Welcome (back) to AI Minds, a newsletter about the brainy and sometimes zany world of AI, brought to you by the Deepgram editorial team.

In this edition:

• 🎥 Sam Altman's method for clear thinking

• 🧠 What norms human-AI cooperation should entail, according to researchers

• 🖼️ Improving AI-generated image detection by removing content bias

• 🐦 Social Media Buzz: The five levels of Agentic AI and Jensen Huang on IT’s future

• 👓 A Comprehensive Survey of AI-Generated Content (AIGC): from GAN to ChatGPT

• ⚡ Technical Webinar: How to Build Responsive Voice Agents with Vonage & Deepgram

• 📲 Three new, trending AI apps for you!

• 🎤 The AI Minds Podcast with Brent Pretty, CEO and Founder at Retellio

• 🎙️ Jensen Huang and Arthur Mensch on a16z podcast

• 🏥 Why AI is the Future of Medical Documentation

• ✏️ Harvard’s Guide to Writing Prompts for GenAI

• 📚Few Shot Learning: A Deep-Dive

Thanks for letting us crash your inbox; let’s party. 🎉

🎥 Sam Altman's Method for Clear Thinking

In this clip, the CEO of OpenAI goes over his note-taking habits and how he thinks about writing. Do you agree with his methods? Would his techniques work for your lifestyle and personality? Check out this video to find out!

🔍  What Norms Human-AI Cooperation Should Entail & How to Improve AI-generated Image Detection

Relational Norms for Human-AI Cooperation - As AI agents powered by large language models are increasingly designed to serve roles analogous to human positions—such as assistant, tutor, or even romantic partner—it is imperative to examine whether and how human relational norms should extend to human-AI interactions.

This analysis explores how differences between AI systems and humans, such as the absence of conscious experience and immunity to fatigue, may affect an AI's capacity to fulfill relationship-specific functions and adhere to corresponding norms.

SFLD: Reducing the content bias for AI-generated Image Detection - Our current methods for identifying AI-generated images are biased towards (read: against) unseen content, or are vulnerable to common image degradations, such as JPEG compression. To address these issues, this paper proposes a novel approach, SFLD, which incorporates PatchShuffle to integrate high-level semantic and low-level textural information.

🐝 Social Media Buzz: The Five Levels of Agentic AI and Jensen Huang on the Future of IT

⚡ Technical Deep Dive: How to Build Responsive Voice Agents with Vonage & Deepgram

You have one more day to sign up for this webinar!

If you want to learn how to build human-like voice agents for customer support, appointment scheduling and more, sign up for our March 26th technical webinar with Vonage.

When: Wednesday 26th March 2025, 10:00 PT / 12:00 ET / 17:00 GMT

Where: Online

Save your spot here!

Hosted by Benjamin Aronov (Developer Advocate at Vonage), Tony Chan (Senior Solutions Engineer at Vonage), and Damien Murphy (Applied Engineer at Deepgram).

📲 Three new, trending AI apps for you!

Coora AI represents a significant leap forward in the world of e-commerce and artificial intelligence. This innovative startup has carved out a niche for itself by creating AI-driven applications that transform the way consumers and businesses engage with online shopping. At the core of Coora AI's offerings is the AI Discovery Shopping app, which personalizes the shopping experience by tailoring product recommendations to individual preferences and behaviors.

GAJIX is an AI-powered learning assistant created by [company name]. It uses advanced natural language processing algorithms to analyze text and provide insights, summaries, and learning tools on any subject or webpage.

PandaChat-AI is an AI-powered platform that allows users to upload any data like documents, presentations, spreadsheets, images, websites etc. and ask questions about the content. It uses natural language processing to provide quick answers by analyzing the uploaded data.

🎙️ The AI Minds Podcast

Brent Pretty, CEO and Founder at Retellio. Retellio distills thousands of hours of calls into 30 minute podcasts. You don't have to stay tethered to your computer. You can listen to customers where and when it's convenient for you.

In this episode, Brent Pretty shares his unconventional journey from growing up in a small town in Newfoundland, dropping out of high school, and eventually finding his way into strategy consulting after a professor recognized his talent for strategic thinking.

🤖 Bonus Bits and Bytes!

• 🎤 Jensen Huang and Arthur Mensch on a16z podcast: The founder of NVIDIA and the cofounder of Mistral sit down to discuss sovereign AI, national IA strategies, and why every country must take ownership of its digital intelligence.

• 🏥 Why AI is the Future of Medical Documentation - AI-powered solutions like improve efficiency, reduce errors, and allow healthcare providers to focus more on patient care. Here’s how.

• ✏️ Harvard’s Guide to Writing Prompts for GenAI - In about 1000 words, this guide will help you improve your prompt-writing skills. While it’s not an in-depth guide to advance prompt engineering, these rules are crucial to know.

• 📚Few Shot Learning - This glossary entry discusses the fundamentals of few shot learning, uncovering its principles, methodologies, and real-world applications