title
15. How Business Thinkers Can Start Building AI Plugin| Andrew Ng | DeepLearning.ai - Full Course

description
The course comes from [https://learn.deeplearning.ai/microsoft-semantic-kernel/lesson/1/introduction](https://learn.deeplearning.ai/microsoft-semantic-kernel/lesson/1/introduction) created by Andrew Ng The content introduces a course on incorporating large language models (LLMs) into applications using Microsoft's SDK called Semantic Kernel. The course, taught by John Maeda, focuses on using LLMs to accelerate businesses, combining traditional business thinking with LLM capabilities. It covers topics like design thinking, SWOT analysis, semantic functions, and AI plugins. The content provides hands-on examples using Semantic Kernel, demonstrating how to create semantic functions, leverage native functions, and organize AI tools for business applications. Get free course notes: https://t.me/NoteForYoutubeCourse

detail
{'title': '15. How Business Thinkers Can Start Building AI Plugin| Andrew Ng | DeepLearning.ai - Full Course', 'heatmap': [], 'summary': "The course introduces llms using microsoft's semantic kernel sdk, open source ai models, coding for business, ai in design thinking, feedback analysis, text analysis services, memory store, vector search, ai planning tools, and multi-step plans with semantic kernel, aiming to make ai accessible and improve revenue growth.", 'chapters': [{'end': 155.393, 'segs': [{'end': 30.2, 'src': 'embed', 'start': 0.129, 'weight': 0, 'content': [{'end': 2.652, 'text': 'As LMs continue to develop,', 'start': 0.129, 'duration': 2.523}, {'end': 10.119, 'text': 'finding ways to efficiently incorporate their capabilities into applications for many users has become increasingly urgent.', 'start': 2.652, 'duration': 7.467}, {'end': 16.845, 'text': "I'm delighted today to introduce a course with one of the leading global application developers, Microsoft,", 'start': 10.139, 'duration': 6.706}, {'end': 23.972, 'text': 'which has been developing an SDK called Semantic Kernel that can enable you to rapidly use LMs in your applications.', 'start': 16.845, 'duration': 7.127}, {'end': 30.2, 'text': 'The instructor for this course is John Maeda, Vice President for Design and AI at Microsoft.', 'start': 24.713, 'duration': 5.487}], 'summary': 'Microsoft has developed an sdk called semantic kernel to rapidly use lms in applications.', 'duration': 30.071, 'max_score': 0.129, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV0129.jpg'}, {'end': 119.744, 'src': 'embed', 'start': 92.259, 'weight': 2, 'content': [{'end': 97.425, 'text': 'That encouragement took me to MIT in my adulthood and then a conventional computer science path.', 'start': 92.259, 'duration': 5.166}, {'end': 102.55, 'text': "But much later in life, I discovered that it wasn't the technology that interests me the most.", 'start': 97.745, 'duration': 4.805}, {'end': 106.153, 'text': 'It was instead what you could do with the technology.', 'start': 103.211, 'duration': 2.942}, {'end': 114.42, 'text': 'And oftentimes it required exposing the technology to non-technologists, whether children, designers or artists, or even business people,', 'start': 106.414, 'duration': 8.006}, {'end': 115.741, 'text': 'to find out how best to really use it.', 'start': 114.42, 'duration': 1.321}, {'end': 119.744, 'text': 'So I have to say, Andrew, I was really inspired by your TED Talk on AI.', 'start': 115.761, 'duration': 3.983}], 'summary': "Inspired by ted talk on ai, emphasizing technology's application beyond just technical aspects.", 'duration': 27.485, 'max_score': 92.259, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV092259.jpg'}], 'start': 0.129, 'title': 'Llms using semantic kernel', 'summary': "Introduces a course on llms using microsoft's semantic kernel sdk, taught by john maeda, vice president for design and ai at microsoft, focusing on making ai accessible to a diverse audience and incorporating lms into applications for various users.", 'chapters': [{'end': 155.393, 'start': 0.129, 'title': 'Llms using semantic kernel', 'summary': "Introduces a course on llms using microsoft's semantic kernel sdk, taught by john maeda, vice president for design and ai at microsoft, focusing on making ai accessible to a diverse audience and incorporating lms into applications for various users.", 'duration': 155.264, 'highlights': ["John Maeda, Vice President for Design and AI at Microsoft, introduces a course on LLMs using Microsoft's Semantic Kernel SDK, emphasizing making AI accessible to a diverse audience and incorporating LMs into applications for various users.", 'Microsoft has been developing an SDK called Semantic Kernel that can enable users to rapidly use LMs in applications.', 'John Maeda discusses his fascination with the enabling power of computation and his goal of making technology accessible to non-technologists, such as children, designers, artists, and business people.']}], 'duration': 155.264, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV0129.jpg', 'highlights': ["John Maeda, Vice President for Design and AI at Microsoft, introduces a course on LLMs using Microsoft's Semantic Kernel SDK, emphasizing making AI accessible to a diverse audience and incorporating LMs into applications for various users.", 'Microsoft has been developing an SDK called Semantic Kernel that can enable users to rapidly use LMs in applications.', 'John Maeda discusses his fascination with the enabling power of computation and his goal of making technology accessible to non-technologists, such as children, designers, artists, and business people.']}, {'end': 1142.961, 'segs': [{'end': 392.607, 'src': 'embed', 'start': 341.821, 'weight': 3, 'content': [{'end': 354.112, 'text': "On the Microsoft side, we'd like to thank Sam Scalace, Umesh Madan, Devesh Lukato, Evan Sharkey, Tim Laverty, Harleen Tind, Abby Harrison,", 'start': 341.821, 'duration': 12.291}, {'end': 359.837, 'text': 'Sean Caligari, Matthew Boulanis and the entire Semantic Kernel community and team.', 'start': 354.112, 'duration': 5.725}, {'end': 364.641, 'text': 'And on the deeplearning.ai side, Jeff Ludwig and Dialla Ezzedine.', 'start': 360.597, 'duration': 4.044}, {'end': 367.924, 'text': 'So John, this sounds like a really exciting set of topics.', 'start': 364.901, 'duration': 3.023}, {'end': 369.145, 'text': "Let's get started.", 'start': 368.384, 'duration': 0.761}, {'end': 372.948, 'text': "So there's a ton of AI content out there.", 'start': 370.265, 'duration': 2.683}, {'end': 378.073, 'text': "You're hoping this is the one that's going to help you figure this stuff out.", 'start': 373.588, 'duration': 4.485}, {'end': 381.997, 'text': "You know it has to be hands-on, otherwise it's just blah, blah, blah.", 'start': 378.753, 'duration': 3.244}, {'end': 383.538, 'text': "That's why we're here.", 'start': 382.858, 'duration': 0.68}, {'end': 392.607, 'text': "Symantec Kernel has been made for large-scale enterprises, but it's also available to you, someone who's trying to figure things out.", 'start': 384.199, 'duration': 8.408}], 'summary': 'Acknowledgments to microsoft and deeplearning.ai teams. discussion on ai content and hands-on approach.', 'duration': 50.786, 'max_score': 341.821, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV0341821.jpg'}, {'end': 480.808, 'src': 'embed', 'start': 448.129, 'weight': 0, 'content': [{'end': 453.692, 'text': 'So the kernel is the essential glue, the orchestrator of different tasks.', 'start': 448.129, 'duration': 5.563}, {'end': 455.113, 'text': 'So super important.', 'start': 454.212, 'duration': 0.901}, {'end': 457.715, 'text': "Secondly, it's semantic.", 'start': 455.733, 'duration': 1.982}, {'end': 459.176, 'text': 'Why is it semantic??', 'start': 458.235, 'duration': 0.941}, {'end': 460.777, 'text': 'Why did we use that word??', 'start': 459.556, 'duration': 1.221}, {'end': 465.359, 'text': "It's because Sam Scalace, the person who got this all started,", 'start': 461.457, 'duration': 3.902}, {'end': 479.667, 'text': 'said this kind of computation that uses AI models and memory models that use language is inherently semantic, not syntactic, not brittle,', 'start': 465.359, 'duration': 14.308}, {'end': 480.808, 'text': 'more flexible.', 'start': 479.667, 'duration': 1.141}], 'summary': 'Kernel is crucial, semantic computation is more flexible.', 'duration': 32.679, 'max_score': 448.129, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV0448129.jpg'}, {'end': 745.34, 'src': 'embed', 'start': 681.714, 'weight': 1, 'content': [{'end': 687.217, 'text': "But isn't it wonderful to use a free model? Okay, got my parens.", 'start': 681.714, 'duration': 5.503}, {'end': 690.519, 'text': 'You made a.', 'start': 689.138, 'duration': 1.381}, {'end': 697.858, 'text': 'Made an open source kernel using an open source AI model.', 'start': 692.236, 'duration': 5.622}, {'end': 701.339, 'text': "Whoa And let's see here.", 'start': 698.278, 'duration': 3.061}, {'end': 704.18, 'text': 'This has to have a few libraries loaded, of course.', 