In this week's podcast, we discuss everything from Google's IO conference and the exciting AI-related updates it brought, to Anthropic's expansion of Claude, and Meta's announcement of generative AI for advertisers. More details below:

• Google's IO conference was full of AI-related update, with the launch of the 540-billion parameter model PaLM 2, which is now being used to fuel a variety of Google tools. We discuss the impressive demos presented, including Google Maps' fly-through city feature and video translation with lip sync. We also look at Google infusing generative AI tools across Gmail, Docs, Sheets, Slides and more.

• Anthropic expands Claude, announcing a significant upgrade to the chatbot, which can now process up to 75,000 words of information, three times more than OpenAI. We delve into the implications this has for the marketing industry and discuss potential use cases when you can provide an LLM tool with this much info and context.

• Stability AI launches text-to-animation application called Stable Animation SDK, a tool designed for artists and developers to create animations through various methods. We share our thoughts on the tool and how it further highlights the move beyond static image generation into the world of video and animation.

• Meta announces generative AI for advertisers, "AI Sandbox," which aims to streamline the process of creating ad creatives for advertisers by generating several versions of ad text and a range of assets for campaigns. We discuss the potential benefits this could bring to marketers, from efficiency and creativity to customization and real-time adaptation.

• Chatbot ads are coming... Microsoft's advertising division has indicated that advertisements delivered within AI chatbot conversations are on the horizon. We share our thoughts on what this could mean for marketers and the potential implications for the landscape of search advertising.

In our brief stories section, we cover a variety of topics, including an interesting article from Bloomberg which reveals that Buzzfeed readers spend 40% more time on AI-generated quizzes than traditional ones, and a leaked memo from Google questioning the ability of large companies like Google and OpenAI to innovate and differentiate their products from open source alternatives.

We also discuss the release of Midjounrey 5.1, which is producing higher quality images, especially when working with shorter prompts compared to previous versions, and a unique app in development that is aiming to train a text-to-bark model so you can communicate with your dog!

Finally, we round off this episode with our tool of the week, Firefly. Tune in to find out more.