- Projects of AI
- Posts
- Google dropped a next-gen lightweight AI model Gemma 3 for single GPU
Google dropped a next-gen lightweight AI model Gemma 3 for single GPU
Plus: Gemini-powered robots
Amandeep Singh
March 13, 2025
Welcome Humans🤖,
Here is what we have Today:
📝 Japanese AI passes peer review while professors weep.
📊 Eraser's new tool makes developers nervous.
💻 AI runs on one GPU, still smarter than your entire IT team.
🤖 Robot revolution: Google's Gemini enters the physical world.
💼 New Job Opportunities
Google just dropped Gemma 3, a next-gen lightweight AI model that packs heavyweight performance—running efficiently on just a single GPU or TPU while competing with much larger models.
Image Source: Google
The breakdown:
✅ Comes in four sizes (1B, 4B, 12B, 27B) for everything from phones to laptops
✅ 27B model beats bigger rivals like Llama-405B, DeepSeek-V3, and o3-mini in human preference rankings
✅ 128K token context window, 140 languages, and multimodal abilities for images, text, and videos
✅ ShieldGemma 2 (4B model) filters explicit content for safer visual applications
Why does it matter?
Gemma 3 is a game-changer. It outperforms models twice its size, runs on a single GPU, and remains open-source, fast, and multimodal. This could redefine AI deployment across devices, from smartphones to enterprise applications.
Robots are about to get a major AI upgrade. Google has partnered with robotics developer Apptronik (known for its work with NASA and NVIDIA) to launch Gemini Robotics and Gemini Robotics-ER—two AI models built on its most advanced AI yet, Gemini 2.0. The goal? Helping robots tackle real-world tasks like never before.
Image Source: Google
🔑 What’s New?
Smarter Robots: Gemini-Robotics lets robots understand and react to new environments—even ones they haven’t been trained on.
Real-World Skills: In tests, AI-powered robots followed voice commands and performed delicate tasks like folding paper, putting glasses into a case, and removing bottle caps.
Enhanced Awareness: The Gemini Robotics-ER model specializes in spatial awareness and task planning. Select developers can integrate it into their own robotics systems.
🚦 When Can We See Them in Action?
Google hasn't announced a release date yet. But it’s training the models to evaluate task safety and developing an AI safety benchmark. This follows last year’s "Robot Constitution"—a set of Isaac Asimov-inspired safety rules for AI robots.
The future of robotics? Closer than you think.
Trending Today
Nicoo, the most realistic robotic puppy has the ability to walk, bark and behave has a real puppy, made up of soft strong material that make it similar to a real puppy.
Eraser, an AI-powered tool that converts your company’s data into visuals, just introduced a new feature that can “generate live diagrams from your codebase that stay up-to-date automatically.”
Sakana, a Japanese AI startup announced that its AI system successfully generated a scientific paper that passed peer review, with the company calling it the first fully AI-authored paper to clear the scientific bar.
Snap released its first AI Video Lenses, powered by its own in-house model, offering premium subscribers three AR animations with new options planned to launch weekly.
Recommended Reading
If WE had to recommend other newsletters
AI Tool Report is an Newsletter, we enjoy reading daily. AI Tool Report delivers top techniques on how AI can transform your business filled with practical tips, real-world examples.
A Smart Bear is for people who love thinking about strategy, startups, product, marketing, decision-making, and founder psychology.
Job Opportunities
AI News
Source: Ideogram
Ideas? Comments? Complaints?
We read your emails, comments and poll replies daily.
Get the most important AI, tech, and science news in a free daily email.
New Here? Subscribe!
Sponsorship Slots Open for March and Reach over 800+ active readers. (Now 40% off) 🤯
What`d you think of today`s edition? |
Until next time, Stay Informed!
Reply