🤝 AidfulAI Newsletter #7: Elevate Your Podcast Experience with AI Notes
Dear curious minds,
Welcome to the ultimate newsletter for those interested in Artificial Intelligence (AI) and Personal Knowledge Management (PKM).
Major AI News
🎥🔮 Gen-2: This AI turns your words into videos
Runway, a leading creative AI research lab known for its participation in the initial release of the text-to-image AI Stable Diffusion, has unveiled Gen-2, the next iteration of their text-to-video generator.
This AI system can craft novel videos from text prompts, images, or video clips. With Gen-2, you can simply input a text prompt and the system will transform your words into a consistent, realistic video in any style you envision. While the technology continues to evolve, and the results might not be perfect, it's a significant step forward to realistic video synthesis. Currently, Gen-2 can produce videos up to four seconds in length, and for longer videos, you have to stitch together generated segments.
To learn more about Gen-2, explore videos it has produced and to try it for free, visit the Runway website. Besides their webpage, you can use the text-to-video AI Gen-2 also in the official iOS app.
Small sidenote for my colleagues at TOMRA: Take a look at the company logo of Runway. It is remarkably similar to the TOMRA logo. 🤔
🎨🔄 StyleDrop: A new AI technique by Google for easy style transfer in images
Google Research published a research paper describing a new technique called StyleDrop. The approach can quickly and easily transfer the style of one image or design to another. To use StyleDrop, simply upload an image or design that you want to use as the source style. StyleDrop will then generate new images for your text inputs in this style.
StyleDrop is still under development and currently not accessible by the public, but it has the potential to be a powerful tool for designers, artists, and anyone who wants to create visually appealing content.
More about StyleDrop and especially a lot more stunning examples on the corresponding website and in the paper linked above.
🔖🤖 Unmasking AI: EU Calls for Clear Labelling of AI-Generated Content
The EU has called on tech giants such as Google and Facebook to start labelling content and images that are created or manipulated by AI tools.
The move is part of a wider effort to tackle online disinformation and fake news. The EU wants users to be able to distinguish between real and AI-generated content, such as text, video, and audio. The EU also warned Twitter that it faces sanctions if it does not comply with new digital content laws that will come into force in August. The EU is a leader in regulating tech companies and is developing new laws on AI. The EU hopes that by labelling AI-generated content, it will help users make informed decisions and prevent the spread of harmful or misleading information, as stated by The Guardian.
Privacy-Friendly AI
🚀 🤖 The LLM RedPajama 7B is Here to Make AI More Accessible
Together, a company that develops open-source generative AI models, has released RedPajama 7B, a large language model (LLM) with 7 billion parameters.
Besides the base model, there are fine-tuned versions for instruction following and chat. The models are available for free and can be used locally on your PC. RedPajama 7B was trained on a massive dataset of text and code, and it can generate text, translate languages, write different kinds of creative content, and answer your questions informative. It has been benchmarked against other open-source 7B LLMs, and it outperforms them on several tasks. You can give it a try without installing anything and create feedback to improve the open-source AI research on the OpenChatKit hosted on HuggingFace.
Together is committed to making AI accessible to everyone. The company believes that open-source LLMs are essential for innovation and creativity, and it is working to make them more widely available.
PKM and AI
🎧📝 Get Tailored Podcast Summaries with Snipd's New AI Podcast Notes
In the last issue, I shared with you how you can use AI to create tailored book summaries from your highlights, based on the advice of Tiago Forte.
Today, I have some exciting news to share with you. I've been given first access to a new feature from my favorite podcast player, Snipd. The feature is called AI Podcast Notes, and it creates after listening to an episode a summary from the highlights you created by pressing the back button on your headphone or car entertainment system.
To use AI Podcast notes you have to enter a waitlist, but I received a special sign-up link for my audience which lets you skip the line. Once you're activated, new highlights you create will be used to create a summary. This is triggered when the episode is marked as completed, which happens automatically after you finish listening. You'll receive an email with your personal AI Podcast Notes within 30 minutes. Furthermore, you can also see your AI Podcast Notes in the app if you go to the “Saved” screen and select a show followed by an episode. There, a button named “✨ Your episode takeaways" is shown.
A few additional insights from my exchange with the Snipd team:
If you edit the snip titles manually, they are passed to the AI and may influence the summary.
The feature is triggered by setting the complete flag, which happens automatically when you finish listening to an episode. If you edit your snips afterward, this will not be reflected in the AI Podcast Notes. However, it is planned to allow re-triggering the creation in the future.
The URLs in the email are unique and cannot be guessed by anyone else.
One last thing: check your spam folder if you don't receive your AI Podcast Notes email. For me, the first summaries ended up there. And now, if not already done, go to the special sign-up link and start getting ahead by receiving tailored AI notes for the podcast episodes you listened to.
Podcasts
🤖✍️ How Notion AI is Redefining AI Interfaces
In the podcast episode “Building the AI × UX Scenius — with Linus Lee of Notion AI”, the team at Latent Space hosted Linus, a key member of the Notion AI team. Linus joined the Notion AI team in January, contributing to a series of discussions and articles on the challenges and possibilities of AI interfaces, including the limitations of text-generation tasks and prompt engineering.
🎧📰 Get More AI News with the “Last Week in AI” Podcast
If you want to get more insights from AI news than I provide in the first section of this newsletter, you should take a look at the “Last Week in AI” podcast.
The summary of my highlights from the 125th episode created by Snipd are shared here.
Disclaimer: This newsletter is written with the help of AI. I use AI as an assistant to generate and optimize the text. However, the amount of AI used varies depending on the topic and the content. I always curate and edit the text myself to ensure quality and accuracy. The opinions and views expressed in this newsletter are my own and do not necessarily reflect those of the sources or the AI models.