Has AI hallucination gotten out of control? In episode 54 of Mixture of Experts, host Tim Hwang is joined by&nbsp;Kate Soule, Skyler Speakman and Kaoutar El Maghraoui to analyze reasoning models and rising hallucinations.&nbsp;Next, as IBM Think 2025 wraps, the experts unpack the biggest highlights from IBM’s biggest show of the year:&nbsp;new AI agents, Ferraris and ... penguins? Then, OpenAI is making moves&nbsp;with&nbsp;its acquisition of Windsurf. What does this mean? Tune in&nbsp;to this week’s&nbsp;Mixture of Experts for more!&nbsp; 00:01 – Intro&nbsp;01:12 – IBM Think 2025&nbsp;09:27 – Reasoning models and hallucinations&nbsp;19:23 – OpenAI Windsurf acquisition&nbsp;&nbsp;&nbsp;The opinions expressed in this podcast are solely those of the participants and do not necessarily reflect the views of IBM or any other organization or entity.&nbsp;

Kate Soule

Director, Technical Product Management, Granite 

Skyler Speakman

Senior Research Scientist 

Kaoutar El Maghraoui

Principal Research Scientist and Manager, Hybrid Cloud Platform 

IBM Think 2025, OpenAI acquires Windsurf, and does AI reasoning mean more hallucination? Tune-in to this week’s Mixture of Experts to find out!

Episode 54:  IBM Think 2025, OpenAI Windsurf acquisition, reasoning models and hallucinations

IBM Think 2025, OpenAI Windsurf acquisition, reasoning models and hallucinations 

We are celebrating MoE podcast’s one year anniversary! In episode 53 of Mixture of Experts, host Tim Hwang is joined by the O.G. panel of experts from our pilot—Chris Hay, Shobhit Varshney and Kush Varshney. This week, we cover some exciting announcements at LlamaCon. Then, we discuss some new Chinese AI models from Qwen3 to the rumored DeepSeek-R2. Next, J.P. Morgan’s CISO, Patrick Opet, released “An open letter to our third-party suppliers,” covering the need for AI security. Are we doomed? Finally, we look back at some of the topics we discussed in episode 1—the Rabbit AI device, GPT-2 chatbot, Apple Intelligence—after all that, who was the first person to say “agents” on the podcast? &nbsp;Tune in to find out, on today’s one-year celebration of Mixture of Experts.&nbsp; 00:00 -- Intro00:38 -- LlamaCon10:34 -- Qwen3 and DeepSeek-R223:23 -- J.P. Morgan’s open letter&nbsp;39:45 -- One year of MoE The opinions expressed in this podcast are solely those of the participants and do not necessarily reflect the views of IBM or any other organization or entity.&nbsp;

Chris Hay

Distinguished Engineer, CTO Customer Transformation 

Kush Varshney

IBM Fellow, AI Governance 

Shobhit Varshney

Head of Data and AI for the Americas 

It's our 1st birthday! Today on the MoE podcast we cover LlamaCon, Qwen3, DeepSeek-R2 and J.P. Morgan’s open letter on AI security.

LlamaCon, Qwen3, DeepSeek-R2 rumors and JP Morgan’s open letter on AI

Is OpenAI going to enter the social media game? In episode 52 of Mixture of Experts host, Tim Hwang is joined by Gabe Goodhart, Kate Soule and Marina Danilevsky. First, Sam Altman is rumored to be testing an internal prototype social network; why is this a potential next move for the AI giant? Next, for our paper of the week, we analyze Anthropic’s study on chain-of-thought reasoning, “Reasoning Models Don’t Always Say What They Think.” Then, AI scraping puts a strain on Wikimedia; what’s the impact of this? Finally, China held a humanoid robot half-marathon, where humans raced alongside robot competitors. Who wins this AI race? All that and more on today’s Mixture of Experts.&nbsp; 00:41 -- OpenAI social network&nbsp;10:02 -- Anthropic’s reasoning study&nbsp;20:56 -- AI bots strain Wikimedia&nbsp;31:33 -- Humanoid half-marathon&nbsp; &nbsp;The opinions expressed in this podcast are solely those of the participants and do not necessarily reflect the views of IBM or any other organization or entity.&nbsp;

Director, Technical Product Management, Granite

Gabe Goodhart

Chief Architect, AI Open Innovation

Marina Danilevsky

Senior Research Scientist

Episode 52 of Mixture of Experts covers OpenAI’s social network, Anthropic’s reasoning study, AI bots on Wikimedia, and a humanoid robot half-marathon.

