As LLMs continue to develop, finding ways to efficiently incorporate their capabilities into applications for many users has become increasingly urgent. I'm delighted today to introduce a course with one of the leading global application developers, Microsoft, which has been developing an SDK called Semantic Kernel that can enable you to rapidly use LLMs in your applications. Your instructor for this course is John Maeda, Vice President for Design and AI at Microsoft. John, welcome. Thank you, Andrew. So glad to be here. John's design work in tech and AI has touched so many lives. For example, his early work at MIT led to his co-creating Stretch, which is a wonderful programming tool for children. Later, John earned an MBA and shifted to the business and venture capital world, which is where we first crossed paths when John was with one of Coursera's investors. So, in this course, I understand you've married two of your favorite topics, teaching programming and accelerating businesses. So, this led to this course on LMs using Semantic Kernel. Well, I know that sounds a bit like a stretch, but I've been fascinated by the enabling power of computation. Computer science is something abstract, but as a kid, I touched a Commodore PET, PET. as a kid growing up in Chinatown of Seattle in a family-run tofu business where parents did just one dream for us, go to college and get out of Chinatown. That encouragement took me to MIT in my adulthood and down a conventional computer science path, but much later in life I discovered that it wasn't the technology that interests me the most, it was instead what you could do with the technology. And oftentimes it required exposing the technology to non technologists, whether children, designers, or artists, or even business people to find out how best to really use it. So I have to say Andrew, I was really inspired by your TED talk on AI. It really struck a chord because you described wanting to bring AI to a business of a pizza shop owner. That's a great vision and it's the kind of things I would have wanted to do for my parents. They're super busy. If AI could have helped them in their business because they've had a better life. I love that thought. So we're going to show how that pizza shop owner can use LLMs in their business. Yeah, thank you, John. I think the work that you and Microsoft are doing to make AI accessible to everyone is really fantastic. Can you say more about what you cover in this course? Absolutely. Well, you know, when you think about Symantec Kernel, it's kind of a fancy name, but just remember it's an open source toolkit. It's the brainchild of Microsoft's deputy CTO. Sam Scalace. Sam is the inventor of something called Google Docs, you may have heard of. There's two underlying concepts to this new wave of AI that I think matter the most. I like to tell people to stick their hands out in front of them. It's hard when you're on a remote, but stick your hands out in front of you, take your right hand, shake it, and call this the completion engine. Whoa, the completion engine. It can finish my sentence. And then take out your left hand, shake it a bit. The left hand is a similarity engine. It's like a magnet, it can sort of pull things out and find things that are similar that are something you wouldn't expect a computer program to find as similar at all. So the ability to compare meaning is very new, at least it feels to me, to some of you as well. And then this completion engine is the kind of wow. And so we sort of focus on the completion engine. oh my god it's so amazing, but the similarity and you need some love. So it's this combination of these two that are really making this wave feel different at least when I talk to businesses of all sizes and I think for small business as well this combination left hand right hand working together it's pretty cool. We see this in places like retrieval augmented generation so-called RAG the ability to connect context with completion but also if you don't have context and you try to complete something, you end up with hallucinations, because the completion engine is running on an empty stomach. So these two together, a combination of two hands, is pretty amazing. And in this course, we're going to try to make sure you don't stay sort of playing with your fingers, but actually coding with Symantec Kernel. You're going to use two AI plugins. Number one, a design thinking plugin, and number two, a business thinking plugin. So in this course, we're going to combine traditional business thinking with the power of LLMs to help make Andrew's pizza shop owner's business life a little better. In the process, you're going to learn about open source semantic kernel. You're going to start by building some semantic functions to summarize some text and do the regular things like chaining. Then we'll jump into design thinking. You're going to take customer feedback and stick it into the plugin and get it to do things and generate the magic of AI. And then we work from top down, we're going to apply SWOT analysis to the pizza shop using the business thinking AI plugin to find ways to improve cost and time efficiency. So moving on, we're going to play with vector memory, of course. Everyone loves that. And use our planner module to finish the meal to get you thinking about the future of this AI revolution. Boy, that's a very exciting set of topics. In addition to John teaching this, we'd also like to acknowledge the many people who've helped make this course possible. On the Microsoft side, we'd like to thank Sam Scalace, Umesh Madan, Devish Lucato, Evan Sharkey, Tim Laverty, Harleen Tind, Abby Harrison, Sean Caligari, Matthew Bolanas, and the entire Semantic Kernel community and team. On the Deep Learning.ai side, Jeff Ludwig, and Diala Ezzedine. So John, this sounds like a really exciting set of topics. Let's get started.
Developing Business Applications using LLMs via Semantic Kernel
- Introduction
- Semantic Kernel is Like Your AI Cooking Kitchen
- Cooking Up Flavorful SWOTs with the Kernel
- Organizing The Tools You Make for Later Reuse
- Frozen Dinner The Design Thinking Meal
- Dont Forget to Save the Generated Dripping or The Gravy
- A Kitchen That Responds to Your I'm Hungry is More Than Feasible
- There's a Fully-Outfitted Professional-Grade Kitchen Ready For You
- Conclusion