Google drops Gemma 3

๐ŸŽค The most realistic text-to-voice ever + ๐Ÿ“บ Hyper-realistic UGC videos with AI

Together with

Hi everyone,

The end of the week always brings some awesome AI updates. 

Today, we have 2 big ones:

  • Google launches Gemma 3, a collection of lightweight but powerful models - making them ideal on phones or self-hosted workstations

  • Captions builds a UGC AI modelโ€ฆ yes, a model that can create super-realistic user-generated-adsโ€ฆ which is actually INSANE!

Letโ€™s get right into it.

In this issue:

  • ๐Ÿ‘๏ธ In Focus: Bring your words to life effortlessly

  • ๐ŸคฟDeep Dive: Google launches Gemma 3

  • ๐Ÿ–ผ๏ธAI Art: Examples of great and trending AI art

  • ๐ŸคฟDeep Dive: Mirage generates hyper-realistic AI videos

  • โš’๏ธTool Snapshots: Tools for AI, no-code, and productivity

  • ๐Ÿ“ฐTop News: News on AI, no-code, and productivity

๐Ÿ‘๏ธIN FOCUS

Struggling to find human-like AI voices for your next big idea?

Murf AI, the award-winning text-to-speech platform, just launched Murf API โ€” a scalable voice solution built for creators, developers, and businesses.

โœ… 130+ voices, 13+ languages โ€” with 15+ speaking styles for lifelike delivery
โœ… MultiNative technology โ€” seamless language & accent switching in the same voice
โœ… Advanced audio duration control โ€” precise timing without losing naturalness
โœ… Developer-friendly integration โ€” RESTful API + Python SDK for easy deployment

๐Ÿ‘‰Early-stage startup? Get $5,000 in free API credits for 3 months
๐Ÿ‘‰Not a startup? Your first 100,000 characters are freeโ€”no strings attached

๐Ÿคฟ DEEP DIVE

Google Introduces Gemma 3 and ShieldGemma 2, Advancing Open AI Models

Intelligence: Google has unveiled Gemma 3, an advanced series of open AI models designed to enhance application development while incorporating rigorous safety protocols.

Source: Google

  • Gemma 3 hits 100+ million downloads, offering lightweight models from 1B to 27B parameters, optimized for various devices like phones to workstations (only needs one GPU to run!)

  • Outperforms rivals like Llama3-405B in early tests, providing enhanced experiences on a single GPU/TPU.

  • Supports 35+ languages, pre-trained in 140, and features complex reasoning, 128k-token context, and automated task functions.

  • Developed with safety protocols, it includes an image safety model for content labeling across categories like danger and violence.

  • Integrates with popular AI tools and offers deployment options optimized for NVIDIA GPUs.

  • The Gemma 3 Academic Program provides $10K in Google Cloud credits to researchers for innovative projects.

๐Ÿ–ผ๏ธ AI ART

Examples of great and trending AI art

Check out this unique twist on "The Wizard of Oz" with a cyberpunk edge, created by Noggahidez using Midjourney.

๐Ÿคฟ DEEP DIVE

Captions Launches Mirage, the First Video Foundation Model for Ultra-Realistic Talking Videos

Intelligence: Captions have announced the launch of Mirage, a groundbreaking video foundation model capable of generating hyper-realistic talking videos from audio files or scripts, without relying on actors or pre-recorded footage.

  • It is the first model specifically designed for creating user-generated content (UGC) style ads and talking content, making it a unique advancement in AI video technology.

  • Mirage offers full-body and facial motion generation that conveys emotion naturally, improving on existing lip-syncing technologies. Users can create videos based on an audio file alone or with minimal visual prompts.

  • Customize the speaker's appearance, including age, gender, clothing, and background through text prompts.

  • Supports video generation in over 29 languages, facilitating authentic content tailored to various markets.

  • Mirage is available for use within the Captions Ad Studio on desktop, targeting brands and marketers to enhance ad performance and creativity.

โš’๏ธ TOOL SNAPSHOTS

Futuristic tools within AI, no-code, and productivity

  • ๐ŸŽ™๏ธ Wispr Flow for Windows - Speak to type, triple your speed. Free option available.

  • ๐ŸŽฎ Break Me - Mind-clearing relaxation through unexpected fun surprises. Free to use.

  • ๐Ÿ“‹ Hoop - Consolidate and track tasks from meetings, email, Slack automatically. Free to try.

  • ๐ŸŒ Yadaphone - Effortless international calls from your browser. Payment required.

  • ๐Ÿ” Read - AI-driven unified search for efficient workplace insights. Free option available.

๐Ÿ“ฐ TOP NEWS

News on AI, no-code, automation, and productivity

OpenAI has called for a U.S. ban on AI models from Chinese lab DeepSeek, citing concerns over state control, data security, and intellectual property risks. The proposal claims DeepSeek operates under Chinese government influence, though no direct link has been confirmed.

Microsoft is introducing AI-powered summaries in Notepad, allowing users to generate concise text summaries with a right-click or shortcut. The update, available for Windows Insiders, also includes a recent files view and an improved Snipping Tool for cleaner annotations.

Patronus AI introduces its Judge-Image tool, the first multimodal AI judge designed to evaluate image captions and detect hallucinations. Already used by Etsy, the tool helps improve the accuracy and reliability of AI-generated content across various industries.

Moonvalley unveils Marey, a clean generative video model designed to give filmmakers full control over their creative process. By providing granular adjustments and respecting creators' work, the technology aims to revolutionize professional video production and empower filmmakers with limited resources.

Browser Use, an AI tool designed to enhance web accessibility for agentic applications, has seen a significant surge in downloads following its mention in a post about the Manus platform. This tool helps AI models interact with website elements and perform tasks autonomously, fueling the rise of web agents.

โ„น๏ธ ABOUT US

The Intelligent Worker helps you to be more productive at work with AI, automation, no-code, and other technologies.

We like real, practical, and tangible use-cases and hate hand-wavy, theoretical, and abstract concepts that donโ€™t drive real-world outcomes.

Our mission is to empower individuals, boost their productivity, and future-proof their careers.

We read all your comments - please provide your feedback!

Did you like today's email?

Your feedback is more valuable to us than coffee on a Monday morning!

Login or Subscribe to participate in polls.

What more do you want to see in this newsletter?

Please vote

Login or Subscribe to participate in polls.