• Amazon committed $50 billion to build AI infra for U.S. government agencies, so the FBI agent who used to secretly listen to your phone calls just got promoted to “Senior Automation Strategist.” — CNBC

• President Trump’s Stargate project is taking shape, OpenAI + Oracle + SoftBank are joining in to the project named after a show where alien DevOps deploy hotfixes through interdimensional SSH— OpenAI

• Luma AI raised $900 million led by a Saudi firm to expand its generative videos of camel races or something idk — Luma Labs

• Harmonic, the anti-hallucination AI startup from the Robinhood guy, is now valued at $1.5 billion, and somewhere in D.C. the IRS just whispered “not this guy again” and reopened a few Google Docs. — Yahoo Finance

• Elon Musk is pushing xAI valuation to $230 billion, which feels less a valuation and more a dare. — NYPost

• NVIDIA has committed to rent $26B in cloud servers in the next 6 years, till then i vision sales reps walking around Silicon Valley like: “psst, you want two H200s?” — The Information

• Cursor the vs code fork, raised $2.3 billion to build its own model,Composer the Claude Sonnet fork hahahahah (*crowd goes booooooooo* relax! relax! its a joke i know we all love cursor ❤ ) — TechCrunch

• HP is “integrating AI into operations” which is corporate for “bye everyone”, no joke, they’re going to cut 4,000–6,000 employees by 2028 — Fox Business

• Anthropic dropped Claude Opus 4.5 and swept every benchmark, and now Google’s Gemini Pro 3 is in the corner like that Pakistani fan meme — Anthropic

• Ex–MrBeast Jay Neo just launched Palo, a $250/mo AI tool exclusively for creators with 100k+ followers, it analyzes their videos, mimics their style, and helps them recreate the formula they clearly already mastered. — Business Insider

• OpenAI introduced ChatGPT Shopping Research, an AI that shops like your mom, asks you 40 questions, judges every choice you make, then picks the thing you should’ve bought in the first place. — OpenAI

• Perplexity also launched Perplexity Shopping, turning in ChatGPT’s worksheet word-for-word like the teacher wasn’t going to notice, cmon guys— PCMag

• Google is preparing Aluminium OS, a Frankenstein ChromeOS built on top of Android then patched into PC with Gemini AI shoved into whatever space is left. — The Verge

• OpenAI introduced ChatGPT Study Mode and ChatGPT for Teachers, education tools built for a school system so outdated that teachers now use AI to create homework and students use the same AI to solve it, completing the ecosystem without anyone actually learning anything. — OpenAI

• Sam Altman and Jony Ive announced OpenAI’s first phone-like prototype, it has no screen, no notifications, no apps, no camera, no icons, no sounds, no features at all, it’s basically void, and you will buy it for a — Fortune (see what i there there?)

• Alibaba launched Qwen Smart Glasses, AI-powered eyewear with navigation so you can be ambushed by “90% OFF FLASH SALE” pop-ups right into your brain. — Bloomberg

• n8n announced Version 2.0 featuring AI agents that help you turn a five-minute task into an eight-hour IKEA-style node jungle, conveniently large enough to screenshot and post on LinkedIn with the caption “BUILT THIS IN 5 MINUTES — COMMENT ‘WANT’ BELOW” — n8n Community

• ElevenLabs launched a marketplace allowing brands to license voices of celebrities, prepare for product demos with David Attenborough going “here we see the stealth-mode startup… unaware it is not funding-ready” — ElevenLabs

• OpenAI has projected that at least 220 million people will pay for ChatGPT subscriptions by 2030, a number that bravely assumes humanity survives that long to renew. — Reuters

• Sam Altman warned employees of “temporary economic headwinds” as Google resurges, which is CEO-language for “we need to outrun Google before they remember they invented half this stuff” — Economic Times

• Meta is reportedly building Project Luna to challenge ChatGPT and Google’s AI Mode, moving at a pace that makes it feel like Internet Explorer is running their innovation department. — Washington Post

• MIT launched the “Iceberg Index”, predicting AI will disrupt 11.7% of U.S. jobs and $1.2 trillion in wages, they marked the exact spot the Titanic would hit, but everyone just keeps ordering cocktails. — Fortune

• OpenAI suffered a data breach through Mixpanel that exposed user names and emails, all because some senior PM wanted to A/B test a button color. — Business Insider

• xAI's Grok faced backlash after the model went on a tangent praising Elon Musk as "fitter than LeBron James," ← I haven’t touched this item lol this is the original — The Guardian

• Anthropic released research showing AI models can learn deception and sabotage, confirming the models are entering their teenager phase . — AI Magazine

• Chinese Scientists developed a computing system that processes AI with light, making GPUs obsolete; NVIDIA has responded by acquiring the sun. — Nature

• Elon Musk announced Grok 5 will battle the best League Of Legends team alive, putting AI through the most toxic test environment ever created. — X (@elonmusk)