'start': 701.379, 'duration': 2.801}, {'end': 706.761, 'text': 'But I think we have them preloaded.', 'start': 704.3, 'duration': 2.461}, {'end': 709.962, 'text': 'And ta-da.', 'start': 708.142, 'duration': 1.82}, {'end': 717.225, 'text': 'So there you have a kernel that uses OpenAI, or Azure OpenAI, or a Hugging Face completion model.', 'start': 710.242, 'duration': 6.983}, {'end': 722.984, 'text': 'Okay, that is fairly pain-free, hopefully, for you.', 'start': 718.041, 'duration': 4.943}, {'end': 731.01, 'text': "One reminder that this is all open source, so if there's any code you're using that you want to learn more about it,", 'start': 723.365, 'duration': 7.645}, {'end': 733.191, 'text': 'just go to the semantic kernel repo.', 'start': 731.01, 'duration': 2.181}, {'end': 739.035, 'text': "It's on GitHub, where all this kind of cool AI code seems to hang out.", 'start': 733.912, 'duration': 5.123}, {'end': 743.619, 'text': "Let's move on to the next section because you made it through the first section.", 'start': 740.256, 'duration': 3.363}, {'end': 745.34, 'text': "Congratulations Let's go.", 'start': 743.739, 'duration': 1.601}], 'summary': 'Demonstrating the use of open source ai models and libraries in a kernel, emphasizing pain-free experience and open source nature.', 'duration': 63.626, 'max_score': 681.714, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV0681714.jpg'}, {'end': 800.378, 'src': 'embed', 'start': 772.085, 'weight': 4, 'content': [{'end': 782.219, 'text': "We're going to cook up SWOTs, strength, weakness, opportunity threats, boring business stuff, and make it sing on your stovetop with AI.", 'start': 772.085, 'duration': 10.134}, {'end': 783.802, 'text': "Let's see it happen.", 'start': 783.301, 'duration': 0.501}, {'end': 794.596, 'text': 'semantic functions, functions made out of words, so-called prompts, and basically do things that you might not expect are possible,', 'start': 785.413, 'duration': 9.183}, {'end': 800.378, 'text': "which is scale the kind of AI you're trying to build, not just scale the number of people that can use it,", 'start': 794.596, 'duration': 5.782}], 'summary': 'Cook up swots with ai to scale ai capabilities beyond user numbers.', 'duration': 28.293, 'max_score': 772.085, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV0772085.jpg'}, {'end': 990.808, 'src': 'embed', 'start': 967.189, 'weight': 5, 'content': [{'end': 976.141, 'text': 'it stuffed this input into the summary function And then it generated Sparkle, this short summary of that long text.', 'start': 967.189, 'duration': 8.952}, {'end': 977.966, 'text': "You know, long text, nothing's wrong with it, long form is good.", 'start': 976.302, 'duration': 1.664}, {'end': 980.051, 'text': 'Okay, how are you feeling so far?', 'start': 978.908, 'duration': 1.143}, {'end': 981.265, 'text': 'Kind of cool, right?', 'start': 980.745, 'duration': 0.52}, {'end': 985.947, 'text': "Let me just tell you there's another shorthand way of calling these different functions without using the kernel.", 'start': 981.425, 'duration': 4.522}, {'end': 990.808, 'text': "In this case, I'm just going to call the function directly, just kind of like calling a procedure.", 'start': 986.807, 'duration': 4.001}], 'summary': 'Summary function generated sparkle, shorthand way of calling functions.', 'duration': 23.619, 'max_score': 967.189, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV0967189.jpg'}, {'end': 1110.162, 'src': 'embed', 'start': 1079.766, 'weight': 2, 'content': [{'end': 1087.792, 'text': "This is kind of, what's it called again? This is not going to feel awesome, but no, this is a big deal.", 'start': 1079.766, 'duration': 8.026}, {'end': 1092.534, 'text': "You're taking a native function and you're wrapping it in a format that Semantic Kernel can use.", 'start': 1088.592, 'duration': 3.942}, {'end': 1097.836, 'text': 'And the same way that Semantic Kernel can process semantic functions, it can process native functions.', 'start': 1092.814, 'duration': 5.022}, {'end': 1110.162, 'text': 'And next up, what we want to do is we want to run a pipeline of functions, the summary function and the Pig Latin function with the input.', 'start': 1098.237, 'duration': 11.925}], 'summary': 'Wrapping native functions for semantic kernel processing is a big deal.', 'duration': 30.396, 'max_score': 1079.766, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV01079766.jpg'}], 'start': 155.813, 'title': 'Semantic kernel and open source ai models', 'summary': "Introduces semantic kernel, an ai toolkit by microsoft's deputy cto, focusing on completion and similarity engines. it also discusses building open source ai models using openai, azure openai, and hugging face models, and explains the process of creating and running semantic functions in a pipeline for business-style tasks.", 'chapters': [{'end': 589.731, 'start': 155.813, 'title': "Semantic kernel: ai's new wave", 'summary': "Introduces semantic kernel, an open source toolkit created by microsoft's deputy cto, sam scalace, focusing on the new wave of ai powered by a completion engine and a similarity engine, aiming to combine traditional business thinking with the power of ai to improve efficiency and create ai recipes for large-scale enterprises.", 'duration': 433.918, 'highlights': ['Semantic Kernel introduces a new wave of AI powered by a completion engine and a similarity engine, aiming to combine traditional business thinking with the power of AI. The combination of completion engine and similarity engine is emphasized as the key concept powering the new wave of AI.', 'The toolkit aims to create AI recipes for large-scale enterprises and improve efficiency in businesses. Semantic Kernel is designed to aid large-scale enterprises and businesses in improving efficiency through AI-powered solutions.', 'Acknowledgment is given to the individuals who have contributed to the development of the course and the Semantic Kernel community. The chapter acknowledges the individuals from Microsoft and deeplearning.ai who have contributed to the development of the course and the Semantic Kernel community.']}, {'end': 844.242, 'start': 589.731, 'title': 'Building open source ai models', 'summary': 'Explains how to create open source kernels using openai, azure openai, and hugging face models, and discusses the use of semantic functions for business-style tasks.', 'duration': 254.511, 'highlights': ['Open source kernels can be created using OpenAI, Azure OpenAI, and Hugging Face models, providing flexibility and options for developers. The chapter demonstrates the process of creating a kernel using OpenAI, Azure OpenAI, and Hugging Face models, showcasing the flexibility and options available for developers.', 'Semantic functions, including SWOT analysis, can be used to scale AI systems and perform business-style tasks. The use of semantic functions, such as SWOT analysis, is discussed for scaling AI systems and performing business-style tasks, offering a versatile approach for developers.', 'The chapter emphasizes the open source nature of the Semantic Kernel, providing access to code on GitHub for further exploration and learning. The open source nature of the Semantic Kernel is highlighted, with an emphasis on accessing the code on GitHub for further exploration and learning opportunities.']}, {'end': 1142.961, 'start': 844.262, 'title': 'Semantic function pipeline', 'summary': 'Introduces the process of creating and running semantic functions and native functions in a pipeline, including summarization and pig latin conversion, with examples and explanations.', 'duration': 298.699, 'highlights': ['The chapter introduces the process of creating and running semantic functions and native functions in a pipeline It covers the steps involved in creating and executing semantic and native functions in a sequence, demonstrating the process of building a pipeline for these functions.', 'The chapter demonstrates the creation of a summary function for text summarization It explains the process of creating a summary function using a templated prompt with parameters like token count, temperature, and top-p for word range, aiming to summarize text in less than 140 characters.', 'The chapter showcases the usage of a native function for pig latin conversion It illustrates the creation and registration of an exotic language plugin for converting text into pig latin, highlighting the process of wrapping a native function into a format compatible with the Semantic Kernel.']