The Data Exchange with Ben Lorica

The SaaSpocalypse Is Coming — But Don't Count Out the Incumbents

AI Agents Are Implemented, Not Adopted

Why Your AI Agent Isn't Ready to Ship (And How to Know When It Is)

Why Foundation Models Haven’t Replaced Classical Machine Learning

Agentic systems are essential for transforming complex, fragmented enterprise data into production-ready machine learning models, particularly for classical tasks like forecasting and fraud detection where foundation models often underperform. Disarray addresses the persistent challenge of data silos by constructing a ...

When "Garbage In, Garbage Out" Gets It Wrong

As Code Generation Speeds Up, Who Tests the Output?

The rapid proliferation of AI-generated code has created a significant verification bottleneck, as traditional testing methods struggle to keep pace with increased development velocity. Ito AI addresses this challenge by integrating automated, agentic quality assurance directly into the pull request workflow. By execut...

The Gap Between AI Hype and Enterprise Reality

Reading the Tea Leaves: What the World's Top AI Researchers Are Really Working On

NeurIPS serves as a critical bellwether for AI research, evolving from an academic gathering into a hub for industry-scale innovation. The rapid transition from laboratory research to production deployment now occurs in weeks, making conference insights essential for business differentiation. Key emerging trends includ...

From Web Video to Real-World Robots

Why Your AI Committee Might Be Your Biggest AI Problem

Building Mathematical Superintelligence

Your First AI Employee Is Already Clocking In

Are Multi-Agent Systems More Complex Than They Need to Be?

Coding Agents Meet Data Science

The podcast explores the use of coding agents in data science, highlighting the challenges and opportunities that arise when applying tools designed for coding to data-centric tasks. It notes that coding agents, while proficient in Python and code decomposition, often lack the skepticism and domain knowledge necessary ...

World Models Are Here—But It’s Still the GPT-2 Phase

Odyssey's CTO Jeff Hawke joins the podcast to discuss world models, a new category of AI models that predict potential futures through continuous streams of interactive pixels. He details Odyssey2 Pro, a frontier model trained on large-scale public video data, distinguishing it from LLMs and generative video models. Ha...

The Hidden Challenges of Running AI at Scale in Production

CoreWeave's SVP of Engineering, Chen Goldberg, discusses the shift from AI experimentation to production across diverse industries. Companies outside Silicon Valley are leveraging their unique data assets to innovate, prompting a need for specialized AI cloud platforms. Goldberg highlights that companies engaged in AI ...

What No One Tells You About Staying Employable in the AI Era

The podcast explores how knowledge workers can remain valuable in the face of increasing AI adoption. It emphasizes adapting work habits to collaborate with AI, rather than resisting it, and shifting focus from routine tasks to problem definition. The discussion highlights the importance of rapid experimentation, orche...

Adaptation: The Missing Layer Between Apps and Foundation Models

The conversation centers on "adaptation" in AI, as an alternative to simply scaling models. Sudip Roy, co-founder and CTO at Adaption Labs, argues that current AI models often fail in the "last 5%" of cases, hindering enterprise adoption. He advocates for gradient-free approaches to lower the unit cost of adaptation, e...

Securing the "YOLO" Era of AI Agents

Jason Martin, Director of Adversarial Research at HiddenLayer, returns to the podcast to discuss OpenClaw, a viral AI agent, and its associated security risks. OpenClaw grants extensive access to a user's system and accounts, making it vulnerable to prompt injection attacks, potentially leading to data exfiltration and...

Building the Open Source Alternative to AWS

The podcast explores the concept of open cloud computing, contrasting it with the closed ecosystems of major hyperscalers like AWS, Azure, and GCP. Umur Choburcu, co-founder of UbiCloud, argues that an open cloud requires an open-source control plane to ensure transparency and portability, allowing users to run cloud s...