Ollion’s Notes from Snowflake Summit Part 1: Latest Innovations
The biggest topic of the conference this year was undoubtedly AI. From significant improvements to general availability of numerous services, the advancements in AI were the highlight. Here’s a breakdown of AI as well as the other the main announcements from the keynote:
AI Developments
Various enhancements and general availability announcements.
Snowflake Arctic
This suite of models features a collection of embeddings optimized for retrieval applications and a versatile large language model (LLM) designed for advanced enterprise tasks.
- Focuses on Enterprise Intelligence and efficiency.
- Not a huge model but highly efficient (will be a great default LLM imho)
Cortex AI
Cortex offers a range of AI capabilities powered by large language models (LLMs). These features are designed to interpret unstructured data, respond to open-ended questions, and deliver intelligent support.
- Cortex Studio: A wizard-like interface for tasks like creating custom LLMs with fine-tuning, usable in downstream apps.
- Cortex Analyst: Enables the creation of chatbots based on your data, allowing chatbot interactions with text data.
- Cortex Search: Facilitates the creation of a search function based on a folder full of documents, simplifying the process.
- There is more too. Cortex is poping up all over the platform and just about every partner tool!
Snowflake Notebooks
Notebooks, available in Snowsight, provide an interactive, cell-based programming environment for developing with Python and SQL.
- New AI notebooks now exist, integrating seamlessly with Git for AI development directly in Snowflake
- They allow scheduling, selecting runtime environments with a simple click, either on a warehouse using Snowflake Anaconda or a container.
- Bonus announcement: Integration with Pandas called Snowpark Pandas API. Doing data prep using python just got easier.
Snowflake Co-Pilot
Copilot is an AI-driven assistant designed to help you write, optimize, and interpret SQL queries
- Inline code generation based on prompts is now a feature.
- Custom instructions can be added, enhancing usability.
Document AI in GA
Document AI is a tool capable of processing various document formats and extracting information using natural language understanding
- Parse out handwritten forms, ask questions in natural language, and validate the answers
Data Management and Streaming
Iceberg and and Polaris
Iceberg Tables are launched in conjunction with the newly announced Polaris Catalog, an open, vendor-neutral catalog implementation for Apache Iceberg.
- Open Source Tables are gaining popularity for raw data storage, with support for Iceberg and Polaris API for cataloging, allowing two-way sync.
- I thought it was great that Snowflake no longer expects data to be landed in the proprietary FDN format. Now, data can be landed in Iceberg (or maybe another another format in the future that will remain unnamed for now unless you have $2B), allowing multiple compute engines to interact, with a clear focus on leveraging Snowflake’s platform.
Dynamic Tables
Dynamic tables in Snowflake automate data transformation, generating tables from query results to simplify data engineering efficiently and cost-effectively
- “Managed Streams and Tasks”
- I will write a follow up blog detailing an interesting use case I saw using dynamic tables
Governance and Compliance
Snowflake Horizon
A suite for managing privacy, security, compliance, and interoperability.
- Also included automatic tagging and classification, along with automated tag propagation, aiding in data lineage and compliance.
Trust Center
Enables Snowflake customers to effectively monitor, identify, and maintain strong security, governance, privacy, and compliance, while simplifying risk management
- Improved role management, focusing on better privacy and security practices.
Internal Improvements
Cost Optimization
- A new management interface in the Admin tab provides recommendations for cost optimization, underscoring its importance.
- A million other “little” things from better execution times to faster loading time.
Applications
Container Support
Snowpark Container Services is a fully managed container solution that simplifies deploying, managing, and scaling containerized services, jobs, and functions.
- General availability on AWS (not yet in Microsoft or GCP)
- Native apps becoming more prominent.
Monitoring Tools
Snowflake Trail
A set of observability features that enable developers to monitor, troubleshoot, and optimize workflows
- Adds metrics, logging, and trace functionalities to databases.
- Compatible with third-party tools like Datadog, Grafana, Slack, etc.
Visual Enhancements
Dark Mode
- A small but welcome addition, providing a more user-friendly interface option.
This keynote highlighted Snowflake’s commitment to AI, governance, cost optimization, and user experience enhancements, positioning it as a robust platform for modern data needs. The introduction of Cortex AI tools, dynamic streaming tables, and advanced governance features demonstrate Snowflake’s ongoing innovation in the data cloud space.
Check out Part 2 of the Summit Series here!
About Ollion
At Ollion, we have been a proud Snowflake Service Partner for almost a decade. Our mission is to connect companies and capabilities worldwide, helping ambitious organizations achieve game-changing breakthroughs without losing sight of the people impacted. We offer a unique point of view as an independent, straightforward partner backed by a global team of client partners, sales, engineering, delivery, and more.
Let me know if you attended and want to talk more about Snowflake Summit 2024!