OpenAI social network, Anthropic’s reasoning study and humanoid half-marathon

OpenAI just dropped o3 and o4-mini! In episode 51 of Mixture of Experts host, Tim Hwang is joined by Chris Hay, Vyoma Gajjar and special guest John Willis, Owner of Botchagalupe Technologies. Today, we analyze Sam Altman’s new AI models, o3 and o4-mini. Next, Google announced that by Q3 you can run Gemini on-prem; what does this mean for enterprise AI adoption? Then, John is on the show today to take us through AI evaluation tools and why we need them. Finally, NVIDIA is planning to move AI chip manufacturing to the U.S. Can they pull this off? All that and more on today’s Mixture of Experts.&nbsp; 00:01 – Intro&nbsp;00:56 – OpenAI o3 and o4 mini&nbsp;14:57 – Google Gemini on-prem&nbsp;23:43 – AI evaluation tools&nbsp;34:59 – NVIDIA's U.S. chip manufacturing&nbsp;&nbsp;&nbsp;The opinions expressed in this podcast are solely those of the participants and do not necessarily reflect the views of IBM or any other organization or entity.&nbsp;

Distinguished Engineer, CTO Customer Transformation

Vyoma Gajjar

AI Technical Solutions Architect 

John Willis

Author, Owner of Botchagalupe Technologies 

In episode 51 of Mixture of Experts, we cover OpenAI o3 and o4-mini, Google Gemini running on-prem, AI evaluation tools and NVIDIA bringing AI chip production back to the U.S. 

o3 and o4-mini, Google Gemini on-prem and NVIDIA’s U.S. chip manufacturing

IBM z17 is here! In episode 50 of Mixture of Experts, host Tim Hwang is joined by Kate Soule, Shobhit Varshney and Hillery Hunter to debrief the launch of a new mainframe with robust AI infrastructure. Next, Meta dropped Llama 4 over the weekend;, how's it going? Then, Shobhit is recording live from Google Cloud Next in Las Vegas, along with Gemini 2.5 Pro. What are some of the most exciting announcements? Finally, the Pew Research Center shows perception of AI, how does this impact the industry? All that and more on today’s 50th Mixture of Experts.&nbsp; 00:01 -- Intro&nbsp;00:55 -- IBM z17&nbsp;11:42 -- Llama 4&nbsp;25:02 -- Google Cloud Next 2025&nbsp;34:29 -- Pew's research on perception of AI&nbsp; The opinions expressed in this podcast are solely those of the participants and do not necessarily reflect the views of IBM or any other organization or entity.&nbsp; Explore the new features of IBM z17: https://www.ibm.com/products/z17 Read the Pew Research: <a href="https://www.pewresearch.org/internet/2025/04/03/how-the-us-public-and-ai-experts-view-artificial-intelligence/" rel="noopener noreferrer" target="_blank">https://www.pewresearch.org/internet/2025/04/03/how-the-us-public-and-ai-experts-view-artificial-intelligence/</a>&nbsp;&nbsp;Subscribe for AI updates: https://ibm.biz/Think_newsletter Visit Mixture of Experts podcast page to learn more AI content: <a href="https://www.ibm.com/think/podcasts/mixture-of-experts" rel="noopener noreferrer" target="_blank">https://www.ibm.com/think/podcasts/mixture-of-experts</a>&nbsp;

VP & Sr. Partner, Americas Head of Data & AI 

Hillery Hunter

IBM Fellow, CTO IBM Infrastructure 

In episode 50 of Mixture of Experts, we debrief many announcements: IBM z17, Meta Llama 4, Gemini 2.5 Pro and more from Google Cloud Next!

