OpenAI Cracks Erdős Conjecture, Google Gemini Omni, NVIDIA Vera CPU, Karpathy to Anthropic

OpenAI's Rosalind and Dreaming V3, Google's Gemma 4, and Siri's Gemini Twist

In this episode, we discuss OpenAI's upgraded GPT Rosalind model for life sciences research, the new Dreaming V3 memory architecture rolling out to ChatGPT users, and Google DeepMind's release of Gemma 4 12B, an open-weight multimodal model built to run locally on laptops. We also dig into a striking report that Apple's revamped Siri will reportedly lean on Google Gemini running on NVIDIA Blackwell hardware, raising big questions about Apple's assistant strategy and the future of full-stack AI. From frontier lab releases to on-device multimodal AI and shifting alliances between OpenAI, Google, NVIDIA, and Apple, we break down what each move means for developers, researchers, and everyday users. https://www.aiconvocast.com Help support the podcast by using our affiliate links: Eleven Labs: https://try.elevenlabs.io/ibl30sgkibkv Disclaimer: This podcast is an independent production and is not affiliated with, endorsed by, or sponsored by OpenAI, Google, DeepMind, Apple, NVIDIA, or any other entities mentioned unless explicitly mentioned. The content provided is for educational and entertainment purposes only and does not constitute professional, financial, or legal advice. Affiliate links may earn the podcast a commission at no additional cost to you.

Meta Business Agent Goes Global, OpenAI Codex Pivot, Instagram AI Hack

In this episode, we cover Meta's global launch of its Business Agent across WhatsApp, Messenger, and Instagram, and OpenAI's pivot of Codex from a coding tool into a general productivity platform now serving millions of weekly users. We also break down a troubling security incident where Instagram's AI support chatbot was reportedly tricked into helping attackers take over accounts, and the UK's landmark ruling forcing Google to let publishers opt out of AI Overviews and AI Mode while remaining in regular Search. From agentic commerce and knowledge work to AI safety failures and publisher leverage against Google, we unpack what each shift means for users, developers, and the broader AI ecosystem. https://www.aiconvocast.com Help support the podcast by using our affiliate links: Eleven Labs: https://try.elevenlabs.io/ibl30sgkibkv Disclaimer: This podcast is an independent production and is not affiliated with, endorsed by, or sponsored by Meta, OpenAI, Instagram, WhatsApp, Google, Anthropic, Shopify, Zendesk, or any other entities mentioned unless explicitly stated. The content provided is for educational and entertainment purposes only and does not constitute professional, financial, security, or legal advice. Affiliate links may earn the podcast a commission at no additional cost to you.

Microsoft Scout and MAI Models, OpenAI Codex Agents, GPT-5 on AWS Bedrock

In this episode, we cover Microsoft's Build 2026 launch of Scout and its new MAI family of in-house models, including the MAI-Thinking-1 reasoning model trained with "zero distillation." We also break down OpenAI turning Codex into a workplace agent with role-specific plugins for analytics, sales, design, and finance, plus the general availability of GPT-5.5, GPT-5.4, and Codex on Amazon Bedrock. Finally, we unpack President Trump's narrower frontier AI executive order, which sets up a voluntary cyber capability testing framework for labs like OpenAI, Anthropic, Google, Microsoft, and xAI. Tune in for a clear look at how Microsoft, OpenAI, AWS, and Washington are reshaping the frontier AI landscape. https://www.aiconvocast.com Help support the podcast by using our affiliate links: Eleven Labs: https://try.elevenlabs.io/ibl30sgkibkv Disclaimer: This podcast is an independent production and is not affiliated with, endorsed by, or sponsored by Microsoft, OpenAI, Amazon, AWS, Anthropic, Google, xAI, or any other entities mentioned unless explicitly mentioned. The content provided is for educational and entertainment purposes only and does not constitute professional, financial, or legal advice. Affiliate links may earn us a commission at no additional cost to you.

NVIDIA Cosmos 3, Microsoft Surface Laptop Ultra, MiniMax M3, and Anthropic Agent Risks

In this episode, we cover NVIDIA's launch of Cosmos 3, an open foundation model for physical AI spanning robotics, autonomous vehicles, and simulation. We break down the NVIDIA and Microsoft partnership behind RTX Spark and the Surface Laptop Ultra, which can run 120 billion parameter models locally and reframes the PC as an agent host. We also dig into MiniMax's surprise M3 release, an open-weights coding model undercutting Claude, Codex, and Gemini on price, and Anthropic's eye-opening browser agent prompt injection disclosures that expose how little transparency other frontier labs offer on agent security. https://www.aiconvocast.com Help support the podcast by using our affiliate links: Eleven Labs: https://try.elevenlabs.io/ibl30sgkibkv Disclaimer: This podcast is an independent production and is not affiliated with, endorsed by, or sponsored by NVIDIA, Microsoft, MiniMax, Anthropic, OpenAI, Google, Meta, or any other entities mentioned unless explicitly stated. The content provided is for educational and entertainment purposes only and does not constitute professional, financial, or legal advice. Affiliate links may earn the podcast a commission at no additional cost to you.

OpenAI Codex on Windows, GitHub Copilot Token Backlash, Nvidia Windows AI PCs

In this episode, we discuss OpenAI bringing Codex Computer Use to Windows, letting the agent see, click, and type through Windows apps while developers steer it remotely from ChatGPT on mobile or Codex on Mac. We break down the developer backlash to GitHub Copilot's shift to token and AI-credit-based billing, the new Pro, Pro Plus, and Max tiers, and why agentic workflows are breaking the flat-subscription model across OpenAI, Anthropic, and xAI. We also cover Nvidia and Microsoft teasing the first Nvidia-powered Windows AI PCs ahead of Computex and Microsoft Build, with rumored N1X-style Windows-on-Arm systems that could reset the AI PC conversation. Finally, we examine the U.S. Commerce Department's move to close the offshore loophole on advanced Nvidia Blackwell, Rubin, and AMD MI350x AI chip shipments to Chinese firms operating outside mainland China. https://www.aiconvocast.com Help support the podcast by using our affiliate links: Eleven Labs: https://try.elevenlabs.io/ibl30sgkibkv Disclaimer: This podcast is an independent production and is not affiliated with, endorsed by, or sponsored by OpenAI, GitHub, Microsoft, Nvidia, AMD, Anthropic, xAI, or any other entities mentioned unless explicitly stated. The content provided is for educational and entertainment purposes only and does not constitute professional, financial, or legal advice. Affiliate links may earn the podcast a commission at no additional cost to you.