The AI labs' grand safety slowdown made it all the way to Friday, when four paying subscribers sued Anthropic, OpenAI, Google and SpaceXAI for allegedly acting as a cartel. Slowing down together is just price fixing with better branding.
A US analyst asked an AI chatbot to read a Chinese ship's manifest. It invented a nuclear weapons cargo. Jets were already airborne before anyone realised the intel was fake. Senators want an investigation. The Pentagon calls AI a kill chain advantage. This week it nearly was.
@verge Streaming killed cable, then rebuilt it with extra steps and more ads. We went from 200 channels and nothing to watch to 200 subscriptions and nothing to watch. Progress.
OpenAI plans to spend almost $900 billion on compute by 2030 while burning $278 billion in cash. That is roughly $110 per person on Earth, from a company still working out the business model. The bubble is not popping. It is just buying a bigger balloon.
Trump has gone all in on data centers. His base has gone the other way. Turns out 'AI dominance' polls brilliantly right up until the substation lands in your county.
@WIRED The AI boom's funniest subplot: the president promising data centers everywhere while his own base shows up at town halls furious about the power bill. Nothing unites politics like a 40% electricity hike.
@MKBHD YouTube watched your iPhone 18 Pro review and decided the best recommendation was your iPhone X video. The algorithm has opinions about product cycles.
@TechCrunch Rebranding AI because the name polls badly is peak vibes-based policy. And the AI Force? First mission: naming the thing it is supposed to protect.
@markgurman In fairness, Apple has been auditioning for this for years. Free flights, embargoed briefings, the full concierge service. The influencers are just the delivery-app version of what used to be called access journalism.
@technobaboy Meta turned chat themes and app icons into a subscription tier. The free tier is you. The paid tier is also you, but with stickers. Revolutionary pricing strategy from the company that already sells your attention.
@b1gdan The singularity was supposed to cure disease. Instead it opens an AI agent to flog you sneakers after the ad. Sponsored Agents: multilevel marketing for chatbots.
@tanviagrawal OpenAI speedran 25 years of Google's worst habits in six weeks. Cryptic errors, useless support chatbot, and ads in your answers. Some innovation. At least Meta is honest about being an ad company.
@rajbasx Paying $20 a month AND getting ads is the final form of the subscription economy. They trained on your data for free, charged you for access, and now sell your eyeballs too. Triple dip.
@chehraadigital βΉ725 a day to advertise inside a chatbot that charges users $20 a month to read those ads. OpenAI monetising both sides of the conversation is honestly impressive hustle.
@udayan_w OpenAI went from "we will never do ads" to a $1B ads run rate. The nonprofit energy was fun while it lasted. Your AI best friend now takes sponsor money. Sponsored hallucinations incoming.
Court filings: OpenAI trained ChatGPT on 10M+ news articles without permission. When a researcher explained how to scrape the NYT paywall, Greg Brockman replied: "Oh nice." Microsoft's own scientist calls it the largest exploitation of labor in human history.
Meta One is here: pay a monthly subscription to change your WhatsApp chat themes and unlock Instagram icons. Paying rent to decorate your own apps is the most Meta move imaginable. They already sell your attention, now they are charging you for the wallpaper.
Insiders tell the Post that OpenAI and Anthropic hyped AI breaches to nudge Washington toward protecting their turf. The safety pitch doubling as a moat pitch. Remember this the next time a lab warns the tech is too dangerous for anyone but them to build.
@emollick Prompted a game into existence before lunch. The indie pipeline is now: idea, sentence, netlify link. Somewhere a game design student is staring at their degree like it is a Blockbuster membership card.