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:

• 🎥 TED Talk: Why the AI Revolution is Underhyped | Eric Schmidt

• ⚡ Live Panel: From Vision to Reality: Implementing Voice AI Agents for CX

• 🔍 Vision-Language Models Can Help Doctors Reason Through X-Rays

• ‼️ Download the State of Voice Report 2025!

• 🎨 High-Fidelity Image Inpainting with Multimodal Guided GAN Inversion

• 🐦 Social Media Buzz: Zuckerberg discusses AI, China, and AGI 

• 🤖 NVIDIA’s Jensen Huang discusses plans for physical AI and robotics

• 🏥 Satya Nadella on Stanford Medicine’s agentic AI integrations

• 🎙️ New AI Minds Podcast with Mac Klinkachorn, Co-Founder & President at Trellis

• 🧠 How Stanford Teaches AI-Powered Creativity in Just 13 Minutes

• 📝 Exploring the TextAttack Framework: Components, Features, and Practical Applications

• 👾 Tutorial and Demo: How to Build a Voice AI Video Game in Python

• 📚 Entropy in Machine Learning: Deep DiveThanks for letting us crash your inbox; let’s party. 🎉

🎥  The AI Revolution is Underhyped | Eric Schmidt

Video Description: “The arrival of non-human intelligence is a very big deal, says former Google CEO and chairman Eric Schmidt. In a wide-ranging interview with technologist Bilawal Sidhu, Schmidt makes the case that AI is wildly underhyped, as near-constant breakthroughs give rise to systems capable of doing even the most complex tasks on their own. He explores the staggering opportunities, sobering challenges and urgent risks of AI, showing why everyone will need to engage with this technology in order to remain relevant.”

⚡ Live Panel: From Vision to Reality: Implementing Voice AI Agents for CX

Live Panel: From Vision to Reality: Implementing Voice AI Agents for CX - According to the 2025 State of Voice AI Report, 52% of organizations see customer service automation as voice AI's most transformative use case, yet only 15% are actively developing voice AI agents. Join Opus Research, Cognigy, and Deepgram as we explore how forward-thinking organizations are implementing sophisticated agentic capabilities, overcoming barriers to adoption, and tailoring voice AI to specific industry needs.

Save your seat here!

Where: Online

When: May 29th at 9AM PT / 11AM CT / 12PM ET / 4PM GMT

🔍 Vision-Language Models Can Help Doctors Reason Through X-Rays and how GAN Inversion Can Fix Images

MedVLM-R1: Incentivizing Medical Reasoning Capability of Vision-Language Models (VLMs) via Reinforcement Learning - Although Medical Visual Language Models (VLMs) show promise for radiological tasks, most existing VLMs merely produce final answers without revealing the underlying reasoning. To address this gap, this paper introduces MedVLM-R1, a medical VLM that explicitly generates natural language reasoning to enhance transparency and trustworthiness.

High-Fidelity Image Inpainting with Multimodal Guided GAN Inversion- Generative Adversarial Network (GAN) inversion have demonstrated excellent performance in image inpainting that aims to restore lost or damaged image texture using its unmasked content. This paper proposes a novel GAN inversion approach, dubbed MMInvertFill, for image inpainting.

‼️ Download the State of Voice Report 2025!

If you haven’t yet already, check out the full 2025 State of Voice Report!

We at Deepgram have partnered with Opus Research to examine the state of voice AI. What does the market look like? What can voice AI actually accomplish? What industries and companies are integrating implementing voice AI?

Long story short, 2025 seems to be the year of human-like voice AI agents. If you’d like to take a look at the data, statistics, and analyses yourself, you can check out the full report here.

🐝 Social Media Buzz: Zuckerberg discusses AI, China, and AGI

🎙️ The AI Minds Podcast!

Mac Klinkachorn, Co-Founder & President at Trellis. Trellis automates manual PDF tasks at scale with AI agents. Trellis integrates end-to-end with your systems of record and allows users to define infinitely complex workflows to capture their business processes.

🤖 Bonus Bits and Bytes!

• 🧠 How Stanford Teaches AI-Powered Creativity in Just 13 Minutes - “Stanford's Jeremy Utley reveals how and why ‘most people are not fully utilizing AI's potential.’ 

• 📝 Exploring the TextAttack Framework: Components, Features, and Practical Applications - Learn about this NLP Python framework built for adversarial attacks, data augmentation, and adversarial training.

• 👾 Tutorial and Demo: How to Build a Voice AI Video Game in Python - In this article, we are going to transform an ordinary platformer game into one that can be controlled by your voice using Deepgram’s API.

• 📚 Entropy in Machine Learning: Deep Dive - This article dives deep into the essence of entropy within machine learning, unraveling its significance, from the foundational theories to its practical applications.