}], 'duration': 987.148, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV0155813.jpg', 'highlights': ['Semantic Kernel introduces a new wave of AI powered by a completion engine and a similarity engine, aiming to combine traditional business thinking with the power of AI.', 'Open source kernels can be created using OpenAI, Azure OpenAI, and Hugging Face models, providing flexibility and options for developers.', 'The chapter introduces the process of creating and running semantic functions and native functions in a pipeline.', 'The toolkit aims to create AI recipes for large-scale enterprises and improve efficiency in businesses.', 'Semantic functions, including SWOT analysis, can be used to scale AI systems and perform business-style tasks.', 'The chapter demonstrates the creation of a summary function for text summarization.', 'Acknowledgment is given to the individuals who have contributed to the development of the course and the Semantic Kernel community.', 'The chapter emphasizes the open source nature of the Semantic Kernel, providing access to code on GitHub for further exploration and learning.']}, {'end': 1569.504, 'segs': [{'end': 1189.704, 'src': 'embed', 'start': 1143.101, 'weight': 0, 'content': [{'end': 1151.468, 'text': "I get lost pretty quickly myself because it's so powerful, but at the end of the day, all you're doing is writing code to solve a problem.", 'start': 1143.101, 'duration': 8.367}, {'end': 1157.874, 'text': 'So the TED talk that Andrew Ng gave about how to help small businesses in particular, a pizza shop owner struck a chord with me.', 'start': 1151.929, 'duration': 5.945}, {'end': 1164.619, 'text': 'And for that reason, I have built this entire lesson for you to not just write code, but think in business terms.', 'start': 1158.274, 'duration': 6.345}, {'end': 1167.362, 'text': "So let's sort of take off by.", 'start': 1165.5, 'duration': 1.862}, {'end': 1170.276, 'text': 'First of all, thinking of the SWOT instrument.', 'start': 1168.715, 'duration': 1.561}, {'end': 1174.597, 'text': 'To create a SWOT, you have to create a list of questions.', 'start': 1171.556, 'duration': 3.041}, {'end': 1179.719, 'text': 'And this is a fairly good list of questions that are composed to figure out, like, the small business owner.', 'start': 1174.998, 'duration': 4.721}, {'end': 1189.704, 'text': 'Like, what are their strengths, weaknesses, opportunities, and threats? And if you organize these in a SWOT chart, the way they look is this.', 'start': 1180.12, 'duration': 9.584}], 'summary': "Andrew ng's ted talk inspired a lesson on coding for small businesses, including the swot analysis for strengths, weaknesses, opportunities, and threats.", 'duration': 46.603, 'max_score': 1143.101, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV01143101.jpg'}, {'end': 1356.748, 'src': 'embed', 'start': 1328.777, 'weight': 2, 'content': [{'end': 1331.598, 'text': 'Let me show you another type of semantic function.', 'start': 1328.777, 'duration': 2.821}, {'end': 1332.798, 'text': 'This is a rewrite.', 'start': 1331.958, 'duration': 0.84}, {'end': 1336.059, 'text': 'Rewrite the text above to be understand by a blank level.', 'start': 1333.038, 'duration': 3.021}, {'end': 1343.413, 'text': "So what we're doing is we're making a new function to be able to change the reading level of whatever comes into it.", 'start': 1336.919, 'duration': 6.494}, {'end': 1350.461, 'text': "And what I'm gonna do is I'm gonna redo the shift domain function, shift the questions into another domain, construction management.", 'start': 1344.214, 'duration': 6.247}, {'end': 1356.748, 'text': "I'm gonna flow that in the pipeline to changing the reading level of whatever comes into it to the level of a child.", 'start': 1351.102, 'duration': 5.646}], 'summary': 'Creating a new function to change reading level, shifting domain to construction management.', 'duration': 27.971, 'max_score': 1328.777, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV01328777.jpg'}, {'end': 1445.347, 'src': 'embed', 'start': 1418.607, 'weight': 4, 'content': [{'end': 1427.054, 'text': "It's like, whoa, all these functions here and there, and this variable here, like how do I organize this? You need plastic containers.", 'start': 1418.607, 'duration': 8.447}, {'end': 1436.922, 'text': "You need ways to organize this AI tool stuff that you're building, plugins, functions, native things, semantic things.", 'start': 1427.374, 'duration': 9.548}, {'end': 1439.283, 'text': 'How do I organize them? No problem.', 'start': 1437.022, 'duration': 2.261}, {'end': 1445.347, 'text': 'We have an organization system for your kitchen to be able to manage your tools.', 'start': 1439.744, 'duration': 5.603}], 'summary': 'Organize ai tools with plastic containers for efficient management.', 'duration': 26.74, 'max_score': 1418.607, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV01418607.jpg'}, {'end': 1497.605, 'src': 'embed', 'start': 1471.219, 'weight': 5, 'content': [{'end': 1477.202, 'text': 'You get both native functions, AI prompts, and it turns out that, you know, large language models are bad at math, generally speaking.', 'start': 1471.219, 'duration': 5.983}, {'end': 1479.403, 'text': 'So you want native functions to do math, woo.', 'start': 1477.482, 'duration': 1.921}, {'end': 1482.604, 'text': 'But if you want to do semantic stuff, where do you go? AI prompts.', 'start': 1480.083, 'duration': 2.521}, {'end': 1488.078, 'text': 'And you can enter things in pure code to build a semantic function,', 'start': 1483.155, 'duration': 4.923}, {'end': 1494.543, 'text': 'or we have an option to package your templated prompt and your configuration information.', 'start': 1488.078, 'duration': 6.465}, {'end': 1497.605, 'text': 'And if you see here, this input is here.', 'start': 1495.683, 'duration': 1.922}], 'summary': 'Native functions for math, ai prompts for semantics. options for entering pure code or templated prompts.', 'duration': 26.386, 'max_score': 1471.219, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV01471219.jpg'}], 'start': 1143.101, 'title': 'Coding and ai for business', 'summary': 'Covers the importance of coding for business success, emphasizing swot analysis and ai solutions such as large language models and semantic kernel to address challenges, aiming to aid small business owners and improve revenue growth and adaptability.', 'chapters': [{'end': 1189.704, 'start': 1143.101, 'title': 'Coding for business success', 'summary': "Emphasizes the importance of coding in solving business problems and highlights the use of swot analysis for small business owners, as demonstrated by andrew ng's ted talk, to help them identify strengths, weaknesses, opportunities, and threats.", 'duration': 46.603, 'highlights': ["The importance of coding for solving business problems, inspired by Andrew Ng's TED talk about helping small businesses, particularly a pizza shop owner.", 'The incorporation of SWOT analysis, with a focus on creating a list of questions to assess the strengths, weaknesses, opportunities, and threats for small business owners.']}, {'end': 1569.504, 'start': 1189.824, 'title': 'Ai solutions for business challenges', 'summary': 'Discusses utilizing large language models and semantic kernel to address business challenges, including transforming swot analysis into different business domains and changing the reading level of texts, to improve ai organization and functionality, aiming to assist businesses in growing revenue and adapting to uncertainties.', 'duration': 379.68, 'highlights': ['The chapter introduces the concept of using semantic functions to transform SWOT analysis into different business domains, such as converting questions geared for a pizza shop into construction management queries, demonstrating the ability to shift the context of the text. ', 'It explores the creation of semantic functions to change the reading level of texts, exemplifying the process of shifting the questions into another domain, such as construction management, and then adjusting the reading level to that of a child, showcasing the flexibility and adaptability of the semantic functions. ', 'The chapter emphasizes the importance of organizing AI tools and functions effectively, mentioning the need for an organization system to manage, improve, and refine the tools, akin to a knife sharpener, to enhance AI development and functionality. ', 'It discusses the usage of native functions and large language models for different AI capabilities, highlighting the role of native functions in mathematical tasks and the utilization of AI prompts and semantic functions for semantic tasks and completion, providing insights into the diverse applications of AI in addressing business challenges. ']}], 'duration': 426.403, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV01143101.jpg', 'highlights': ["The importance of coding for solving business problems, inspired by Andrew Ng's TED talk about helping small businesses, particularly a pizza shop owner.", 'The incorporation of SWOT analysis, with a focus on creating a list of questions to assess the strengths, weaknesses, opportunities, and threats for small business owners.', 'The chapter introduces the concept of using semantic functions to transform SWOT analysis into different business domains, such as converting questions geared for a pizza shop into construction management queries, demonstrating the ability to shift the context of the text.', 'It explores the creation of semantic functions to change the reading level of texts, exemplifying the process of shifting the questions into another domain, such as construction management, and then adjusting the reading level to that of a child, showcasing the flexibility and adaptability of the semantic functions.', 'The chapter emphasizes the importance of organizing AI tools and functions effectively, mentioning the need for an organization system to manage, improve, and refine the tools, akin to a knife sharpener, to enhance AI development and functionality.', 'It discusses the usage of native functions and large language models for different AI capabilities, highlighting the role of native functions in mathematical tasks and the utilization of AI prompts and semantic functions for semantic tasks and completion, providing insights into the diverse applications of AI in addressing business challenges.']