🚀 Google's I/O flex
·
📖4 min read
🚀 Google's I/O 2026 Was Basically an AI Firehose
- Google dropped Gemini 3.5 Flash, a new Gemini Omni video model (text/audio/images → videos), AND cut the AI Ultra plan to $100/month. Also?? You can now talk to your Gmail inbox, "Ask YouTube" stuff, and build Android apps from web prompts.
- Google is carpet-bombing every product you already use with AI to beat OpenAI. It's giving "if you can't out-innovate them, out-distribute them" energy.
🇺🇸 US Tries to Legally Ban "Made in China" AI
- New bipartisan bill wants to block China from selling AI tech abroad - and is throwing a $500 MILLION fund at allies to buy American chips, models, and cloud instead. Basically: anti-Belt-and-Road, but for GPUs.
- The AI race is now straight-up geopolitics. Sir this is no longer about chatbots - it's about who runs the global digital empire.
🇨🇳 Alibaba Ships 560K AI Chips, Says "Nvidia Who??"
- Alibaba launched the Zhenwu M890 - 3x faster than the last one, built for AI agents - and has already shipped 560,000+ units to Chinese companies. Faster chips reportedly coming in 2027 and 2028.
- US export bans were supposed to slow China down. Instead they basically speedran China's domestic chip industry. Great job everyone 👏
🚨 Hackers Are Now Using AI to Break In Stupid Fast
- Per Verizon's new report, exploiting software flaws has officially passed stolen passwords as the #1 hacking method. Time to patch a vuln has shrunk from months to hours. Also, employees secretly using random AI tools at work = data leak buffet.
- We've entered the fight-AI-with-AI era of cybersecurity. Your IT team is now in a Terminator vs. Terminator situation. Not great!!
💬 Meta "Generously" Lets Rivals onto WhatsApp
- Meta is giving OpenAI and other AI chatbots limited free access to WhatsApp in Europe - after a certain message count, they start charging. Smaller AI startups are like "this is fake fairness, bestie."
- This is Meta trying to dodge EU regulators while still controlling the pipes. Classic "open, but on my terms" speedrun.
🧩 Anthropic, OpenAI, and Blackstone Get Tangled Up
- Plot twist: Google DeepMind's boss was an early investor in Anthropic (?!?). Anthropic gave its agents safer ways to run code, OpenAI is adopting Google's image watermarks to spot AI fakes, and Google + Blackstone teamed up on a $5B AI cloud venture.
- The "rival" AI labs are basically one incestuous money-and-talent web. Competition? Sure. But also... brunch.
🤝 Google DeepMind Pulls the Classic "Acqui-Hire" Move
- DeepMind paid up to $90 MILLION to license tech from startup Contextual AI - and casually hired 20+ researchers including the CEO. Totally not buying the company though!! Wink wink.
- This is the new antitrust-proof talent heist playbook. Why acquire a startup when you can just vacuum out its brain legally??
🏦 Banks Are Quietly Replacing Humans With AI
- HSBC's CEO says AI will "create and destroy" jobs - they're retraining 200,000 workers. Meanwhile Standard Chartered is just cutting to the chase and axing 7,000+ jobs by 2030, replacing them with AI.
- The "AI will take jobs" debate is OVER - it's actively happening right now in finance. Learn the tools or get automated into the void.
🏥 Commure Hits $7B Doing the Paperwork Nobody Wants
- Healthcare AI startup Commure just raised $70M, pushing their valuation to a chunky $7 BILLION. Their AI agents handle 85% of patient billing and admin work across thousands of clinics - no humans needed.
- Turns out the killer AI use case isn't curing cancer - it's making medical billing less soul-crushing. We love to see it.
🎮 Status AI Wants You to Roleplay Your Whole Feed
- New app Status AI raised $17M to turn social media into a choose-your-own-adventure. You can role-play as a pop star, live inside your favorite TV show, or even run for president. The AI builds the world around you.
- Doomscrolling is out, becoming the main character of a fake AI universe is in. This is lowkey chaoscore but honestly?? Sign us up.