Is AI in your (near) future? How would you see it fitting in – at any level of functionality?
It’s a huge push for upcoming releases – at Microsoft Build AI is all over the presentations, the tools, the direction for Microsoft. I wrote briefly about it earlier, talking about SQL Server and the integration of AI and machine learning. It’s on the very near horizon in terms of capabilities, but what about application?
Do you have applications that you’ll be looking to apply the technologies? The learning curve is steep, there is no doubt about that. The thought processes that have to change too are not tiny. We have to figure out not only what types of things can be done (and could be done), but how it gets integrated in, and what it means for our systems.
We’re investigating some rudimentary AI in terms of text-based-speech recognition for support and help bots, and we’re looking at some interesting tools to be better know what people are most interested in when it comes to working with SQL Server and data platforms in general. I have to admit, the times spent thinking through and talking about these types of applications of the technology are quite mind-bending.
You have to really start changing how you think about your information – it takes thinking about how you can make that information available for learning, for review. Typical systems we’d be saying “start with the output in mind” and suggesting you work back from there. Now, we’re looking at entirely different questions. Totally different approaches.
Now, we’re far more likely to be asking about how information will be managed so that it’s usable by any number of projects or toolsets. Not only that, but it would be foolish to think that we know all of the ways it can be used. This means we need to be trying to second-guess what will be possible in the future, and how we can manage information in our systems so that those things will be supported.
It’s a bit like cryogenic freezing of your information. The thought process behind cryogenic freezing is that, at some point in the future, whatever the person was dying of will be cured. At that point, they can be brought back to life and then the thing that was killing them would be cured and they could pick up where they left off. OK, possible over-simplification, but you get the idea.
We’re faced with the same clear direction, but unknown path and destination, when it comes to data and our systems. It would be a huge shame if you were storing everything you needed for that really cool AI process to start making really great observations about your business and the needs of your customers — only to find out you were missing some piece that you could have easily grabbed. This is part of what’s leading to the huge boom in data store sizes. People are starting to store *everything* – in hopes of being prepared when it’s time to thaw out that data for a future capability.
It’s a cool prospect, but man. Sitting with someone, trying to fabricate how this impacts your applications, development and database management, can be a very brain-frying time. Cool, eh?