}, {'end': 2391.177, 'segs': [{'end': 1829.582, 'src': 'embed', 'start': 1791.783, 'weight': 0, 'content': [{'end': 1793.844, 'text': 'And also, go ahead and change these strengths if you want to.', 'start': 1791.783, 'duration': 2.061}, {'end': 1797.705, 'text': "You'll just see how this is not canned, it's actually generated.", 'start': 1793.864, 'duration': 3.841}, {'end': 1806.409, 'text': "Magic, right? Okay, so what did we do here? We ran the Business Thinking plugin's capability to gain cost efficiencies.", 'start': 1798.165, 'duration': 8.244}, {'end': 1808.149, 'text': 'You tried out the time efficiency, hopefully.', 'start': 1806.489, 'duration': 1.66}, {'end': 1809.19, 'text': "It's a delicious meal.", 'start': 1808.389, 'duration': 0.801}, {'end': 1817.609, 'text': "and now we're going to actually look at another kind of business lens using strategy, because that's how we roll here.", 'start': 1810.123, 'duration': 7.486}, {'end': 1818.91, 'text': "so let's do that.", 'start': 1817.609, 'duration': 1.301}, {'end': 1829.582, 'text': "so we're going to do is we're going to again use the plugin and the plugin is imported from the directory to use the business thinking plugin.", 'start': 1818.91, 'duration': 10.672}], 'summary': 'Business thinking plugin used to gain cost efficiencies and time efficiency, demonstrating its capabilities.', 'duration': 37.799, 'max_score': 1791.783, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV01791783.jpg'}, {'end': 2012.575, 'src': 'embed', 'start': 1987.405, 'weight': 3, 'content': [{'end': 1994.549, 'text': 'and now you can see that you could apply this method to different interviews and generate incredible advice.', 'start': 1987.405, 'duration': 7.144}, {'end': 1999.972, 'text': "that is the power of this new kind of ai and, if you think about it, there's an easier way to do all this.", 'start': 1994.549, 'duration': 5.423}, {'end': 2002.273, 'text': "if you don't want to live in just jupyter notebooks,", 'start': 1999.972, 'duration': 2.301}, {'end': 2009.798, 'text': 'there is a visual studio code extension for semantic kernel that lets you do all of this prompt tuning from your own ide.', 'start': 2002.273, 'duration': 7.525}, {'end': 2012.575, 'text': 'Okay so we did a lot of code stuff.', 'start': 2010.754, 'duration': 1.821}], 'summary': 'New ai method applies to different interviews, offering incredible advice. visual studio code extension allows prompt tuning from own ide.', 'duration': 25.17, 'max_score': 1987.405, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV01987405.jpg'}, {'end': 2082.947, 'src': 'embed', 'start': 2055.982, 'weight': 5, 'content': [{'end': 2062.485, 'text': "So let's reach into the freezer and try out a kind of frozen dinner, a design thinking plugin.", 'start': 2055.982, 'duration': 6.503}, {'end': 2071.049, 'text': "You figure out the SWOT stuff, let's go grab another one, work with it and keep cooking so that you become the best AI chef ever.", 'start': 2062.645, 'duration': 8.404}, {'end': 2071.929, 'text': "Let's do that.", 'start': 2071.489, 'duration': 0.44}, {'end': 2078.045, 'text': "So we're going to tap into the Design Thinking plugin, get excited.", 'start': 2073.322, 'duration': 4.723}, {'end': 2082.947, 'text': 'Design Thinking is the five-step process, empathize, define, ideate, prototype, test.', 'start': 2078.065, 'duration': 4.882}], 'summary': 'Utilize design thinking plugin for 5-step process: empathize, define, ideate, prototype, test.', 'duration': 26.965, 'max_score': 2055.982, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV02055982.jpg'}, {'end': 2134.118, 'src': 'embed', 'start': 2110.644, 'weight': 4, 'content': [{'end': 2119.95, 'text': 'as you take your AI capabilities and make them into transportable plugins packaged so that someone else can use them,', 'start': 2110.644, 'duration': 9.306}, {'end': 2121.791, 'text': "you'll be able to use them in all kinds of places.", 'start': 2119.95, 'duration': 1.841}, {'end': 2124.633, 'text': "If you tried out the chat GPT plugins, that's pretty cool.", 'start': 2121.851, 'duration': 2.782}, {'end': 2134.118, 'text': 'And at Microsoft Build, you heard the EVP of AI, Kevin Scott, talk about, more ways to use AI plugins across the Microsoft universe.', 'start': 2125.234, 'duration': 8.884}], 'summary': 'Transportable ai plugins enable widespread use, as seen in microsoft build with chat gpt plugins and plans for broader ai plugin integration.', 'duration': 23.474, 'max_score': 2110.644, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV02110644.jpg'}], 'start': 1569.524, 'title': 'Ai in business thinking and design', 'summary': "Delves into leveraging ai for cost and time efficiency through business thinking plugins, and explores ai's role in design thinking, emphasizing the process and benefits with detailed examples and capabilities to process swots and gather customer feedback.", 'chapters': [{'end': 1987.405, 'start': 1569.524, 'title': 'Business thinking plugin for cost and time efficiency', 'summary': 'Discusses using a pre-made business thinking plugin to achieve cost and time efficiency, demonstrating the process through detailed examples and emphasizing the benefits of leveraging ai for business strategies.', 'duration': 417.881, 'highlights': ['The Business Thinking plugin is used to gain cost efficiencies, providing specific suggestions such as menu optimization and cross-training staff. The plugin lays out suggestions for cost efficiency, such as menu optimization and cross-training staff, demonstrating its capability to generate specific business improvement recommendations.', "The plugin is also capable of providing strategies for time efficiency, emphasizing the AI's ability to streamline business decision-making processes. The plugin's capacity to offer time efficiency strategies showcases the AI's potential to streamline business decision-making processes, highlighting the benefits of leveraging AI for business strategies.", 'The analysis provided by the plugin is of extremely high quality and adaptable, emphasizing the fluid nature of AI-driven business strategies. The analysis provided by the plugin is of extremely high quality and adaptable, showcasing the fluid nature of AI-driven business strategies and underscoring the potential for tailored business solutions.']}, {'end': 2391.177, 'start': 1987.405, 'title': 'Ai chef: design thinking and plugin development', 'summary': 'Explores the power of ai in generating advice, the integration of design thinking into ai development, and the process of accessing and running the design thinking plugin, highlighting the ability to process swots and gather customer feedback.', 'duration': 403.772, 'highlights': ["The chapter explores the power of AI in generating advice, demonstrating the capability to apply the method to different interviews and generate incredible advice, showcasing the potential of AI in providing valuable insights. It also mentions an easier way to perform tasks, such as using a visual studio code extension for semantic kernel to facilitate prompt tuning from one's own IDE.", 'The integration of design thinking into AI development is emphasized, with the chapter highlighting the significance of knowing different methods to achieve scale. It also presents the concept of becoming an AI chef through the use of design thinking plugins to enhance AI capabilities, enabling the automation of the design thinking process and the creation of transportable plugins.', 'The process of accessing and running the Design Thinking plugin is detailed, showcasing the ability to generate SWOT analysis, process customer feedback, and run empathy analysis using the plugin. It also highlights the steps involved in importing and running the Design Thinking plugin and demonstrates the categorization of different types of customer comments for analysis.']