Good morning 👋 Let’s start by acknowledging a core Silicon Valley truth: No idea is too absurd if you can sell it. Case in point, Colossal Biosciences just raised $120 million at a $10 billion valuation to reincarnate the dodo, history’s cautionary mascot of extinction. They’re literally disrupting evolution. Suddenly your crypto-powered, AI-driven robot pizza startup doesn’t sound so crazy.
⚡ OpenAI and NVIDIA have teamed up to build 10GW of AI datacenters. NVIDIA intends to invest up to $100 billion, scaling up its investment as each new gigawatt comes online. The project is so massive it’ll use 4-5 million GPUs and require as much electricity as 10 nuclear reactors just to keep things running. Even after NVIDIA takes its notorious 60% chip tax, the deal gives OpenAI enough leverage to keep investors throwing money at AI for years. At this rate, NVIDIA can keep the AI bubble floating forever.
✨ Silicon Valley was founded by idealistic, anti-establishment rebels with big dreams. But lately it seems that tech leaders have ditched their founding values for toxic masculinity, Neo-Nazi madness, and shady political favors. Billionaire CEOs openly court controversial politicians, trading principles for short-term gains, while once-outspoken insiders quietly plot their exits. Silicon Valley’s rebels didn’t just join the establishment. They became exactly what they once despised.
💰 PsiQuantum raised $1 billion promising quantum computing by 2027. Right now, quantum computers struggle to do basic math, and even industry giants like Google and IBM admit reliable quantum tech is decades away. Yet PsiQuantum somehow con-vinced investors it’ll leapfrog the entire field in just two years. Silicon Valley’s never met a billion-dollar pipe dream it didn’t like, but this quantum fantasy sets a new standard for lighting cash on fire.
🤜🤛 Together with Julius
Don’t get stuck in the past
Unlock the power of fresh data, automatically. From spreadsheets to databases, Julius AI pulls all your data together 24/7 to deliver smarter business insights automatically. No coding required.
Why you need it: Instead of manually pulling data every time you need an update, Julius puts analysis on autopilot. It lets you monitor your key metrics across data sources continuously so you can focus on strategy, not building reports.
Never miss a trend. Schedule reports to run overnight and wake up to fresh summaries, forecasts, and performance metrics in your inbox or Slack.
🎯 MIT researchers just boosted small AI models’ accuracy by 66%. Until now, smaller AI systems struggled with complex, multi-step planning, often scoring under 5% accuracy. MIT’s new method, called PDDL-INSTRUCT, teaches models logical thinking step by step, checking their work at each stage with external validators. On tough benchmarks, MIT’s tiny model jumped from near-zero to 94% accuracy. Soon, cheap AI might start tackling problems only billion-dollar models could solve.
🌮 Whatever happened to drone-delivered burritos? The FAA just loosened drone rules, but instead of dinner drops, we’re getting police surveillance. Companies like Skydio and Flock are already deploying autonomous drones to track suspects and silently scoop up massive amounts of data. Authorities promise they’ll respect our privacy, but we’ve heard that one before. Apparently, burritos aren’t as profitable as watching everything we do.
🛝 Silicon Valley is obsessed with virtual playgrounds for AI. Startups like Mechanize, Axiom, and Prime Intellect are cashing in on environments where AI agents practice coding, shopping, and basically being human. Anthropic alone is planning to dump a billion dollars into these digital sandboxes next year, hoping they’ll unlock smarter AI without torching the power grid. Now we just wait for agents to figure out how to game these playgrounds, too.
fun stats
of people cheat when they use AI for tasks 🤨 A new study suggests ethics go out the window when using AI, with honestly dropping up to 30%.
[Futurism]
Infinite money glitch 🤑