"Dream big, little explorer!" 🚀🌌
Meet Billie, on his journey to Mars...
"My REDDY Buddy Mars", a short AI film made entirely with grok Imagine.
Original story, editing & music by @Xiat
Watch & let me know below! ❤️
#AIFilm#GrokImagine#MarsDreams@xai
🚨BREAKING: Nvidia just unveiled a laptop chip that can run a 200-billion-parameter AI model without ever touching the cloud.
Jensen Huang unveiled it this morning at Computex in Taipei. RTX Spark is Nvidia's first chip built for Windows laptops and small desktops. One petaflop of AI performance, a quadrillion operations per second, inside a thin notebook.
Until now the heavy work happened elsewhere. You type into ChatGPT, your words go to a data center, GPUs do the thinking, the answer comes back. You rent that compute monthly and everything you send passes through someone else's servers.
RTX Spark moves the work onto the machine in front of you. A 20-core Arm processor, a Blackwell GPU, up to 128GB of unified memory. That shared memory is what holds a 200-billion-parameter model and a million-token context locally, no round trip to a server.
For you that means an AI agent that reads your files and drafts your work without uploading anything, a coding assistant that runs offline, Photoshop and Premiere about twice as fast. No subscription, no latency, no question about where your data went, because it never left the device.
It ships this fall in laptops from ASUS, Dell, HP, Lenovo, Microsoft Surface, and MSI. A direct shot at Apple silicon and Qualcomm's Snapdragon.
Watch the price. Nvidia is launching into a memory shortage pushing DRAM costs up, and 128GB of fast memory is the component getting expensive. The desktop version retails near $4,700. These laptops will be premium.
For two years AI meant bigger models in bigger data centers pulling more power from the grid. RTX Spark points the other way. The cloud is not going away, but serious AI no longer has to live somewhere else.
Source: Nvidia, Tom's Hardware, Phoronix, Computex 2026