}], 'duration': 821.653, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV01569524.jpg', 'highlights': ['The Business Thinking plugin provides specific suggestions for cost efficiencies, such as menu optimization and cross-training staff.', "The plugin offers strategies for time efficiency, showcasing the AI's potential to streamline business decision-making processes.", 'The analysis provided by the plugin is of extremely high quality and adaptable, underscoring the potential for tailored business solutions.', 'The chapter explores the power of AI in generating advice and providing valuable insights through different interviews.', 'The integration of design thinking into AI development is emphasized, enabling the automation of the design thinking process and the creation of transportable plugins.', 'The process of accessing and running the Design Thinking plugin is detailed, showcasing the ability to generate SWOT analysis, process customer feedback, and run empathy analysis.']}, {'end': 2799.239, 'segs': [{'end': 2418.042, 'src': 'embed', 'start': 2391.177, 'weight': 0, 'content': [{'end': 2401.86, 'text': 'we have a design thinking instant empathize, magical ai generative moment that took in these comments and it generated a sentiment map.', 'start': 2391.177, 'duration': 10.683}, {'end': 2403.86, 'text': 'so now that we have that, you want to remember.', 'start': 2401.86, 'duration': 2}, {'end': 2405.38, 'text': 'what is design thinking?', 'start': 2403.86, 'duration': 1.52}, {'end': 2408.821, 'text': 'and design thinking is a simple set of steps.', 'start': 2405.38, 'duration': 3.441}, {'end': 2409.841, 'text': 'there are five steps.', 'start': 2408.821, 'duration': 1.02}, {'end': 2418.042, 'text': "people argue that there's no steps in design thinking, that It's a set of activities you can sort of do in any order.", 'start': 2409.841, 'duration': 8.201}], 'summary': 'Design thinking uses 5 steps to generate sentiment map from comments.', 'duration': 26.865, 'max_score': 2391.177, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV02391177.jpg'}, {'end': 2580.977, 'src': 'embed', 'start': 2553.777, 'weight': 1, 'content': [{'end': 2558.399, 'text': 'And then you want to prototype it, and you also want to test it with real customers.', 'start': 2553.777, 'duration': 4.622}, {'end': 2559.46, 'text': 'Does this empathize?', 'start': 2558.74, 'duration': 0.72}, {'end': 2563.002, 'text': 'define, ideate, prototype, test five different components of design thinking?', 'start': 2559.46, 'duration': 3.542}, {'end': 2566.464, 'text': 'Well, you can do it with the design thinking plugin as well.', 'start': 2563.462, 'duration': 3.002}, {'end': 2573.872, 'text': "Let's show you it running with four faces all chained together.", 'start': 2567.364, 'duration': 6.508}, {'end': 2580.977, 'text': "We're gonna empathize, we're going to define the problem, we're gonna ideate solutions and we're going to suggest prototyped.", 'start': 2574.072, 'duration': 6.905}], 'summary': 'The design thinking process involves empathizing, defining, ideating, and prototyping, and can be facilitated by the design thinking plugin.', 'duration': 27.2, 'max_score': 2553.777, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV02553777.jpg'}, {'end': 2684.174, 'src': 'embed', 'start': 2658.647, 'weight': 2, 'content': [{'end': 2664.229, 'text': "Now, you might say, hold it, John, you're getting off the hook because you didn't let me test the idea.", 'start': 2658.647, 'duration': 5.582}, {'end': 2681.493, 'text': 'Well, it turns out that more and more services are emerging where you could basically give a large language model a role like a 40-year-old person with a certain situation and suggest how they would respond to this prototype.', 'start': 2665.109, 'duration': 16.384}, {'end': 2684.174, 'text': "Let's do some of that, shall we? So let's make a prompt.", 'start': 2681.513, 'duration': 2.661}], 'summary': "Emerging services allow testing large language model's responses to prompts.", 'duration': 25.527, 'max_score': 2658.647, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV02658647.jpg'}], 'start': 2391.177, 'title': 'Design thinking and ai in feedback analysis', 'summary': 'Explores the application of design thinking and ai in analyzing customer feedback, focusing on empathize and define steps, and emphasizing sentiment map generation and problem identification. it also discusses the use of large language models in testing prototypes and generating ideas.', 'chapters': [{'end': 2531.597, 'start': 2391.177, 'title': 'Design thinking and ai in feedback analysis', 'summary': 'Discusses the use of design thinking and ai to analyze customer feedback, emphasizing the steps of empathize and define, and highlighting the generation of a sentiment map and problem identification.', 'duration': 140.42, 'highlights': ['The use of design thinking and AI to analyze customer feedback, particularly focusing on the steps of empathize and define, with an emphasis on the generation of a sentiment map and problem identification.', 'The discussion of design thinking as a simple set of steps, including the argument about the absence of steps and the use of large language model AI to summarize feedback.', 'The process of chaining together the empathize and define components in design thinking, highlighting the sequential processing and hypothesizing of potential problems based on customer feedback.']}, {'end': 2799.239, 'start': 2531.637, 'title': 'Design thinking and innovation', 'summary': 'Discusses the process of design thinking, including empathizing, defining, ideating, prototyping, and testing, and highlights the use of large language models in testing prototypes and generating ideas.', 'duration': 267.602, 'highlights': ['The chapter discusses the process of design thinking, including empathizing, defining, ideating, prototyping, and testing. The chapter outlines the five components of design thinking: empathize, define, ideate, prototype, and test.', 'The use of large language models in testing prototypes and generating ideas is highlighted. The use of large language models is emphasized in testing prototypes and generating ideas, showcasing how it can simulate reactions to prototypes.', 'The chapter emphasizes the potential of generating ideas and testing prototypes using synthetic persons. The chapter highlights the potential of generating ideas and testing prototypes using synthetic persons, demonstrating the application of large language models in simulating reactions to prototypes.']}], 'duration': 408.062, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV02391177.jpg', 'highlights': ['The use of design thinking and AI to analyze customer feedback, particularly focusing on the steps of empathize and define, with an emphasis on the generation of a sentiment map and problem identification.', 'The chapter discusses the process of design thinking, including empathizing, defining, ideating, prototyping, and testing. The chapter outlines the five components of design thinking: empathize, define, ideate, prototype, and test.', 'The use of large language models in testing prototypes and generating ideas is highlighted. The use of large language models is emphasized in testing prototypes and generating ideas, showcasing how it can simulate reactions to prototypes.']}, {'end': 3118.508, 'segs': [{'end': 2831.697, 'src': 'embed', 'start': 2799.799, 'weight': 1, 'content': [{'end': 2804.321, 'text': 'Or I might be interested because I might want to be someone who takes advantage of being a regular customer.', 'start': 2799.799, 'duration': 4.522}, {'end': 2814.947, 'text': 'So these are really useful responses that you can imagine applying to any of the prototyped ideas and a simple example of how to use this design thinking plugin,', 'start': 2804.341, 'duration': 10.606}, {'end': 2820.449, 'text': 'its different functions and just really helping to test ideas around the small business.', 'start': 2814.947, 'duration': 5.502}, {'end': 2823.331, 'text': 'owner, pizza shop, owner, construction company, you name it.', 'start': 2820.449, 'duration': 2.882}, {'end': 2831.697, 'text': "so let's remember what's happened so far in this lesson, which is uh, i want to keep saying this over and over,", 'start': 2824.091, 'duration': 7.606}], 'summary': 'Design thinking plugin benefits small business owners across industries.', 'duration': 31.898, 'max_score': 2799.799, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV02799799.jpg'}, {'end': 2902.632, 'src': 'embed', 'start': 2870.2, 'weight': 0, 'content': [{'end': 2871, 'text': "It's all over the place.", 'start': 2870.2, 'duration': 0.8}, {'end': 2872.121, 'text': 'You have so much of it.', 'start': 2871.281, 'duration': 0.84}, {'end': 2879.978, 'text': 'What if you could harvest that, reuse it, make it a powerful partner to the completion engine?', 'start': 2873.072, 'duration': 6.906}, {'end': 2886.243, 'text': 'What if you could use the similarity engine that we talked about to be able to play with your gravy?', 'start': 2880.458, 'duration': 5.785}, {'end': 2887.944, 'text': 'Sounds a little bit abstract.', 'start': 2887.084, 'duration': 0.86}, {'end': 2891.968, 'text': "Let's get concrete and let's work with our gravy.", 'start': 2888.425, 'duration': 3.543}, {'end': 2896.932, 'text': "Let's review inventory.", 'start': 2895.731, 'duration': 1.201}, {'end': 2899.354, 