Google releases DiffusionGemma.✨
The new 26B-A4B diffusion text model runs locally on 18GB RAM.
It supports high-speed text generation, thinking, image, video and 256K context.
Run and train via Unsloth Studio.
GGUF: https://t.co/ZH0dCJQ59P
Guide: https://t.co/wYLfJWE6kG
@bcherny It's less tied to a discipline anymore. The boundary weakens as general purpose AI models are capable to generate quality outputs. Current conversation about AI outputs revolves around taste and quality and whether evaluation is comparing against the right metrics.
Introducing autoresearch for arXiv papers
Change 'arxiv' to 'autoarxiv' in any paper URL
An agent deploys to resolve setup issues on the codebase, run a minimal reproduction, and estimate full replication cost. Read more below
Latest microVM by AWS. "For interactive use cases such as interactive development environments and AI sandboxes, ... MicroVMs can be suspended when idle, preserving memory and disk state while reducing costs." https://t.co/iyBXpbh9MB
Introducing Sakana Fugu: A full multi-agent orchestration system accessible via a single model API.
Our ‘Fugu Ultra’ model matches the performance of Fable and Mythos, delivering frontier capability without the risk of export controls.
Try it: https://t.co/hhO6qTawgb 🐡
When you think building website is solved (with AI). You have apps like @ployai innovating on the space. I love it how it "slurps" - getting context of existing website and build something consistent.
OpenAI's market share drops below 50% for the first time. Google is quickly eating into it.
The pure LLM business doesn't have stickiness. Regular people don't see a difference between ChatGPT and Gemini, and because they already have their entire online ecosystem with Google, it's just simpler not to use other LLMs.
Whoever controls the browser wins.
@jackprice it's about taste and it's something that a human needs to iterate over. I design the logo and page for my open source project to fit the narrative of what it does at https://t.co/ZKj0aa8P6D Not sure if this looks like AI slop because it can be very subjective.
@sidchatterjee So it’s empheral? Let’s say I want to run a one-off task and forget I don’t have to worry about the setup and tear down? Sounds like a perfect setup for test automation and many more.
We built an AI that can draw on your screen.
It's a true personal tutor.
Using Claude Opus we're able to draw polygons, point with pixel perfect accuracy, and walk users through complex steps directly on their screen.
Here's me learning Pythagorean Theorem + FL Studio.
Demo: