Good morning 👋 and happy Property Rights Awareness Day! It’s not actually a thing, but maybe it should be after Elon Musk parked his spaceships all over Cards Against Humanity’s Texas land. The infamous gamemaker originally bought the plot as part of a fan-funded stunt, then promptly sued SpaceX for squatting, demanding $15M. The good news: SpaceX agreed to clear off its space junk. The better news: instead of cash, fans get limited-edition cards specifically designed to mock Musk.
⚡ China’s new MiniMax M2 sets the benchmark for open source LLMs, rivaling Claude at 8% cost. And that’s exactly the problem. A16z partner Anjney Midha warns that open source AI is quickly becoming “China’s game,” creating a geopolitical headache for the US. While American companies kept their AI models locked down and pricey, Chinese startups jumped ahead by launching powerful, cheap open source models like DeepSeek and Qwen. Now we’re racing to play catch-up before our allies realize they can self-host these efficient models for free.
🧠 A new open source robot brain sees the world in full 3D. Until now, robot foundation models mainly learned from flat, labeled images, limiting real-world interactions. SPEAR-1 changes that, adding 3D training data to give industrial robots an intuitive grasp of how physical objects actually behave and move through space. This may one day enable general humanoids to quickly adapt to new situations. Early tests show the free open-source model performs nearly as well as billion-dollar proprietary ones. The best part? It’s available to any startup or researcher who wants to play.
❤️🩹 Over a million people talk to ChatGPT about suicide every week. Another 560,000 show signs of psychosis or mania, while roughly 2.4M express suicidal thoughts or emotional dependence on the chatbot. After getting sued in August, OpenAI worked with mental health experts to tweak GPT-5 to better recognize serious distress, cutting harmful responses by 52%. Perhaps growing awareness of AI psychosis will push OpenAI to become better stewards of our sanity.
🤜🤛 Together with
✨ Your guide to agentic AI innovation
✨ Your guide to agentic AI innovation
Looking to adopt and scale agentic AI solutions?
This report from Glean and Gartner combines innovative insights with practical steps: how to tame agent sprawl, set guardrails, combine LLMs, and ensure reliable performance at scale.
Zero in on opportunities and solve critical challenges.
🦾 Robot hands will unlock a $5 trillion market – once they actually work. Researchers say we’re at least 10 years away from robotic hands truly human-like. Why so long? Because hands that can accurately sense and apply the perfect amount of force, like gently holding a pencil without crushing or dropping it, is insanely complicated. Roboticists admit they still need significantly more sensors beneath artificial skin just to mimic basic touch. And they’re nowhere close to replicating human skin that heals itself or joints that naturally lubricate. No wonder thumbs were the original killer app.
🧪 Sam Altman wants to get inside your head. Brain implants always felt sketchy because surgery and neurons don’t exactly mix. But Sam’s new startup, Merge Labs, believes ultrasound is the safer BCI play. Instead of wires in your skull, Merge uses gene therapy to make brain cells responsive, allowing signals to be read without surgery. Altman wants you to be able to think something and have ChatGPT respond. With $250 million from OpenAI, your inner voice might soon have a co-pilot.
👾 AI might make cybersecurity teams obsolete. Vibe coding ships fast, buggy software loaded with vulnerabilities, making life easy for hackers. After all, why invent sophisticated new attacks when you can exploit predictable MITRE vulnerabilities from 20 years ago? Former CISA head Jen Easterly argues this isn’t a cybersecurity problem, it’s a software quality crisis. She believes that AI is uniquely suited to rapidly find and eliminate flaws people can’t keep up with. Maybe someday breaches will become rare anomalies instead of daily business expenses.
fun stats
people actively use ChatGPT every week 👍 Launched in Nov 2022, the chatbot has now passed the bar exam, helped a user win the lottery, and prompted 200 complaints to the FTC for psychological harm.
[Wired]