'text': 'Design plugin, you did that.', 'start': 2897.472, 'duration': 1.882}, {'end': 2902.632, 'text': 'Did that congrats.', 'start': 2901.209, 'duration': 1.423}], 'summary': 'Harvest and reuse excess for powerful partnership, using similarity engine to work with gravy.', 'duration': 32.432, 'max_score': 2870.2, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV02870200.jpg'}, {'end': 2960.154, 'src': 'embed', 'start': 2922.963, 'weight': 2, 'content': [{'end': 2925.064, 'text': "Don't forget that you've used a lot of tokens to make all the information.", 'start': 2922.963, 'duration': 2.101}, {'end': 2930.426, 'text': "Wouldn't it be nice if we could soak it all up and reuse it somehow? Think about those two dimensions.", 'start': 2925.084, 'duration': 5.342}, {'end': 2932.607, 'text': "Remember that? There's the two engines.", 'start': 2930.466, 'duration': 2.141}, {'end': 2935.148, 'text': 'There is the semantic completion engine.', 'start': 2932.807, 'duration': 2.341}, {'end': 2936.009, 'text': "You've used it a lot.", 'start': 2935.168, 'duration': 0.841}, {'end': 2939.15, 'text': 'And now you want to use the semantic similarity engine.', 'start': 2936.869, 'duration': 2.281}, {'end': 2950.872, 'text': 'And that is slightly different, but so similar that you may not have noticed it as a AI chef that was in training like six months ago.', 'start': 2939.79, 'duration': 11.082}, {'end': 2953.012, 'text': 'And now I can make a pretty good AI meal.', 'start': 2950.972, 'duration': 2.04}, {'end': 2960.154, 'text': "I've been doing this for so long that I realized, wait, there's a lot of completion, but I'm getting a lot of similarity now.", 'start': 2954.253, 'duration': 5.901}], 'summary': 'Transition from semantic completion to semantic similarity engines for ai training and meal preparation.', 'duration': 37.191, 'max_score': 2922.963, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV02922963.jpg'}, {'end': 3118.508, 'src': 'embed', 'start': 3074.817, 'weight': 4, 'content': [{'end': 3083.083, 'text': 'And I attached a memory storage to the kernel so I can hold onto vectors that I generate from the embedding conversion process.', 'start': 3074.817, 'duration': 8.266}, {'end': 3086.226, 'text': 'Just as sort of a caveat, because we want this notebook to be useful to you.', 'start': 3083.603, 'duration': 2.623}, {'end': 3095.655, 'text': "Let's say you run this and your memory store, which is going to be stored in my memories as directory, starts to give you errors.", 'start': 3087.307, 'duration': 8.348}, {'end': 3099.88, 'text': 'So this is a little bit of code that will delete that memory store.', 'start': 3096.196, 'duration': 3.684}, {'end': 3100.861, 'text': "Let's run it.", 'start': 3100.48, 'duration': 0.381}, {'end': 3102.502, 'text': "Let's say delete dir equals true.", 'start': 3100.881, 'duration': 1.621}, {'end': 3107.284, 'text': 'If I run this, it just deleted the memory store folder.', 'start': 3103.243, 'duration': 4.041}, {'end': 3109.845, 'text': 'You might say like, wait, bring it back.', 'start': 3108.685, 'duration': 1.16}, {'end': 3115.667, 'text': "Well, first off, let's just go back here and run this code above and no worries, it's there.", 'start': 3110.305, 'duration': 5.362}, {'end': 3118.508, 'text': "And let's just ignore this code, walk over it, look aside, look aside.", 'start': 3115.787, 'duration': 2.721}], 'summary': 'A memory storage is attached to the kernel to hold vectors; code to delete and restore the memory store is provided.', 'duration': 43.691, 'max_score': 3074.817, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV03074817.jpg'}], 'start': 2799.799, 'title': 'Implementing design thinking and text analysis services', 'summary': 'Discusses implementing a design thinking plugin for small businesses, utilizing completion and semantic similarity engines, and embedding services for text analysis, including the process of attaching a memory store to the kernel for vector storage.', 'chapters': [{'end': 2960.154, 'start': 2799.799, 'title': 'Design thinking plugin for small business', 'summary': 'Discusses the use of a design thinking plugin for testing ideas in small businesses, emphasizing the importance of utilizing the completion engine and semantic similarity engine to generate and reuse information effectively.', 'duration': 160.355, 'highlights': ['The importance of utilizing the completion engine and semantic similarity engine to generate and reuse information effectively, especially in small businesses.', 'Emphasizing the value of collecting and reusing the generated output from the completion engine to enhance the design thinking process.', 'Encouraging the application of the design thinking plugin to various business types such as pizza shops and construction companies for testing and refining ideas.', 'Highlighting the upcoming focus on utilizing the semantic similarity engine to enhance the design thinking process and generate new insights.', "Explaining the concept of 'drippings' from the completion engine and the potential for reusing the information generated by it, illustrating the significance of maximizing resources.", 'Reflecting on the AI chef analogy to convey the evolution of skills and the increasing focus on utilizing the semantic similarity engine in the design thinking process.']}, {'end': 3118.508, 'start': 2960.514, 'title': 'Embedding services for text analysis', 'summary': 'Explains the concept of text embeddings, their role in creating completion and similarity services, and the process of attaching a memory store to the kernel to hold onto vectors generated from the embedding conversion process.', 'duration': 157.994, 'highlights': ['The chapter explains the concept of text embeddings and their role in creating completion and similarity services. It describes how text embeddings can be used to create a completion service and a similarity service, providing two powers: completion power and similarity power.', 'The process of attaching a memory store to the kernel to hold onto vectors generated from the embedding conversion process is detailed. It discusses the steps involved in attaching a memory store to the kernel, such as registering a memory store and persisting the vector database, ensuring that the data generated from the embedding conversion process can be held onto.', "A code snippet for deleting the memory store folder is provided, along with instructions for restoring it if needed. It includes a code snippet for deleting the memory store folder with instructions to restore it if necessary, ensuring the notebook's utility by addressing potential errors in the memory store."]}], 'duration': 318.709, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV02799799.jpg', 'highlights': ['Emphasizing the value of collecting and reusing the generated output from the completion engine to enhance the design thinking process.', 'Encouraging the application of the design thinking plugin to various business types for testing and refining ideas.', "Explaining the concept of 'drippings' from the completion engine and the potential for reusing the information generated by it, illustrating the significance of maximizing resources.", 'Reflecting on the AI chef analogy to convey the evolution of skills and the increasing focus on utilizing the semantic similarity engine in the design thinking process.', 'The process of attaching a memory store to the kernel to hold onto vectors generated from the embedding conversion process is detailed.', 'A code snippet for deleting the memory store folder is provided, along with instructions for restoring it if needed.']}, {'end': 3633.62, 'segs': [{'end': 3197.205, 'src': 'embed', 'start': 3140.483, 'weight': 0, 'content': [{'end': 3141.524, 'text': "They haven't gone anywhere.", 'start': 3140.483, 'duration': 1.041}, {'end': 3150.795, 'text': "OK, so let's now put all of those strengths, weaknesses, opportunities, and threats into memory.", 'start': 3141.985, 'duration': 8.81}, {'end': 3153.017, 'text': "So I'm going to add them to memory collection named SWAT.", 'start': 3151.015, 'duration': 2.002}, {'end': 3161.746, 'text': "I'm going to loop over the different arrays of strings and let's just neatly put them all into, there we go.", 'start': 3153.478, 'duration': 8.268}, {'end': 3163.127, 'text': "So now it's sitting in the vector store.", 'start': 3161.846, 'duration': 1.281}, {'end': 3164.989, 'text': 'Fantastic Okay.', 'start': 3163.548, 'duration': 1.441}, {'end': 3168.112, 'text': "Now what? Let's now use the magnet.", 'start': 3165.49, 'duration': 2.622}, {'end': 3170.935, 'text': 'You know how much I love the magnet symbol? Magnet.', 'start': 3168.332, 'duration': 2.603}, {'end': 3176.854, 'text': "Let's use a magnet.", 'start': 3176.254, 'duration': 0.6}, {'end': 3181.317, 'text': "Okay, so I'm now going to look at this SWAT.", 'start': 3177.415, 'duration': 3.902}, {'end': 3186.419, 'text': "The SWAT's all in vector memory, and I'm now going to ask questions of it.", 'start': 3181.737, 'duration': 