The Intersection of Data and Generative AI
Microsoft's Buck Woody breaks down how SQL Server is leveraging gen AI, and what skills you will need to stay competitive.
No pairing has made more sense than the capabilities generative AI can bring to enterprise data. However, with its ability to transform and add to the overall data, keeping a handle of it all may have some IT pros who are thrown in the deep end.
Microsoft has thrown IT a lifeline with AI integration into SQL Server, including its Machine Learning Services feature, which allows for the execution of R and Python scripts directly within the database server, and supports the training and scoring of machine learning models.
But that's not the only AI hook SQL Server is packing. Breaking it all down for us, ahead of his Live! 360 session, titled "Generative AI: Concepts, Tools, and Applications for Data Professionals," is Microsoft's Buck Woody. Woody, an applied data scientist working on the Azure Data Services team, shares some of the new capabilities that have come to SQL Server and breaks down what skills IT needs to have to stay successful when big data meets AI.
And for more great insights, don't forget to attend Woody's upcoming session at Live! 360 in Orlando, Fla. this November. Register now!
Redmond: What sorts of updates or changes has Microsoft made in SQL Server to make it more primed to work with generative AI?
Woody: We've recently published guidelines on using the Machine Learning Services built into SQL Server with Azure OpenAI, In fact, we have an entire Notebook tutorial on how to do that in a practical application. We've also added a new feature to Azure SQL Databases, where you can make REST calls to OpenAI, or actually any AI service that exposes an endpoint.
Between SQL Server and Azure SQL, which one is more compatible with generative AI technologies? Are there advantages and disadvantages with each?
Both work with OpenAI. If you're interested in keeping all your data on-premises, then SQL Server is a good option. If you have the ability to leverage the Cloud, then Azure SQL Database has so many features that interact with that service.
What are some of the coolest use cases you've seen that have been the result of the combination of SQL and generative AI?
Every day a new one comes out. The examples I mentioned above take a simple product name and short description from the product catalog in a set of SQL Server Tables, and then sends that to OpenAI to generate an entire Marketing Campaign for the team to review! By putting this in a Stored Procedure, you could send the entire catalog and get thousands of campaigns in just a few minutes.
There's a lot of talk about how tools like ChatGPT are able to democratize data analysis, machine learning, etc., and make it accessible to people who don't have years of expertise in these subjects. But…this isn't the end of the data scientist, is it? What's your opinion on how traditional data scientists will evolve as these generative AI tools become more widespread?
We take a "Copilot" view -- it's a tool that helps you do your job, but our view is that it won't replace the data professional.
What kinds of skills do you suggest that data pros (or even IT pros in general) brush up on so they can take the fullest advantage of the generative AI technologies that you'll be talking about?
The typical advice is three things: coding, math (heavy on the stats, please) and "domain knowledge," or knowing your business and its goals. I would add to that grammar -- yes, grammar -- so that you can learn how to leverage large language models.