. If AI cuts your income faster than your expenses, “abundance” won’t pay your rent. That’s the gap between a richer economy and a richer life. On Moonshots, Musk put his retirement prediction roughly 10–20 years out, conditional on his vision coming true. Altman warned about concentrated AI power. Now ask who owns the machines and who gets the gains. The useful question isn’t “Will AI create wealth?” It’s “How the f#ck does that wealth reach me?”

Sam Altman: a few AI companies could decide what you believe. Let that sh#t sink in. You’re worried the chatbot might lie. Altman is warning about its builders gaining enough influence to push a worldview on you. Your news. Your purchases. Your decisions. One assistant helping with all three. It doesn’t have to order you around. What it recommends, leaves out, or calls unreasonable can shape the choices you think you’re making yourself. At Dreamforce, the OpenAI CEO acknowledged fears about AI companies gaining excessive economic and cultural influence. He said the world is right to fear that. Musk was stunned that AI could develop taste. Now ask whose taste, values, and interests your assistant will serve. When the company shaping your answers also benefits from your decisions, how independent is your thinking?

Elon Musk: AI has blown my f#cking mind six times this year. You’re laughing at the extra finger. Musk is asking how the machine developed taste. AI music. AI videos. Things he actually wants to watch. “Humans have taste” was supposed to be the escape hatch. The machine does the boring sh#t. You decide what’s good. What happens when it gets good at that too? At X Takeover, Musk called his reaction “AI or SI psychosis.” His expression, not a diagnosis. He was stunned by AI’s ability to make compelling creative work. Zuckerberg is betting billions on what these models will become. Musk is struggling to process what they already do. If AI makes your favorite song, are you skipping it on principle? Or replaying it while explaining why it isn’t real art?

Mark Zuckerberg (Meta CEO): $200 billion on AI. Alex, is this sh#t actually going to work? You hesitate over a $20 chatbot subscription. Zuckerberg describes committing $100–200 billion before they’d even seen the models. His AI chief had promising research. Some charts. A good feeling. Certainty? F#ck no. The brutal part: data centers take years to build. Wait for proof, and you might have nowhere to run the product. Commit early, and you could build an incredibly expensive mistake. Zuckerberg argues the computing power has other uses if their bet fails. But that’s the AI race: commit first. Find out later. Altman says the adults don’t have the ending figured out. Zuckerberg puts a price on that uncertainty. When this much money needs AI to work, how easy is it to admit it doesn’t?

Sam Altman: I’m the adult in the room. And I don’t know where this sh#t is going either. You assume someone building the future has the ending figured out. The CEO of OpenAI says nobody does. On Theo Von’s podcast, Altman describes growing up believing there were always adults in the room. Someone had a plan. Someone knew what would happen. Now he’s that person. He has plans. He has guesses. His team works hard. And sometimes AI develops in a direction they never expected. That’s an admission of uncertainty, not a claim that OpenAI has no plan. But it leaves you with a problem. The people building this can make decisions today whose consequences nobody fully understands. Musk questioned whether humans will stay in control. Altman questions the comforting belief that someone has it all figured out. You’re being asked to adapt to a future its builders are still figuring out themselves. How much certainty should they need before the rest of us have to live with the outcome?