The Cloud Pod | Weekly AI & Cloud News on AWS, Azure & GCP · Justin Brodley, Jonathan Baker, Ryan Lucas and Matt Kohn | Cloud Computing & AI News

333: The Cloud Pod Goes Nano Banana

·1 hr 3 min·4 clips
Sam Altman issued an internal code red memo after Gemini gained 200 million users in three months.
1. The Cloud Pod episode 333, “The Cloud Pod Goes Nano Banana,” centers on Google, Anthropic, OpenAI, and AWS-related cloud news from December 2nd, 2025. 2. Hosts Justin, Jonathan, Ryan, and Matthew lead the show, and Ryan and Matt handle most of the back-and-forth while the episode title is framed around Google’s Nano Banana announcement. 3. The episode is for listeners who want a weekly cloud-news roundup that moves from AI model releases to networking, infrastructure, and platform pricing. 4. The hosts start with a post-Thanksgiving check-in and say the week has felt “weird,” “blah,” and hard to get into gear. 5. They note that re:Invent is happening “underneath us right now,” and they postpone AWS coverage until next week. 6. The Nano Banana segment identifies Google’s “Nano Banana Pro” as Gemini 3 Pro image editing in Vertex AI and Google Workspace. 7. They say the model supports up to 14 reference images, 4K resolution outputs, multilingual text, and Google Search grounding for factual accuracy. 8. The hosts also mention SynthID watermarking, copyright identification at general availability, and enterprise integrations with Adobe Firefly, Photoshop, Canva, and Sigma. 9. They add that Klarna, Shopify, and Wayfair are already using the model for product visualization and marketing asset generation. 10. Anthropic’s Claude Opus 4.5 segment gives the model’s GA pricing as $5 per million input tokens and $25 per million output tokens. 11. The hosts say the model identifier is “claude-opus-4-5-20251101,” and they call out that the name is hard to say. 12. They describe a new effort parameter in the API that lets developers choose between speed and capability. 13. They note testing claims that Opus 4.5 matched Sonnet 4.5’s best score with 76 fewer output tokens at medium effort and exceeded it by 4.3 points at higher effort with 48 fewer tokens. 14. The episode highlights Cloud Code plan mode, which creates editable plan.md files before execution and now appears in the desktop app for parallel sessions. 15. The hosts also mention that Claude for Chrome and Claude for Excel are expanding to all Max, Team, and Enterprise users. 16. OpenAI’s “code red” story says Sam Altman issued an internal memo after Gemini 3 reached the top of the LMArena leaderboard and Gemini gained 200 million users in three months. 17. The show’s tone is conversational and informal, with jokes about “stupid rocks floating in space,” bill shock from token usage, and turning down the volume on the microphone. 18. The format is a fast-moving panel discussion that mixes product announcements, personal usage stories, and side comments about cloud bills and tooling choices. 19. Listeners who track AI model releases and cloud platform changes will get the most from this episode. 20. Listeners who want a tightly focused AWS keynote recap may skip it.
Listen to the show on