Azure & DevOps Podcast

• Richard Hundhausen and Brad Frazer: Who Owns the Code? - Episode 366

  Brad Frazer is a Boise-based attorney and partner at Hawley Troxell, Idaho’s largest law firm, where he leads the Intellectual Property & Internet practice group with a specific focus in IP Strategy, AI, Data Security, and IT. A graduate of BYU, UC Hastings, and the University of Utah, Brad has decades of experience in cybersecurity, internet law, and IP, including roles as Deputy General Counsel at major tech firms. He’s also a published novelist and a recognized expert in emerging technologies like blockchain, NFTs, and Web3.   Richard Hundhausen helps software organizations and teams deliver better products by understanding and leveraging Azure DevOps and Scrum. He is a Professional Scrum Trainer and author of Professional Scrum with Azure DevOps (MS Press). As a software developer and consultant with over 30 years of experience, he understands that software is built and delivered by people and not by processes or tools.   Topics of Discussion: [2:24] Richard and Brad share their enthusiasm for AI, and they discuss the legal implications of using AI-generated code. [4:00] Brad explains that ownership of AI-generated code is academic until legal issues arise, such as lawsuits or investment rounds. [5:29] Richard explains the process of using AI tools to create code, emphasizing the iterative nature of the process. [8:02] The nuances of copyright law, including the need for human authorship to establish ownership. [10:57] How one gets a registered copyright. [14:19] The different things that AI-driven development can mean. [19:44] Risk avoidance practices as a coder. [23:46] Brad advises software developers to be aware of the legal environment and the potential risks of using AI tools. [24:59] What is an AI output, and what is just the computer being helpful? [32:35] Brad shares a real-world example of a $20 million deal where the seller did not own the code, highlighting the potential risks. [38:38] Brad mentions the Anthropic case, where the company was sued for training its LLM on copyrighted material. [41:22] Richard and Brad discuss the importance of raising awareness and providing resources to help developers understand the legal implications of using AI tools.   Mentioned in this Episode: Clear Measure Way Architect Forum Software Engineer Forum Brad Frazer LinkedIn Richard Hundhausen LinkedIn Thaler vs. Perlmutter Bartz v. Anthropic Who Owns the Code?   Want to Learn More? Visit AzureDevOps.Show for show notes and additional episodes.

  --------  

  45:49

  --------

  45:49
• Tony Cardella: .NET Testing using NCrunch

  Tony Cardella is a seasoned software engineer based in Houston, Texas. With a robust background in enterprise development, Tony brings deep expertise in the .NET Framework (C#), Python, and cloud platforms including Microsoft Azure and Amazon Web Services. His technical repertoire spans both relational databases — such as SQL Server, MySQL, and PostgreSQL — and NoSQL solutions like Azure Cosmos DB.   Tony is a strong advocate for developer productivity tools, frequently leveraging JetBrains products including ReSharper, DataGrip, PyCharm, and Rider, as well as Visual Studio. Outside the world of code, Tony is equally passionate about strength training, whether he’s lifting weights himself or coaching others in the discipline.   Topics of Discussion: [1:34] Tony shares his career journey, starting with a consulting company that reached out to him while he was job hunting. [3:17] NCrunch is an automated testing tool that runs unit tests continuously, focusing on impacted tests. [5:08] Challenges and benefits of NCrunch, and why would we need to use it? [7:44] Tony shares his approach to unit testing, focusing on covering 80% of the code with minimal effort and addressing the remaining 20% as needed. [8:51] The importance of not over-investing in unit tests that may not provide significant value. [11:47] Tony explains how Ncrunch provides code coverage metrics and visual indicators of covered and uncovered code. [12:59] The tool’s ability to show exactly where unit tests are failing, without needing to dive into stack traces. [13:51] Distributed processing and integration tests. [27:44] The challenges of running integration tests with external dependencies, such as databases. [29:18] Exploratory testing and code quality. [32:34] Tony emphasizes the value of unit tests in codifying tribal knowledge and ensuring code quality.   Mentioned in this Episode: Clear Measure Way Architect Forum Software Engineer Forum Tony Cardella Lightning Talks! The Code Gorilla Survey: Fixing Bugs Stealing Time from Development NCrunch   Want to Learn More? Visit AzureDevOps.Show for show notes and additional episodes.

  --------  

  39:24

  --------

  39:24
• Jonathan Peppers: GitHub Copilot for Maui Applications - Episode 364

  Today’s guest is Jon Peppers, Principal Software Engineer on the .NET MAUI team at Microsoft. Before building developer tools, Jonathan was a Xamarin MVP and the lead developer behind various cross-platform Maui apps. With a deep background in C#, from WPF-based self-checkout systems to home automation software featured on Extreme Home Makeover, Jonathan brings a wealth of experience in both app development and the frameworks that power them.   Topics of Discussion: [1:59] Jonathan recounts his first job after college, working on C# for self-checkout software and migrating to WPF. [4:40] How much on the continuum are we right now with Copilot agent mode? [7:11]  The process of setting up Maui development, including installing Visual Studio and the Maui workload. [12:40] Using Copilot for multi-language debugging. [18:42]  Copilot's effectiveness in deleting unnecessary files and finding errors in string localization files. [19:10] Copilot coding agent. [21:20] The process of assigning issues to Copilot, which creates a branch, opens a pull request, and updates the description with its plan. [27:36]  The availability of different models in VS Code, including Claude and GPT, and the anticipation of new models being released. [31:36] The potential for using LLMs on-device for privacy concerns, especially in healthcare. [35:01] Jonathan encourages developers to try Copilot in their IDEs and explore its code completions and suggestions. [35:17] Jonathan’s Cat Swipe dating site!   Mentioned in this Episode: Clear Measure Way Architect Forum Software Engineer Forum How the .NET Maui Team uses GitHub Copilot for Productivity Jonathan on LinkedIn Jonathan Peppers   Want to Learn More? Visit AzureDevOps.Show for show notes and additional episodes.  

