33 - On Working With LLM Assistants

33 - On Working With LLM Assistants

In this conversation, Paul and Tim discuss their experiences with AI assistance in coding and its impact on their workflows. They explore the influence of AI assistance on learning and skill development, as well as the challenges and adjustments required in their coding processes. The conversation also delves into the reliability and trust in AI assistance, as well as the concept of authorship and the influence of AI assistance on coding practices. The conversation delves into the themes of code ownership and responsibility, as well as the impact of AI on coding practices. It also explores the concept of the Centaur and the augmentation of human intelligence by AI. Additionally, it discusses the role of experience and expertise in coding, highlighting the challenges and opportunities for new programmers entering the industry. The conversation delves into the application of agile principles in legacy code, emphasizing the importance of incremental development and the concept of walking skeletons. It also explores the resurgence of XP principles and their relevance in modern agile adoption. Additionally, the discussion touches on the cost and viability of AI code assistants, highlighting the need to balance their utility with associated expenses. Takeaways * AI assistance impacts coding workflows * AI assistance influences learning and skill development Code ownership and responsibility * Impact of AI on coding practices Incremental development * Agile principles * Cost of AI code assistants Chapters * 00:00 Authorship and Influence of AI Assistance * 33:28 Code Ownership and Responsibility * 42:34 Experience and Expertise in Coding * 01:08:20 Working with Legacy Code and Agile Principles * 01:16:19 Old Becoming New Again: XP Principles and Agile Adoption * 01:35:45 Cost of AI Code Assistants Find Paul: https://feedbackdriven.dev/ [https://feedbackdriven.dev/] https://www.youtube.com/@FeedbackDrivenDev [https://www.youtube.com/@FeedbackDrivenDev] https://www.linkedin.com/in/paul-hammond-bb5b78251/ [https://www.linkedin.com/in/paul-hammond-bb5b78251/] Find Tim: https://www.linkedin.com/in/agileotter [https://www.linkedin.com/in/agileotter] agileotter.blogspot.com [http://agileotter.blogspot.com] https://leanpub.com/u/tottinge [https://leanpub.com/u/tottinge] https://techhub.social/@tottinge [https://techhub.social/@tottinge] Notes and links: https://www.industriallogic.com/blog/work-to-be-interruptible/ [https://www.industriallogic.com/blog/work-to-be-interruptible/] https://agileotter.blogspot.com/2022/03/splitting-stories-resource-list.html [https://agileotter.blogspot.com/2022/03/splitting-stories-resource-list.html] https://www.geepawhill.org/2021/10/26/mmmss-a-closer-look-at-steps/ [https://www.geepawhill.org/2021/10/26/mmmss-a-closer-look-at-steps/] The Systems Bible - By John Gall. https://www.industriallogic.com/blog/story-points-why-is-this-so-hard/ [https://www.industriallogic.com/blog/story-points-why-is-this-so-hard/]

32 - Hot Tech-e: We are failing too late

In this episode of Hot Tech(e), Philippe Van Eerdenbrugghe and I talk about failure. Mostly about how we keep saying that we should fail early, and then somehow manage to do the opposite. We talk about bugs, TDD, experiments that become permanent by accident, product ideas that are not quite dead enough to remove, and code that was “good enough for now” until “now” became eighteen months. We also talk about the emotional side of failure. Because a failing test does not feel like shame. But a failing feature? A failing design? A failing product bet? That is harder. It was a really nice conversation about failing sooner, learning faster, and maybe not making engineers silently pay for every bad decision forever. FIND PHILIPPE Philippe on LinkedIn: https://www.linkedin.com/in/philippevaneerdenbrugghe [https://www.linkedin.com/in/philippevaneerdenbrugghe] Nodalview: https://www.linkedin.com/company/nodalview/ [https://www.linkedin.com/company/nodalview/] Provide SRL: https://www.provide-srl.com [https://www.provide-srl.com] If you are looking for guidance, Philippe is a great coach for software engineers, tech leads, and CTOs.

