Zac Zuo

Hi everyone! Check out Instella, a new family of 3B language models from AMD. These models are interesting because they're achieving performance comparable to larger, open-weight models, but with a smaller footprint. Key points: 📈 Strong Performance: Outperforms other 3B models, and rivals some larger open-weight models like Llama-3.2-3B and Gemma-2-2B. ✨ Reasoning Focus: Second stage of...

Hi everyone!DuckDuckGo brings Duck.ai, a free and private AI chatbot! It's now out of beta and offers a really interesting alternative to other AI chat services.What makes it different:🛡️ Privacy-Focused: Chats are anonymized and not used for training the AI models. DuckDuckGo proxies the requests, so your IP address isn't shared.🤖 Multiple Models: You can choose from a selection of models,...

Hi everyone! HunyuanVideo-I2V is a new open-source image-to-video model from Tencent, and it's a significant addition to the growing AI video space! This builds on their previous HunyuanVideo project, but focuses specifically on turning still images into videos. Here's the detail: 🖼️ Image-to-Video: Give it an image and a text prompt, and it generates a video. 🌅 High Resolution: Can generate...

Hi everyone! Hailuo AI is now going to mobile! It's like having a complete AI video studio right in your pocket! This app, from MiniMax, lets you create amazing videos using just text or images. The really cool part is the new "Director Mode", powered by their newest I2V-01-Director and T2V-01-Director models. You could control camera movements (like pan, zoom, tilt, dolly, even Hitchcock...

Hi everyone!Check out QwQ-32B, a new open-source language model from the Qwen team. It's achieving something remarkable: reasoning performance comparable to DeepSeek-R1, but with a model that's 20 times smaller (32B parameters vs. 671B)!This is a big deal because:🤯 Size/Performance Ratio: It punches way above its weight class in reasoning, math, and coding tasks.🧠 Scaled Reinforcement Learning:...

Hi everyone!AI Mode is clearly the punch back to Perplexity. Google is trying to take the search experience to the next level, using a custom version of Gemini 2.0 to tackle complex, multi-part questions that might normally require multiple searches.It's currently available as an opt-in experiment in Labs for Google One AI Premium subscribers. They're gradually rolling it out to more...

Hi everyone!Check out Aya Vision, a new set of open-weights models from Cohere For AI, and this is a significant step towards making AI truly global! Most vision-language models are heavily biased towards English. Aya Vision tackles this head-on by supporting 23 languages spoken by over half the world's population.Here's why it's important:🌍 Multilingual by Design: Excels at understanding and...

Hi everyone!Sharing Colab Agent, a new tool integrated directly into Google Colab that could seriously speed up data science workflows! It's powered by Gemini 2.0 and lets you generate entire, runnable Colab notebooks just by describing your analysis goals in natural language.Here's how it works:1. Upload Data: CSV, JSON, or .txt files (up to 1GB).2. Describe Your Goal: Tell it what you want to...

Hi everyone!Sharing Sesame's Conversational Speech Model (CSM), and this is a big step beyond typical text-to-speech. The goal is to achieve what Sesame calls "voice presence": making spoken interactions feel real, understood, and valued.A PH version of this model System Card is :)😃 Emotional Context: It tries to understand and respond to the emotion in the conversation.⏱️ Conversational...

Hi everyone!Sharing CogView4, a new open-source text-to-image model from the ChatGLM team, and it's got some seriously impressive capabilities!What stands out:🖼️ Native 2K Resolution: Generates images at 2048x2048 natively – no need for upscaling.📝 (Almost) Unlimited-Length Prompts: Supports verrrrrrrrry long and detailed prompts, in both Chinese and English.🔠 In-Image Text Generation: Can...