  --------  

  36:51

  --------

  36:51
• Micah Martin: Clean Coders - Episode 363

  Micah Martin is a co-founder, CEO, and Studio Director for Clean Coders. He’s a 2nd-generation coder, serial entrepreneur, family man, pilot, airplane builder, engineer, author, and just a normal guy trying to enjoy life to the fullest.   He is also a co-author of Agile Principles, Patterns, and Practices in C#, along with his father, Robert C. Martin, also known as Uncle Bob, and a contributor to the popular testing framework FitNesse, among other things.   Topics of Discussion: [1:52] Micah talks about early influences from his father, Uncle Bob, including childhood “robot” games and learning to code on a Commodore 64 and the first Macintosh. [3:47] First job working alongside industry legends like Kent Beck and Ward Cunningham, and the humbling realization that college hadn’t prepared him to build real software. [5:00] Founding Clean Coders to distribute training videos and later expanding into custom software development with Clean Coders Studio. [6:54] Why apprenticeship is key to developing strong software engineers and how Pete McBreen’s Software Craftsmanship influenced his approach. [8:20] Parallels between martial arts training under a sensei and learning software from a master craftsman. [11:23] How Clean Coders apprentices learn new languages like Clojure through Project Euler challenges and Koans, and why maintaining and extending code is essential training. [15:13] The origins and purpose of FitNesse, acceptance testing, and the need for a modern replacement. [18:43] The gap in tooling for non-programmers to write executable tests, and AI’s potential role in bridging it. [20:35] The role of bullet-point test scenarios in developer/customer collaboration. [21:07] The decline of Agile’s technical focus and the creation of the Software Craftsmanship Manifesto at a summit in Libertyville, IL. [25:29] Carrying forward the Clean Code discipline in both training and client projects. [27:11] Plans to offer a digital apprenticeship experience through CleanCoders.com. [28:17] How Micah uses AI for algorithms, test data generation, and client projects, plus its current limitations. [36:37] Lessons from aviation autopilot systems and why humans remain essential in software development.   Mentioned in this Episode: Clear Measure Way Architect Forum Software Engineer Forum Agile Principals, Patterns, and Practices in C# Clean Coders   Want to Learn More? Visit AzureDevOps.Show for show notes and additional episodes.  

  --------  

  39:16

  --------

  39:16
• Burke Holland: GitHub Copilot Agent - Episode 362

  Burke Holland is a Principal Developer Advocate at Microsoft who leads the Developer Tools Advocacy team. He’s done so much in developer tools, like growing VS Code from 6M to 40M+ users. He’s a frequent speaker at conferences like Microsoft Build, Ignite, GitHub Universe, QCon, and VS Code Day, which he helped create. He’s led many product launches for GitHub Copilot, Codespaces, and more. He has a very popular YouTube channel talking about developer tech.   Topics of Discussion: [3:33] Burke describes his transition from being a good developer to an evangelist, inspired by a Visual Studio evangelist he met. [6:27] The rapid pace of change in the software development world compared to other industries. [9:22] AI-driven development and the various modes available in VS Code, including ask mode, edit mode, and agent mode. [15:41] Burke describes the current moment in AI developer tooling — no one really knows yet what the end product is supposed to be. Everyone agrees AI will help developers, but exactly how is still being figured out. [16:39] What are the right questions to ask AI? [17:41] The importance of providing the AI with the right context to ensure accurate and efficient development. [25:05] AI’s unpredictability makes it difficult to rely on it for consistent development tasks, which is frustrating and foreign to most developers. [32:18] Burke explains that while local AI models can handle small, scoped tasks like generating a function’s contents, they still fall far short of the performance needed for more complex jobs compared to models like GPT-4 or Claude. [37:18] Co-Pilot’s competition. [38:23] Inspiration to people that as long as you are the software architect, you can do anything.   Mentioned in this Episode: Clear Measure Way Architect Forum Software Engineer Forum Programming with Palermo — New Video Podcast! Email us at [email protected]. VS Code Day 2024 Youtube.com/@BurkeHolland/videos Burkeholland.github.io/resume/ Build.microsoft.com/en-US/speakers/0e476452-35ca-4750-ac78-393c0d8c4cb3 Linkedin.com/in/burkeholland/   Want to Learn More? Visit AzureDevOps.Show for show notes and additional episodes.

  --------  

  41:41

  --------

  41:41

Mais podcasts de Notícias

Podcasts em tendência em Notícias

Sobre Azure & DevOps Podcast