- The Rundown AI
- Posts
- 🤖 OpenAI's newest product update!
Sign Up | Advertise | Tools | View Online
Welcome, AI enthusiasts.
We just witnessed another classic Tuesday in the world of AI. Huge developments from Meta, Open-sourced AI, OpenAI, Midjourney, and Snap. Let’s get into it.
In today’s rundown:
🗣️ Speak 100 languages with Meta AI
🎥 CoDeF makes AI video edits seamless
🛠️ 9 new AI tools
🔌 5 game-changing Figma AI plugins
🤖 3 quick AI updates
Read time: 4 minutes
LATEST DEVELOPMENTS
META
Image source: Meta
The Rundown: Meta has open-sourced an AI model called SeamlessM4T, which translates and transcribes nearly 100 languages across text and speech.
Details:
SeamlessM4T combines speech recognition, translation, and text-to-speech in one model.
Text-to-speech you to switch languages mid-sentence.
The model was trained using tens of billions of scraped sentences and four million hours of speech data.
Why it matters: This multilingual model brings us one step closer to Star Trek-style universal translators — and continued advances could soon make global language barriers a thing of the past.
TOGETHER WITH INNOVATING WITH AI
The Rundown: Launching an AI product is tough, but Innovating with AI managed to launch two new AI products in 1 month.
Luckily, they created a free video lesson and cheat sheet to help you do the same, and are giving it away for free.
Here’s what’s inside:
The AI Business Model Matrix: How to combine your skills, interests, and advanced AI tools in unique and exciting ways
The AI Venture Mindset: How to use rapid experimentation and industry uncertainty to your advantage
The AI Marketing Cookbook: Where to share and promote your new AI-powered business for maximum growth
AI RESEARCH
Image Source: CoDeF Research Paper
The Rundown: Researchers have just developed CoDeF, an extremely impressive open-sourced AI system that allows for seamless style transfers and editing across video frames.
Key points:
CoDeF introduced a new method that captures the static elements of a video with a reference image and tracks how this image changes or morphs over time to match each video frame.
The system offers superior consistency across video frames than other methods and is capable of tracking non-rigid objects like water and smog.
The model enables easy application of image edits to entire videos and can track key points without training.
Noteworthy: AI has already unlocked some mind-blowing capabilities in video editing, but CoDeF might be the most impressive model we’ve seen so far. Check out some of the wild sample video outputs here.
TRENDING TOOLS
🖼️ OpenArt- Create stunning AI images without prompts (link)*
🌈 Rainbow AI- Precise weather forecasting with AI (link)
🪓 Pickaxe- Embed AI anywhere on the web (link)
📜 SEC Insights AI- Document analysis lightning fast (link)
📊 Speak AI- Turn your language data into insights, fast and with no code (link)
✍️ Rytr- An AI writing assistant that helps you create high-quality content, in just a few seconds (link)
✨ Magic Studio- Create amazing product photos in minutes (link)
🐦 TweetHunter- All-in-one Twitter growth tool to create content using AI, build an audience, and schedule posts (link)
🔮 Supertools- Organizes 100’s of AI tools for you in one spot (link)
* = This is sponsored content
AI TRAINING
Do you spend a lot of time in Figma? These AI plugins can give you a serious productivity boost:
FigGPT- Connect ChatGPT in Figma to generate and refine copy (link)
Automator- One-click task automation to speed up workflows (link)
Magician- AI-powered design tool for stunning visuals (link)
PixCut- Auto background remover for free (link)
AI Image Upscaler- Enables 8X upscaling with AI for crisp graphics (link)
Try weaving AI into your design process to unlock new creative possibilities and take your skills to the next level.
QUICK HITS
OpenAI just launched a massive product update, enabling fine-tuning for GPT-3.5 Turbo. It allows developers to customize models for better performance on specific tasks. Early tests show the fine-tuned GPT-3.5 Turbo can match GPT-4, with use cases including improved steerability, output formatting, and custom tone. Fine-tuning for GPT-4 is reportedly coming this fall.
Midjourney launched a new inpainting feature called Vary (Region), letting users easily modify parts of a generated image by highlighting a section and describing edits. It solves the headaches of re-generating or using external editing tools for minor tweaks. The new tool can be found as an option after upscaling an image on the platform.
Snapchat is back at it again, reportedly brewing a new AI feature called ‘Dreams’, allowing users to turn selfies into imaginative scenes. The code includes a ‘Dreams With Friends’ element for multiple users in a single image, along with purchasable ‘Dream Packs’ that hint at monetization down the road.
SPONSOR US
🦾 Get your product in front of over 275k+ AI enthusiasts
Our newsletter is read by thousands of tech professionals, investors, engineers, managers, and business owners around the world. Learn more or get in touch today.
THAT’S A WRAP
Keep an eye out and click the star emoji hidden somewhere in today’s email for a free entry into our AI starter pack giveaway (win a free M2 MacBook Air)!
Additionally, you’ll get another free entry for every subscriber who joins our newsletter using your unique link: https://www.therundown.ai/subscribe?ref=PLACEHOLDER
You can track how many referrals you’ve made here. Happy hunting!
If you have anything interesting to share, please reach out to us by sending us a DM on Twitter: @rowancheung & @therundownai ⭐️
How was today's newsletter?Feedback helps us improve! |
❤️ Rundown review of the day
Join the conversation