What’s next for AI? How are large language models evolving? What does reinforcement learning mean for the future of AI training? These are the kinds of questions we find ourselves asking every day. AI discussions are everywhere, from research papers to boardrooms and headlines, but meaningful conversations that break down its real impact are harder to find.

That’s why we’re launching Gradient Descent, a podcast that dives deep into AI’s most important questions, one episode at a time. Hosted by Vishnu Vettrivel (Founder & CTO) and Alex Thomas (Principal Data Scientist), this series will explore AI models, training methodologies, industry impact, and the research driving the future of artificial intelligence. You can subscribe now on your favorite podcast platform.

Communicating AI’s Complexity

Artificial intelligence is one of the most transformative technologies of our time, but keeping up with its rapid advancements can be overwhelming, even for those working in the field. 

With Gradient Descent, our goal is to cut through the noise and engage in structured, research-backed discussions about the latest developments in AI. The series will focus on how new models are built, the choices behind their training methodologies, and their implications for businesses and society.

Each episode features a deep dive into a specific AI topic, covering both the technical foundation and the broader impact. After months of curating topics that highlight the excitement and challenges of AI research today, we’re proud to bring you a podcast designed with you in mind. 

Whether you’re an AI professional looking to deepen your knowledge or someone seeking a clearer understanding of AI’s real-world implications, Gradient Descent is here to provide valuable insights. We aim to make complex AI topics accessible while offering technical depth that helps you stay ahead of the curve in an ever-evolving field.

What to Expect in Season One

In each episode, you’ll hear from Vishnu and Alex as they break down the research, models, and industry shifts defining AI today.

The first season will feature topics like:

• AI Model Architectures: How new large language models are designed, trained, and deployed.

• The Economics of AI: How scaling laws, GPU shortages, and major tech investments shape the industry.

• Training Methodologies: The role of reinforcement learning (RL), supervised fine-tuning (SFT), and hybrid approaches.

• Reliability & Ethics: Bias, safety considerations, and the trade-offs in AI model alignment.

In our first episode, we dive into DeepSeek, an emerging AI phenomenon from China. We explore:

• How DeepSeek’s training differs from traditional fine-tuning models

• The market impact of large-scale AI models

• What DeepSeek’s approach means for the future of AI research and deployment

And that’s just the beginning.

Where to Listen & How to Get Involved

The podcast is live on Apple Podcast, Spotify, and Amazon Music. To explore all episodes, visit the Podcast section of our website, where you’ll find every discussion in one place.

Each episode also comes with show notes featuring research papers, model breakdowns, and recommended reading. Whether you’re looking for a deep dive or a quick overview, everything you need is right there.

Subscribe to the Gradient Descent Newsletter for more updates and behind-the-scenes content.

Why Did We Start This Podcast?

AI is evolving faster than ever, and we find ourselves constantly trying to make sense of the latest breakthroughs, separating real progress from hype. While there are many discussions about AI, they often feel either too broad or too technical. We wanted to create a space where professionals and enthusiasts could explore AI deeply without needing a PhD to follow along.

Gradient Descent is our way of breaking down AI’s most pressing topics through structured, research-backed discussions. This podcast is dedicated to understanding the 'why' behind the research, the choices driving AI development, and its real-world impact. Most importantly, we want to engage with the AI community, answer your questions, and explore the topics that matter most to you. 

Who is This Podcast For?

While there are plenty of excellent AI podcasts out there, many tend to be either too general or too technical. Gradient Descent is designed for AI professionals, researchers, engineers, and anyone eager to gain a deeper understanding of AI without having to sift through complex research papers on their own.

What Topics We Will Cover?

We focus on how AI research is done, not just its outcomes. There’s a lot of discussion about AI’s risks and potential, but fewer conversations about what goes into designing, training, and deploying these systems. Here are the topics we want to cover:

• How scaling laws, compute availability, and major tech investments shape AI’s growth and accessibility.

• The financial impact of AI advancements, from stock movements to industry shifts driven by major breakthroughs.

• The long-term implications of AI development, identifying who benefits and who gets left behind.

• The trade-offs in model reliability, bias mitigation, and ethical considerations in AI deployment.

• How AI is transforming businesses, automation, and the global workforce across various sectors.

• The latest advancements in AI theory, model efficiency, and real-world applications that drive the field forward.

Who Should We Interview?

For now, the podcast features AI experts, Vishnu and Alex, but we’re open to expanding. Let us know if there are researchers, engineers, or industry leaders you’d like to hear from!

Your Feedback Shapes Future Episodes

While we plan to make more episodes, we need your feedback. What do you want to learn more about? What’s missing from AI discussions today? Send us your thoughts, and we’ll shape future episodes based on what you find most valuable.

Feel free to contact us by email at oksana@wisecube.ai.

We’re excited to launch this podcast and can’t wait for you to join us.