AI on IBM z17, Meta's Llama 4 and Google Cloud Next 2025

Will OpenAI be fully open source by 2027? In episode 49 of Mixture of Experts, host Tim Hwang is joined by Aaron Baughman, Ash Minhas and Chris Hay to analyze Sam Altman’s latest move towards open source. Next, we explore Anthropic's mechanistic interpretability results and the progress the AI research community is making. Then, can Apple catch up? We analyze the latest critiques on Apple Intelligence. Finally, Amazon enters the chat with AI agents. How does this elevate the competition? All that and more on today’s Mixture of Experts. 00:01 -- Introduction00:48 -- OpenAI goes open &nbsp;11:36 -- Anthropic interpretability results&nbsp;24:55 -- Daring Fireball on Apple Intelligence&nbsp;34:22 -- Amazon’s AI agents The opinions expressed in this podcast are solely those of the participants and do not necessarily reflect the views of IBM or any other organization or entity. Subscribe for AI updates: <a href="https://www.ibm.com/account/reg/us-en/signup?formid=news-urx-52120" rel="noopener noreferrer" target="_blank">https://www.ibm.com/account/reg/us-en/signup?formid=news-urx-52120</a>Learn more about artificial intelligence → <a href="https://www.ibm.com/think/artificial-intelligence" rel="noopener noreferrer" target="_blank">https://www.ibm.com/think/artificial-intelligence</a>Visit Mixture of Experts podcast page to learn more AI content → <a href="https://www.ibm.com/think/podcasts/mixture-of-experts" rel="noopener noreferrer" target="_blank">https://www.ibm.com/think/podcasts/mixture-of-experts</a>

Ash Minahs

Lead AI Advocate

Aaron Baughman

IBM Fellow, Master Inventor

OpenAI goes open, Anthropic on interpretability, Apple Intelligence updates and Amazon AI agents

OpenAI goes open, Anthropic mechanistic interpretability, Apple Intelligence updates and Amazon AI agents

What’s the best open-source model? In episode 48 of Mixture of Experts, host Tim Hwang is joined by Kate Soule, Kush Varshney and Skyler Speakman to explore the future of open-source AI models. First, we chat about the release of DeepSeek-V3-0324. Then, more announcements coming out of Google including Gemini Canvas and Gemini 2.5. Next, Extropic has entered the chat with a thermodynamic chip. Finally, AI image generation is on the rise as OpenAI released GPT-4o image generation. All that, and more on today’s Mixture of Experts. 00:01 – Intro 00:42– DeepSeek-V3-0324 09:48 – Gemini 2.5 and Canvas 21:27– Extropic’s thermodynamic chip 30:20 – OpenAI image generation The opinions expressed in this podcast are solely those of the participants and do not necessarily reflect the views of IBM or any other organization or entity.

What’s the best open-source model? In episode 48 of Mixture of Experts we discuss a new release of DeepSeek V3, Google’s Gemini 2.5 and Canvas, Extropic’s thermodynamic chip, and OpenAI’s GPT-4o image generation.

DeepSeek-V3-0324, Gemini Canvas and GPT-4o image generation

Podcasts at IBM

Learn more about AI

Welcome to Mixture of Experts, your weekly deep dive into the ever-evolving landscape of artificial intelligence—bringing you insightful discussions on the latest AI trends, innovations, and their impact on business. From breakthrough research to practical applications, each episode offers a balanced blend of expertise and analysis. Explore how AI is reshaping industries, driving efficiency, and unlocking new opportunities for growth. Whether you're a seasoned professional seeking to stay ahead of the curve or an enthusiast curious about the future of technology, Mixture of Experts delivers the perfect mix of insights and practical knowledge. Tune in and stay informed as we navigate the dynamic intersection of AI and business.

Episode 54: IBM Think 2025, OpenAI Windsurf acquisition, reasoning models and hallucinations

DESCRIPTION

Today's Host

Tim Hwang

Today's Guests

Kate Soule

Skyler Speakman

Kaoutar El Maghraoui

Recent Episodes

Episode 54: IBM Think 2025, OpenAI Windsurf acquisition, reasoning models and hallucinations

DESCRIPTION

Today's Host

Tim Hwang

Today's Guests

Kate Soule

Skyler Speakman

Kaoutar El Maghraoui

Related Resources

Recent Episodes

LlamaCon, Qwen3, DeepSeek-R2 rumors and JP Morgan’s open letter on AI

OpenAI social network, Anthropic’s reasoning study and humanoid half-marathon

o3 and o4-mini, Google Gemini on-prem and NVIDIA’s U.S. chip manufacturing

AI on IBM z17, Meta's Llama 4 and Google Cloud Next 2025

OpenAI goes open, Anthropic on interpretability, Apple Intelligence updates and Amazon AI agents

DeepSeek-V3-0324, Gemini Canvas and GPT-4o image generation