Tangent - Creating Crossword Puzzles with Rachel

In this special episode, Rachel shares insights into the world of crossword puzzles, discussing collaboration in puzzle creation, the selection and tightness of themes, and the process of creating a crossword puzzle. Takeaways * Collaboration in crossword puzzle creation * Theme selection and tightness * The process of creating a crossword puzzle Chapters * 00:00 Introduction to Crossword Puzzles * 05:50 Creating a Crossword Puzzle Find all about creating NYT crossword puzzles here:  https://www.nytimes.com/article/submit-crossword-puzzles-the-new-york-times.html [https://www.nytimes.com/article/submit-crossword-puzzles-the-new-york-times.html] Find Rachel at:  https://www.linkedin.com/in/rachel-grey-b69319/ [https://www.linkedin.com/in/rachel-grey-b69319/]  https://bsky.app/profile/greyoctagon.bsky.social [https://bsky.app/profile/greyoctagon.bsky.social]

31 - Hot Tech-e: Estimates are the wrong solution to a problem you don't have - Kevin

The conversation delves into the nature of estimates, the conflation of complexity and time, implicit contracts and expectations, the impact of inaccurate estimates, the usefulness of estimates in business, hierarchy and decision-making, and diverse perspectives on estimates. The conversation delves into the complexities of estimation, planning, and decision-making in software development. It explores the impact of estimates on deadlines, reassurance, and miscommunication, as well as the challenges of standardization and long-term efforts. The speakers discuss the value of small tasks, strategies for handling external deadlines, and the influence of business realities on planning and prioritization. Takeaways * Estimates conflate complexity and time * Implicit contracts and expectations arise from estimates Estimates impact decision-making * Challenges with standardization and long-term efforts Chapters * 00:00 The Nature of Estimates * 08:12 Implicit Contracts and Expectations * 13:48 Usefulness of Estimates in Business * 19:58 Diverse Perspectives on Estimates * 25:02 The Process of High-Level Architecture * 30:23 Challenges with Management Layers and Miscommunication * 35:30 Realistic Considerations for Deadlines and Prioritization * 41:02 Granularity and Handling Small Tasks * 47:08 The Challenge of Standardization and Long-Term Efforts ---------------------------------------- Find Kevin at: * vzb.io [http://vzb.io] * prologuekit.com [http://prologuekit.com] * spott.hr [http://spott.hr]

30 - Hot Tech-e: I can do TDD on the Front End (and AI) - Paul

Paul discusses the possibility and effectiveness of TDD on the front end, emphasizing the importance of testing based on behavior rather than syntax. He shares his experiences with TDD at the BBC and the early days of React, highlighting the philosophy behind front-end testing and the efficiency and confidence it brings. The conversation covers topics related to Test-Driven Development (TDD), AI, mutation testing, and the application of Extreme Programming (XP) and Agile practices in the context of AI. It explores the challenges and benefits of these practices, as well as the limitations and adaptations required when applying them to AI development. The discussion provides valuable insights into the intersection of AI and traditional software development practices, highlighting the potential for improved efficiency and quality in software development processes. Takeaways * TDD on the front end is possible and works brilliantly * Behavior-driven development (BDD) is about testing based on behavior, not just using given-when-then syntax AI and TDD * Mutation Testing * XP and Agile Practices Chapters * 00:00 Efficiency and Confidence in Front-End Testing * 35:21 TDD and Describe Blocks in Testing * 43:50 Mutation Testing * 57:14 XP and Agile Practices ---------------------------------------- Find Paul at: https://feedbackdriven.dev [https://feedbackdriven.dev] and https://www.linkedin.com/in/paul-hammond-bb5b78251/ [https://www.linkedin.com/in/paul-hammond-bb5b78251/] Referenced videos: https://youtu.be/E2ZlufvErk8 [https://youtu.be/E2ZlufvErk8] https://youtu.be/T5wwVK0Xcxo [https://youtu.be/T5wwVK0Xcxo] Pauls Dotfiles: https://github.com/citypaul/.dotfiles [https://github.com/citypaul/.dotfiles] Ian Cooper's TDD talk: https://www.youtube.com/watch?v=EZ05e7EMOLM [https://www.youtube.com/watch?v=EZ05e7EMOLM]