4.682}, {'end': 3190.562, 'text': "So what are the easiest ways to make more money is the question I'm gonna ask.", 'start': 3186.92, 'duration': 3.642}, {'end': 3193.583, 'text': "And I'm gonna do the same kind of memory search async.", 'start': 3191.342, 'duration': 2.241}, {'end': 3197.205, 'text': "I'm going to pluck out the different memory results.", 'start': 3194.304, 'duration': 2.901}], 'summary': 'Using memory collection named swat, looped over arrays to store in vector memory, then used magnet symbol to ask questions and search async for different memory results.', 'duration': 56.722, 'max_score': 3140.483, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV03140483.jpg'}, {'end': 3402.122, 'src': 'embed', 'start': 3376.491, 'weight': 3, 'content': [{'end': 3382.881, 'text': "And now that you've done this, we're going to take you into the next chapter, which is all about something at a whole different level.", 'start': 3376.491, 'duration': 6.39}, {'end': 3391.814, 'text': "Once you master plugins and similarity, completion all this world, you suddenly discover that it's time to make a plan, make an AI plan,", 'start': 3383.361, 'duration': 8.453}, {'end': 3394.839, 'text': 'generate a way to solve a goal.', 'start': 3391.814, 'duration': 3.025}, {'end': 3402.122, 'text': 'Instead of just sift through your plugins by hand, have the AI go look through your giant Lego bit of plugins.', 'start': 3395.52, 'duration': 6.602}], 'summary': "After mastering plugins and similarity completion, it's time to make an ai plan for handling plugins efficiently.", 'duration': 25.631, 'max_score': 3376.491, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV03376491.jpg'}, {'end': 3511.45, 'src': 'embed', 'start': 3484.578, 'weight': 2, 'content': [{'end': 3489.86, 'text': 'What do you do? You use retrieval augmented generation to kind of find similar things from your knowledge base.', 'start': 3484.578, 'duration': 5.282}, {'end': 3496.589, 'text': 'And then you might have different things that plugins could do, like write marketing copy or send an email.', 'start': 3490.802, 'duration': 5.787}, {'end': 3499.091, 'text': 'You can make plugins that do all kinds of things, native or semantic.', 'start': 3496.629, 'duration': 2.462}, {'end': 3505.385, 'text': 'And once you do that, what happens is, Each plugin will have a description of some form.', 'start': 3499.872, 'duration': 5.513}, {'end': 3511.45, 'text': 'And all you have to do is use that similarity engine to find the different tools in your tool bucket.', 'start': 3506.065, 'duration': 5.385}], 'summary': 'Implement retrieval augmented generation to find similar items and create various plugins for different tasks.', 'duration': 26.872, 'max_score': 3484.578, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV03484578.jpg'}], 'start': 3119.228, 'title': 'Memory store, vector search, ai plugin, and similarity engine', 'summary': 'Covers techniques for storing data in memory, performing vector searches with a relevant match score of 1, and retrieving specific results. it also discusses using ai plugins and similarity engines to save time and generate professional presentations, ai plans, and complete tasks.', 'chapters': [{'end': 3268.629, 'start': 3119.228, 'title': 'Memory store and vector search', 'summary': "Explains how to store data in memory and perform vector searches, achieving a relevant match score of 1 and retrieving relevant results such as 'catering potential' and 'partnerships' for specific queries.", 'duration': 149.401, 'highlights': ['The chapter demonstrates storing strengths, weaknesses, opportunities, and threats in memory collection named SWAT, achieving a relevant match score of 1 for specific queries.', "The chapter discusses using the magnet symbol for vector searches, retrieving relevant results such as 'catering potential' and 'partnerships' for specific queries.", 'The chapter explains the process of comparing queries to items in the vector store, achieving a relevant match score of 0.8 and 0.6 for specific queries.']}, {'end': 3633.62, 'start': 3269.629, 'title': 'Ai plugin and similarity engine', 'summary': 'Demonstrates the use of ai plugins and similarity engines to save time and generate professional presentations, create ai plans, and complete tasks by leveraging retrieval augmented generation and similarity engines.', 'duration': 363.991, 'highlights': ["Retrieval Augmented Generation The AI plugin uses a similarity engine to find and generate professional presentations, providing solutions to business owners' concerns, showcasing RAG's capability.", 'AI Plan Generation After mastering plugins and similarity engines, the chapter emphasizes the importance of creating AI plans to solve goals and tasks efficiently, reducing the need for manual plugin sifting.', 'Effortless Completion and Solutions The transcript discusses the capability of LLM to complete tasks and generate solutions with minimal effort using plugins and retrieval augmented generation, enabling time-saving and effortless task completion.']}], 'duration': 514.392, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV03119228.jpg', 'highlights': ['The chapter demonstrates storing strengths, weaknesses, opportunities, and threats in memory collection named SWAT, achieving a relevant match score of 1 for specific queries.', "The chapter discusses using the magnet symbol for vector searches, retrieving relevant results such as 'catering potential' and 'partnerships' for specific queries.", "Retrieval Augmented Generation The AI plugin uses a similarity engine to find and generate professional presentations, providing solutions to business owners' concerns, showcasing RAG's capability.", 'AI Plan Generation After mastering plugins and similarity engines, the chapter emphasizes the importance of creating AI plans to solve goals and tasks efficiently, reducing the need for manual plugin sifting.']}, {'end': 4596.554, 'segs': [{'end': 3881.438, 'src': 'embed', 'start': 3854.68, 'weight': 0, 'content': [{'end': 3861.185, 'text': 'But when you can do that, a simple planner, you can imagine a planner that is much more powerful.', 'start': 3854.68, 'duration': 6.505}, {'end': 3867.649, 'text': 'And so the action planner is good for like a very basic searching through, find just one function.', 'start': 3861.865, 'duration': 5.784}, {'end': 3875.113, 'text': "But what if I wanted to do a multi-step plan that's automatically generated? Right? Let's do that.", 'start': 3868.049, 'duration': 7.064}, {'end': 3881.438, 'text': "So what we're going to do is we're going to pull in the sequential planner.", 'start': 3875.893, 'duration': 5.545}], 'summary': 'Introducing a more powerful sequential planner for generating multi-step plans automatically.', 'duration': 26.758, 'max_score': 3854.68, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV03854680.jpg'}, {'end': 4148.587, 'src': 'embed', 'start': 4123.582, 'weight': 1, 'content': [{'end': 4130.269, 'text': "And it's also pulling out the right plugins through the descriptions to be able to execute a plan.", 'start': 4123.582, 'duration': 6.687}, {'end': 4139.036, 'text': "And what does this mean? It means that in the future, you won't go into the kitchen and make dish one, dish two, dish three.", 'start': 4131.43, 'duration': 7.606}, {'end': 4146.104, 'text': "You'll just say I'm hungry and it'll make you maybe five dishes if it has the right ingredients to do that, and it'll make it the ones that you want.", 'start': 4140.037, 'duration': 6.067}, {'end': 4148.587, 'text': "Because it's doing completion.", 'start': 4146.465, 'duration': 2.122}], 'summary': 'Ai pulls out plugins from descriptions to execute plans, enabling it to make multiple dishes based on ingredients and preferences.', 'duration': 25.005, 'max_score': 4123.582, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV04123582.jpg'}, {'end': 4396.431, 'src': 'embed', 'start': 4366.978, 'weight': 2, 'content': [{'end': 4372.164, 'text': 'I can use the chat copilot to understand things like plugins.', 'start': 4366.978, 'duration': 5.186}, {'end': 4374.063, 'text': 'plans and personas.', 'start': 4373.103, 'duration': 0.96}, {'end': 4375.224, 'text': 'Let me walk you through that for a second.', 'start': 4374.243, 'duration': 0.981}, {'end': 4384.487, 'text': "It's super important, because what happens is, if you wanted to build a what's called a chat GPT compatible plugin, you might ask yourself wow,", 'start': 4375.924, 'duration': 8.563}, {'end': 4386.188, 'text': 'I wrote that code, how do I test it?', 'start': 4384.487, 'duration': 1.701}, {'end': 4389.169, 'text': "Well, that's what this chat compiler is there for.", 'start': 4386.648, 'duration': 2.521}, {'end': 4396.431, 'text': 'It lets you basically click on the upper right corner and add your ChatGPT plugin you made in Symantec Kernel.', 'start': 4389.709, 'duration': 6.722}], 'summary': 'Chat copilot helps test and add chatgpt plugins in symantec kernel.', 'duration': 29.453, 'max_score': 