In today’s email:

  • 👀 Sam Altman on OpenAI, Future Risks and Rewards, and Artificial General Intelligence

  • 🧸 Grimes is working on an interactive AI toy for kids. Meet Grok.

  • 🧠 A Portable, non-invasive, mind-reading AI turns thoughts into text

  • 🧰 14 new AI-powered tools and resources. Make sure to check the online version for the full list of tools.

Top News

The new research by OpenAI in the field of superhuman AI alignment discusses “superalignment.” The primary challenge in this area is supervising AI systems that are significantly smarter than humans. The team behind this research proposes an analogy to study this problem: using smaller, less capable models to supervise larger, more capable ones.

The research shows that a GPT-2-level model can be used to supervise GPT-4, eliciting performance close to GPT-3.5, even on complex problems where the small model fails. This demonstrates that strong models can generalize according to the weak supervisor’s intent, leveraging their capabilities beyond the supervisor’s limitations.

The article addresses the limitations of current alignment methods like reinforcement learning from human feedback (RLHF), which may not scale well to superhuman models. The proposed method, however, shows promise in improving weak-to-strong generalization, suggesting a potential pathway for aligning future superhuman AI models.

To foster further research in this area, the team is releasing open-source code for weak-to-strong generalization experiments and launching a $10 million grants program for research in superhuman AI alignment. The goal is to ensure that advanced AI systems remain safe and beneficial to humanity, and this new research direction offers a promising avenue for empirical progress in AI alignment.

Frustrated by biased news? Tired of trying to sift through 100s of news sources to find the facts? Leave behind the opinions and receive a daily digest of pure facts. Say hello to 1440! Delve into politics, business, science, sports, culture, and more. Join the other 3 million Americans who read 1440 each morning. 

Virtual try-on has become a transformative technology, empowering users to experiment with fashion without ever having to physically try on clothing. However, existing methods often struggle with generating high-fidelity and detail-consistent results. Diffusion models have demonstrated their ability to generate high-quality and photorealistic images, but when it comes to conditional generation scenarios like virtual try-ons, they still face challenges in achieving control and consistency. Outfit Anyone addresses these limitations by leveraging a two-stream conditional diffusion model, enabling it to adeptly handle garment deformation for more lifelike results. It distinguishes itself with scalability—modulating factors such as pose and body shape—and broad applicability, extending from anime to in-the-wild images. Outfit Anyone’s performance in diverse scenarios underscores its utility and readiness for real-world deployment.

Instagram introduced a new AI-powered background editing tool for U.S.-based users, allowing them to change image backgrounds in Stories using prompts. Meta’s generative AI lead, Ahmad Al-Dahle, announced this feature, which includes ready prompts like “On a red carpet” or “Surrounded by puppies,” and also allows users to write their own. The tool adds a “Try it” sticker to Stories with these backgrounds, encouraging others to use the feature.

Snapchat also released a tool for paying users to create and send AI-generated images. They previously launched a feature called Dreams for creating themed images, and allowed Snapchat+ users to use prompt-based backgrounds.

Additionally, Meta expanded the availability of its 28 AI-powered characters to all its apps in the U.S., including WhatsApp, Messenger, and Instagram. These characters support Bing search and have an improved context window. Meta also launched a standalone AI-image generator, Imagine with Meta, powered by their Emu model.

Other stuff

GPT Store now in Superpower ChatGPT. 🎉

1000s of custom GPTs right inside ChatGPT and adding more every day

Superpower ChatGPT Extension on Chrome
Superpower ChatGPT Extension on Firefox
Tools & LinkS

Editor’s Pick ✨

Music.AI – The single source for multiple state-of-the-art AI models

Collov AI – Discover the Future of Home Design

Promptbase is an evolving collection of resources, best practices, and example scripts for eliciting the best performance from foundation models like GPT-4.

Tynal – Draft contracts 10x faster

Labophase – Generate text responses from every top AI model in one place. – AI based platform, over 1500 AI tools, no coding

TubeOnAI – Summarize & Listen to YouTube & Podcasts in 30 Seconds

Unwrapped AI – Spotify-like recap for your Github contributions in 2023

Music.AI – The single source for multiple state-of-the-art AI models

WooWell – Generative AI-powered dating solution for modern men

Delphi – Use AI to build the digital version of yourself

Heatseeker AI – Build product conviction with data and testing in the market

ChatGPT Tips and Tricks – Part 2: Open Plugins by default, Hide example prompts, and pin messages

Unclassified 🌀 

⚡️ Be the Highlight of Someone’s Day – Think a friend would enjoy this? Go ahead and forward it. They’ll thank you for it!

Hope you enjoyed today’s newsletter

⚡️ Join over 200,000 people using the Superpower ChatGPT extension on Chrome and Firefox.

Superpower ChatGPT Extension on Chrome
Superpower ChatGPT Extension on Firefox

Source link

What's Your Reaction?

hate hate
confused confused
fail fail
fun fun
geeky geeky
love love
lol lol
omg omg
win win
The Obsessed Guy
Hi, I'm The Obsessed Guy and I am passionate about artificial intelligence. I have spent years studying and working in the field, and I am fascinated by the potential of machine learning, deep learning, and natural language processing. I love exploring how these technologies are being used to solve real-world problems and am always eager to learn more. In my spare time, you can find me tinkering with neural networks and reading about the latest AI research.


Your email address will not be published. Required fields are marked *