AI-podden

In this special two-part Dreamforce episode, our host Ather Gattami sits down with Itai Asseo, VP at Salesforce AI Research, and Dr. Franny Hsiao, Chief AI Architect at Salesforce, to explore how Salesforce is approaching the next phase of enterprise AI. Across the two conversations, they cover Koa, Salesforce's new CRM-specific reasoning model, the role of customer data and privacy, and why businesses should choose the right-sized model for each task rather than defaulting to the largest frontier models. They also discuss dynamic AI interfaces, self-evolving enterprises, and why successful AI adoption depends as much on people, organisational change and continuous learning as it does on the technology itself.

In this August AI news roundup, our host Ather sits with Anders Arpteg to unpack the latest developments in frontier models, physical AI, autonomous agents and the growing race for AI infrastructure. The conversation covers breakthroughs in robotics and latent-space reasoning, fierce model and pricing competition from US and Chinese players, major changes at Google, and the escalating battle for chips and compute. They also explore the limits of current AI reasoning, emerging security risks as agents become more autonomous, and why efficiency, reliability and infrastructure are becoming just as important as raw model performance.

In this July AI news roundup, our host Ather sits with Anders Arpteg to unpack the latest frontier model releases, the rapid rise of open-source competitors, and the growing importance of enterprise agent platforms. The conversation explores how innovation is shifting from raw model performance towards infrastructure, governance, efficiency and latent-space reasoning, while Chinese models, energy constraints and increasingly capable autonomous agents reshape the competitive landscape.

In this June news episode, host Ather Gattami sits down with Anders Arpteg to unpack the biggest AI developments from the past month. The discussion looks beyond frontier model hype to the infrastructure, platforms and specialised models that are starting to define real-world adoption. From GPT 5.6, Claude and China's fast-rising open models to custom chips, enterprise agents, geopolitical pressure, continual learning and AI-enabled cyberattacks, the episode explores where advantage in AI is really shifting: from model weights alone to data, orchestration, efficiency and domain expertise.

In this May news episode, our host Ather Gattami sits with Anders Arpteg to discuss how the industry accelerating on several fronts at once: coding, agentic workflows, enterprise tooling and research capability. As Anthropic, Google, Microsoft, Mistral and others push new models and platforms, the competitive edge is shifting from raw model performance toward speed, integration, transparency and data sovereignty. At the same time, breakthroughs in mathematical proof, AI-assisted research and formal reasoning suggest the technology is beginning to move beyond pattern recognition toward genuinely novel insight and more autonomous problem-solving.