Vertex AI from Google Announces Massive Upgrades

New Features

Google has unveiled significant updates to Vertex AI, its platform for developing, training, and deploying machine learning models. These updates include improved AI models for text, image, and code generation, with a notable 25% quality improvement in the Kodi code generator model. Additionally, the Imgen model now supports style tuning, allowing for the creation of brand-aligned images with minimal reference pictures 1.

   

Model Improvements

The platform's Palm 2 language model has expanded its language understanding capabilities, now covering 38 languages in general availability and over 100 in preview. It also features a 32,000-token context window, enabling it to process up to 25,000 words before generating additional text. This balance between new capabilities and competitive price-to-performance ratios is crucial for enterprises 1.

The choice of a 32,000 token Windows kind of strikes a balance between new capabilities and competitive price to performance ratios.

--- Jaeden Schafer

These improvements are part of Google's broader strategy to integrate AI more deeply into business functions 2.

   

Integration

Vertex AI is also enhancing its integration capabilities by incorporating third-party models like Anthropic's Claude 2 into its model garden. This move aims to offer a collection of pre-built models that can be customized to meet enterprise needs, directly challenging platforms like Amazon's Bedrock. Google is also introducing extensions and data connectors to Vertex AI, enabling developers to connect models to real-time data and third-party apps 2.

This move is seen as a direct challenge to Amazon, Bedrock AWS's equivalent platform.

--- Jaeden Schafer

These advancements reflect the industry's trend towards deeper AI integration in various business functions 2.