📉 OpenAI's vibes are off

📉 OpenAI Finally Has a Bad Quarter (??)

• A new report says OpenAI missed its targets for both new users and revenue. Their CFO is sweating about how to pay for those enormous data centers if growth keeps slowing.

• Why it matters: This is a massive speedbump for OpenAI's planned IPO. Turns out even the AI golden child has to do math eventually.

---

💸 The Entire AI Stock Market Catches a Cold

• Oracle and SoftBank stocks tanked on the OpenAI news, and chipmakers Nvidia and AMD got dragged down with them. Sir, this is supposed to be a bubble - bubbles don't deflate.

• Why it matters: It's giving single point of failure vibes. The way that the entire tech market rides on one company's growth chart is lowkey terrifying.

---

⚖️ Elon Takes the Stand for $130 BILLION

• Musk is officially in court for his $130 billion lawsuit, claiming Sam Altman stole a charity to turn it for-profit. OpenAI's lawyers fired back that this is just "sour grapes" because Elon dipped before the company blew up.

• Why it matters: If Elon wins, OpenAI might have to undo its for-profit conversion entirely. The audacity I-

---

🚀 NVIDIA Drops Nemotron 3 Nano Omni

• NVIDIA released a new open multimodal model that handles vision, audio, and text all at once - and runs at 9x the speed of other open multimodal models.

• Why it matters: Free, fast, and multimodal. Devs eating good this week.

---

🎨 Claude Just Moved Into Blender, Adobe, and SketchUp

• Anthropic launched tools that let Claude work directly inside creative software. You can now ask it to write scripts, edit 3D models, or process files without leaving the app.

• Why it matters: Claude went from "chatbot" to "actual coworker who knows Blender" real fast. Designers either rejoicing or updating their resumes - possibly both.

---

📱 Xiaomi Casually Open-Sources a 1 TRILLION Param Model

• Xiaomi dropped MiMo-V2.5-Pro - a 1.02 trillion parameter open-source model with a 1 million token context window. It ties with top models on coding benchmarks. For free.

• Why it matters: Open-source is catching up to closed models at an alarming rate. We love to see it (if you're a dev, not OpenAI).

---

🌐 SpAItial's Echo-2 Generates Entire 3D Worlds From a Prompt

• New model Echo-2 turns a text prompt or photo into a fully explorable 3D world - and unlike older models, the spaces are physically consistent instead of janky video loops.

• Why it matters: HUGE for video games, architecture, and robot training. Building virtual worlds just got... terrifyingly easy.

---

🪖 600+ Google Employees Beg Sundar to Stop Pentagon Deal

• Google signed a classified deal letting the Pentagon use its AI models for any legal government purpose. Over 600 Googlers wrote a letter begging the CEO to stop.

• Why it matters: It's 2018 all over again. The "Don't Be Evil" → "Actually, Define Evil" pipeline continues.

---

🇨🇳 Chinese Tech Giants Stampede for Huawei Chips

• ByteDance and Alibaba are rushing to buy Huawei's Ascend 950 AI chips, with demand exploding after DeepSeek V4 launched optimized for Huawei hardware.

• Why it matters: China is successfully decoupling from Nvidia. The chip war just got a real second player.

---

🌍 US Drops €50B on a Croatian Data Center

• The US signed massive Balkan deals including a €50 billion AI data center in Croatia - partly to push out Russian influence in the region.

• Why it matters: AI infrastructure is now literal geopolitical strategy. Data centers are the new aircraft carriers.

---

🇪🇺 EU Talks for 12 Hours, Agrees on Nothing

• EU lawmakers debated changes to the AI Act for 12 hours and... failed to agree. Some countries want industry exemptions to compete with the US and Asia.

• Why it matters: Pure regulatory chaos for European companies, while Big Tech enjoys the extended grace period :)

---

🏛️ Congress Actually Wants to Regulate AI Chatbots

• New bipartisan bills would require family accounts to make AI chatbots safer for kids, plus tax breaks for companies funding employee cybersecurity training.

• Why it matters: Lawmakers might finally do something about kids and AI without nuking innovation. Bestie, the bar is on the floor but we'll take it.

---

🏦 Banks Are Adopting AI 2x Faster Than Regulators

• A new study found banks are using AI twice as fast as their regulators, with models like Anthropic's Mythos making cyber risks nearly impossible to monitor.

• Why it matters: If watchdogs don't catch up with agentic AI, the next financial crisis is going to be... interesting. No because WHY are we speedrunning 2008??

---

📦 Amazon Replaces Human Recruiters with AI Interviewers

• Amazon is rolling out software where AI agents conduct face-to-face job interviews with zero humans involved. They're calling the design "humorphism."

• Why it matters: Millions of job seekers about to get rejected by a literal machine. Not great!!

---

🎥 YouTube Tests "Ask YouTube" Search

• YouTube's testing a feature that gives step-by-step answers instead of a list of videos - mixing text, short clips, and long videos tailored to your question.

• Why it matters: YouTube is quietly turning into a search engine. Google searching Google for Google answers.

---

👻 Snapchat Drops Conversational AI Ads

• Snapchat launched "AI Sponsored Snaps" in the main chat tab, letting users talk directly to brand AI bots and ask product questions inside the ad.

• Why it matters: Ads you can chat with. The future of advertising is talking to a corporate chatbot pretending to be your friend. Cool!!

---

👨‍⚖️ Lawyer With 30-Year Career Gets COOKED by ChatGPT

• A federal judge publicly destroyed a US prosecutor for filing a court brief stuffed with fake quotes and made-up laws generated by AI. The lawyer lost his job.

• Why it matters: 30-year career → gone in one hallucinated citation. Stop 👏 trusting 👏 AI 👏 blindly.

---

📻 Researchers Built an AI That Sounds Like It's From 1931

• Scientists built "Talkie" - an AI trained only on books and news from before 1931, using vintage etiquette manuals to teach it how to chat.

• Why it matters: Since every modern AI sounds exactly the same, this proves we can build chatbots with actual personality. It's giving Great Gatsby-core.

---