Hello world!

The strongest 8B model in the world isn't trained on Nvidia hardware Notable development

We are beyond thrilled to share our first flagship models, Rnj-1 base and instruct 8B parameter models. Rnj-1 is the culmination of 10 months of hard work by a phenomenal team, dedicated to advancing American SOTA OSS AI. Lots of wins with Rnj-1. 1. SWE bench performance close

We started building in @ziglang 3 years ago This week, @AnthropicAI bet on Zig with their acquisition of @bunjavascript We're extremely bullish on this language We wrote up everything we've learned building a browser in Zig ⤵️ https://t.co/fLb73GDCom

Today at #NeurIPS2025, we present Titans, a new architecture that combines the speed of RNNs with the performance of Transformers. It uses deep neural memory to learn in real-time, effectively scaling to contexts larger than 2 million tokens. More at: https://t.co/pEuI5oDkkn

Today, Netflix announced our acquisition of Warner Bros. Together, we’ll define the next century of storytelling, creating an extraordinary entertainment offering for audiences everywhere. https://t.co/rXPFMNIs1A

CONFIRMED: Oxford University Press has named ‘rage bait’ as the Oxford Word of the Year 2025. #OxfordWOTY

1/ 10 We are pleased to introduce Z-Image, an efficient 6-billion-parameter foundation model for image generation. Through systematic optimization, it proves that top-tier performance is achievable without relying on enormous model sizes, delivering strong results in

Super cool to see FLUX.2 builds on @MistralAI Mistral Small 3 24B parameter VLM with a rectified flow transformer. The VLM brings real world knowledge and contextual understanding, while the transformer captures spatial relationships, material properties, and compositional

🏆 We are incredibly honored to announce that our paper, "Gated Attention for Large Language Models: Non-linearity, Sparsity, and Attention-Sink-Free" has received the NeurIPS 2025 Best Paper Award! A huge congratulations to our dedicated research team for pushing the boundaries

being the first astronaut that goes on the Mars mission must be so fun, because you get to be away from civilization for so long that maybe you'd want finally want to just live on that planet alone

Sharing moments shouldn’t depend on the phone you have. Starting today with the Pixel 10 family, Quick Share now works with AirDrop, making secure file transfers between Android phones and iPhones more seamless. This builds on our commitment to cross-OS compatibility to bridge

Today we are releasing & open-sourcing Segment Anything 3 (SAM 3). It is a state-of-the-art model for image & video segmentation, and builds upon the work of SAM & SAM 2. SAM3 will also power features in Edits, Meta AI, & Facebook Marketplace soon. https://t.co/ufppJ6gR3c

meta

⚡️OSS Project Spotlight⚡️ Below is a powerful resource monitoring tool for Linux that provides deep insights into system performance, including CPU, memory, I/O, and more — all from an intuitive terminal interface. Learn more:

twitter dying was not on my to do list today

tfw haven't opened Cursor in months

agents are just the prompts we made along the way

It's SOTA, not only open weights SOTA :)

Today, we're releasing Kimi K2 Thinking, our best open-source model. What makes it different isn't just the benchmarks, though it achieves SOTA results on Humanity's Last Exam, BrowseComp, and other challenging tests. What matters is how it thinks. It reminds me of the minds on

Midjourney runs fully on Bun Server side routing: Bun Runtime: Bun Client side bundling: Bun Scripts: Bun Real-time generation previews: Bun No framework beside vanilla React and no data management library. This enabled us to scale the site to millions with only ~5 full-timers.