Good morning 👋 Ever wondered if the latest novel you grabbed on Amazon was spit out by a chatbot? Good news. A new startup just launched the “Organic Literature” certification, guaranteeing your book was written by an authentic human. Several indie publishers are already onboard, with the first certified AI-free books hitting shelves soon. Because the best stories come from real authors, caffeine, and deadlines.
🚩 The US is pushing “sovereign AI” but China may already be winning. Washington is selling foreign governments on the idea of building their own national AI, promising control, privacy, and independence. In reality, sovereign AI means buying American. Servers from Nvidia, models from OpenAI, and data centers built to US specs. It’s digital diplomacy disguised as democracy. Meanwhile, China skipped the sales pitch and started giving its tech away. Open source models from Alibaba and DeepSeek are spreading worldwide, offering actual independence without the middleman.
🪩 Sam Altman wants to rebuild the world in his image. He’s shaking hands and signing deals for a staggering 26 GW of computing power, roughly equal to the output of 20 nuclear reactors. Each GW costs around $50 billion, with total costs exceeding $1.3 trillion. To pull it off, OpenAI is counting on unproven financing strategies, borrowed infrastructure, and tech breakthroughs that don’t yet exist. Altman openly admits profitability isn’t even on his radar. If he pulls this off, we get the singularity; if he’s wrong, we get the world’s biggest GoFundMe.
🪞 MIT says the AI industry’s obsession with giant data centers is pure BS. A new report found that throwing massive compute at AI isn’t paying off anymore. Performance gains from huge models are rapidly shrinking, even as costs explode. Meanwhile, smaller models like Anthropic’s new Claude Haiku 4.5 are stealing the show. Haiku 4.5 matches Anthropic’s best previous model on coding tasks, yet it runs twice as fast at just a third of the cost. Imagine all the money we could save if we didn’t have to build 26GW of data centers!
🤜🤛 Together with
Work at the speed of thought
Wispr Flow turns your words into edit-free text in every app with zero setup.
Why leaders love it: Typing is tiring. Flow let’s you dictate emails, docs, and code 4x faster. On the go or at your desk, it gets your words out and makes it sounds like you, not a bot.
Ditch your keyboard. Use Flow to fly through Slack notifications, brain dump with ease in Notion, or give better prompts to Claude or Cursor. Used by pros at OpenAI, NVIDIA, and Loveable to get more done.
🎞️ Veo 3.1 just broke what was left of reality. Google’s upgraded AI video model makes it easier than ever to create shorts that are virtually indistinguishable from reality. The latest version offers hyper-realistic audio, seamless object insertion, and the ability to upload multiple image references. Veo 3.1 can even clone existing videos and let you rewrite scenes in seconds, raising the bar for realism and misinformation simultaneously. Reality was overrated anyway.
🪖 Trump’s AI czar has declared war on Anthropic. David Sacks, venture capitalist turned White House “Crypto and AI Czar,” accused Anthropic of running a sophisticated fear-mongering campaign to shape AI regulation. Anthropic, once loudly anti-military and now cozy with the Pentagon, has nearly tripled its lobbying budget, pushing rules critics say will choke out smaller startups. Apparently the real threat to AI is someone mentioning safety.
🤔 How do you give kids career advice in an AI world? Trades like plumbing and HVAC sound great until everyone has the same idea – there are only so many toilets to fix. Liberal arts degrees might build critical thinking skills, but the tuition doesn’t always match the paycheck. Specialized paths like asteroid mining or human rights law might pay off big if you guess right. Many parents say networking and being outgoing matter, but with AI rewriting the rules every day, giving teens career advice feels like throwing darts blindfolded.
fun stats
of people around the world don’t trust the US to regulate AI 🤞 China gets even lower marks, with 60% of people having little to no trust in them.
Who can you trust? According to a new global survey, more than half of people believe their own country can do it best, followed closely by the EU.