4366.978, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV04366978.jpg'}, {'end': 4498.665, 'src': 'embed', 'start': 4466.52, 'weight': 3, 'content': [{'end': 4467.721, 'text': 'You can turn on telemetry.', 'start': 4466.52, 'duration': 1.201}, {'end': 4472.705, 'text': "There's also content safety, meaning content filtering for more AI safety, responsible AI.", 'start': 4468.121, 'duration': 4.584}, {'end': 4476.248, 'text': "There's ways to import different types of documents into your vector stores.", 'start': 4473.345, 'duration': 2.903}, {'end': 4477.129, 'text': "And there's also OCR.", 'start': 4476.288, 'duration': 0.841}, {'end': 4479.151, 'text': 'We have some multimodal experiments running as well.', 'start': 4477.149, 'duration': 2.002}, {'end': 4482.754, 'text': 'So you can also work with images in the system.', 'start': 4479.611, 'duration': 3.143}, {'end': 4489.261, 'text': "Again, our focus has been large language models, but it's been kind of cool that we have different things that just sort of come for free, basically.", 'start': 4482.794, 'duration': 6.467}, {'end': 4490.762, 'text': "And they're there for you.", 'start': 4489.882, 'duration': 0.88}, {'end': 4494.544, 'text': "So if you look at the config file of the backend server, you'll see some of the goodness.", 'start': 4491.543, 'duration': 3.001}, {'end': 4498.665, 'text': 'You can set a default configuration for a different model for completion of anything in Planner.', 'start': 4494.584, 'duration': 4.081}], 'summary': 'Platform offers telemetry, content safety, ocr, multimodal support, and configurable model options.', 'duration': 32.145, 'max_score': 4466.52, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV04466520.jpg'}], 'start': 3633.62, 'title': 'Implementing ai planning tools', 'summary': 'Explores the implementation of an action planner for a computational engine, generating multi-step plans with semantic kernel, and introducing an open-source chat compiler with various backend and frontend chat behaviors, providing insights into future ai capabilities in kitchen automation and multimodal experiments.', 'chapters': [{'end': 3853.58, 'start': 3633.62, 'title': 'Ai action planner for computational engine', 'summary': 'Discusses the implementation of an action planner for a computational engine, adding tools and creating a single function from a vat of plugins to perform tasks like math operations and text manipulation, demonstrating its limited but practical capabilities.', 'duration': 219.96, 'highlights': ['The implementation of an action planner for a computational engine, adding math, file reading, time-telling, and text manipulation tools to a vat of plugins, demonstrating practical but limited capabilities.', 'Demonstrating the capability of the action planner to identify and utilize the most relevant function from the available plugins by comparing function descriptions, such as finding the addition function for summing numbers and the today function for getting the current date.', 'Highlighting the limited but practical nature of the action planner, as it may struggle with complex tasks but excels in simple functions, such as writing text to a file using the file IO skill.']}, {'end': 4266.811, 'start': 3854.68, 'title': 'Generating multi-step plans with semantic kernel', 'summary': 'Demonstrates the use of the sequential planner, showcasing the creation of a multi-step plan to automatically generate a poem and translate it to french, with the system autonomously selecting the required plugins and executing the plan based on similarity and completion, offering insights into the future capabilities of ai in kitchen automation.', 'duration': 412.131, 'highlights': ["The system autonomously selects required plugins and executes the plan based on similarity and completion The chapter demonstrates the system's capability to autonomously select the required plugins for a multi-step plan to create and translate a poem, showcasing the potential for future AI capabilities in kitchen automation.", "Creation of a multi-step plan to automatically generate a poem and translate it to French The demonstration includes the creation of a multi-step plan to autonomously generate a poem and translate it to French, showcasing the sequential planner's ability to handle complex tasks.", 'Introduction of the sequential planner, a gen two planner, and its capabilities compared to the action planner The chapter introduces the sequential planner as a gen two planner that is much more powerful than the action planner, highlighting its enhanced capabilities for creating multi-step plans and executing complex tasks.']}, {'end': 4596.554, 'start': 4266.811, 'title': 'Open-source chat compiler', 'summary': 'Introduces an open-source chat group pilot that offers backend and frontend chat behaviors for free, including a token meter, aiux behaviors, and the ability to create gpt-compatible plugins and plans, while also providing backend server features such as content safety, ocr, and multimodal experiments.', 'duration': 329.743, 'highlights': ['The open-source chat group pilot offers backend and frontend chat behaviors for free, including a token meter and AIUX behaviors. The open-source chat group pilot provides backend and frontend chat behaviors for free, such as a token meter and AIUX behaviors around latency.', 'The chat compiler repository is available to easily create the chat application, with server backend, web API, and frontend web app components. The chat compiler repository is available to easily create the chat application, containing server backend, web API, and frontend web app components.', 'The chat compiler allows the creation of GPT-compatible plugins and plans, enabling users to test and generate plans within the chat context. The chat compiler enables the creation of GPT-compatible plugins and plans, allowing users to test and generate plans within the chat context.', 'The backend server features various functionalities such as content safety, OCR, and multimodal experiments, providing a quick overview of production use cases and support for working with images. The backend server offers functionalities like content safety, OCR, and multimodal experiments, providing a quick overview of production use cases and support for working with images.', 'The completion engine and similarity engine, along with code, enable the creation of unique and unimaginable applications, encouraging users to share their recipes with others. The completion engine and similarity engine, combined with code, enable the creation of unique and unimaginable applications, encouraging users to share their recipes with others.']}], 'duration': 962.934, 'thumbnail': 'https://coursnap.oss-ap-southeast-1.aliyuncs.com/video-capture/aPQBulU8sV0/pics/aPQBulU8sV03633620.jpg', 'highlights': ['The sequential planner is much more powerful than the action planner, highlighting its enhanced capabilities for creating multi-step plans and executing complex tasks.', 'The system autonomously selects required plugins and executes the plan based on similarity and completion, showcasing the potential for future AI capabilities in kitchen automation.', 'The open-source chat compiler enables the creation of GPT-compatible plugins and plans, allowing users to test and generate plans within the chat context.', 'The backend server offers functionalities like content safety, OCR, and multimodal experiments, providing a quick overview of production use cases and support for working with images.']}], 'highlights': ['2 Semantic Kernel introduces a new wave of AI powered by a completion engine and a similarity engine, aiming to combine traditional business thinking with the power of AI.', '2 Open source kernels can be created using OpenAI, Azure OpenAI, and Hugging Face models, providing flexibility and options for developers.', "3 The importance of coding for solving business problems, inspired by Andrew Ng's TED talk about helping small businesses, particularly a pizza shop owner.", '3 The incorporation of SWOT analysis, with a focus on creating a list of questions to assess the strengths, weaknesses, opportunities, and threats for small business owners.', '4 The Business Thinking plugin provides specific suggestions for cost efficiencies, such as menu optimization and cross-training staff.', '5 The use of design thinking and AI to analyze customer feedback, particularly focusing on the steps of empathize and define, with an emphasis on the generation of a sentiment map and problem identification.', '6 Emphasizing the value of collecting and reusing the generated output from the completion engine to enhance the design thinking process.', '7 AI Plan Generation After mastering plugins and similarity engines, the chapter emphasizes the importance of creating AI plans to solve goals and tasks efficiently, reducing the need for manual plugin sifting.', '8 The sequential planner is much more powerful than the action planner, highlighting its enhanced capabilities for creating multi-step plans and executing complex tasks.']}