Today, Google released generative AI functionality in Vertex AI, its machine learning platform. Vertex AI lets developers use Google’s PaLM 2, Imagen, and Codey models to generate and classify text, build ChatGPT-like multi-turn chat experiences, and embed text for semantic search and recommendation engines. The Vertex AI “Model Garden” and the low-code Generative AI Studio will offer these models.
Today’s announcement coincides with the Google Cloud Executive Forum at Mountain View’s Bayview Campus. These skills were previously restricted to trusted testers. Garden debuts Codey, Google’s code completion approach, and PaLM for Chat.
Google partners like GitLab use Codey for their “explain this vulnerability” tool. Canva is leveraging Vertex AI’s translation tool to help non-English speakers and testing the PaLM model to turn short videos into lengthier narratives.
Typeface, which develops creative commercial applications for generative AI, announced a partnership and deep integration with Google’s AI tools today. Typeface, a Google Workspace Marketplace plugin, will use Google’s massive language models. DataStax stated that their AstraDB database service on Google Cloud now supports vector search, allowing AI models to leverage it for long-term memory.
Neo4j is also integrating Google Cloud’s Vertex AI generative AI features. Unstructured data can now be converted into knowledge graphs that users can query using natural language. Vertex AI will enrich these knowledge graphs in real-time, and enterprise customers will appreciate the ability to check huge language model replies against these enterprise knowledge graphs to prevent hallucinations.
Today, Twilio said it tests Vertex-powered features like tailored “best next action” recommendations for contact center agents and automated call summaries.