Software Delivery Notebook

Tag: AI

  • The changing state of Platform Engineering

    The changing state of Platform Engineering
    The Software Delivery Notebook
    The changing state of Platform Engineering
    Loading
    /

    Drawing on the findings of the most recent two volumes of the State of Platform Engineering report, this episodes looks at the difference between shift-left and shift-down, portal traps, and abstract requirement satisfaction. There’s also a look at AI in platforms and platforms for AI.

    There’s also this idea of platform pluralism. We used to chase the mythical single platform to rule them all, the data says that’s the wrong goal.

    You’ll also hear about the measurement crisis, which has been validated in multiple reports from different research teams.

  • State of AI vs Human Code Generation

    State of AI vs. Human Code Generation
    The Software Delivery Notebook
    State of AI vs Human Code Generation
    Loading
    /

    CodeRabbit looked at 470 open-source pull requests and produced the State of AI vs Human Code Generation report with their findings.

    Crucially, AI makes 1.7x more errors, increasing major and critical defects. The AI-authored/assisted changes cause problems with logic errors, security vulnerabilities, and performance regressions (whereas humans are more likely to make spelling mistakes or have testability issues).

    That means the AI is introducing foundational risk that a competent developer should be catching very early. Is the training data just poisoning the model with old insecure patterns?

  • Engineering performance and AI impact 2025 Report

    Engineering Performance and AI Impact 2025 Report
    The Software Delivery Notebook
    Engineering performance and AI impact 2025 Report
    Loading
    /

    This week we look at the LeadDev and Honeycomb Engineering Performance report 2025, which is based on 400+ survey responses from engineering leaders.

    Find out what they’re measuring performance and how they work out whether AI is having an impact.

  • DORA’s State of AI-Assisted Software Development report

    DORA 2025: State of AI-Assisted Software Development
    The Software Delivery Notebook
    DORA’s State of AI-Assisted Software Development report
    Loading
    /

    DORA has published their latest report and it goes deep into AI-Assisted Software development. It covers the extent of adoption, how it moves the needle on outcomes, and (crucially) what you need in place if you want to succeed.

    As well as thoughtful and thorough analysis on software delivery with AI assistance, the report also looks at different team types and how throughput and stability happen together, not in conflict with each other.

    Burnout, friction, and instability are on the list of things to watch out for, so find out how you can avoid amplifying the bad stuff, and boost the positive outcomes instead.

  • AI coding assistants and perceptions of productivity

    AI coding assistants and perceptions of productivity
    The Software Delivery Notebook
    AI coding assistants and perceptions of productivity
    Loading
    /

    A very deep exploration, conducted by METR with 16 open-source developers and 246 real issues, has looked at perceptions and reality of productivity when using AI coding assistants. Titled Measuring the impact of early-2025 AI on experienced open-source developer productivity, the report tackles something we’ve known for a while, our perception of productivity is no indicator for reality.

    We had the same issue with multi-tasking, where people thought they were more productive, but the reality was they were less productive. So, how does this translate to software delivery with AI assistance? The TL;DR is a perception of a 20% less time to complete tasks, but a reality of an additional 19%. Less than half of AI suggestions were accepted by the developers.

    A lot of earlier studies looked at artificial problems, things that were self contained, maybe didn’t reflect the messiness or real code, or they relied on metrics that, honestly, AI could game.

  • The AI playbook for UK Government

    UK Government AI Playbook
    The Software Delivery Notebook
    The AI playbook for UK Government
    Loading
    /

    A tour of the 10 principles that guide safe, responsible, and effective use of AI in government from the Artificial Intelligence Playbook for the UK Government.

    1. You know what AI is and what its limitations are
    2. You use AI lawfully, ethically and responsibly
    3. You know how to use AI securely
    4. You have meaningful human control at the right stage
    5. You understand how to manage the AI life cycle
    6. You use the right tool for the job
    7. You are open and collaborative
    8. You work with commercial colleagues from the start
    9. You have the skills and expertise needed to implement and use AI
    10. You use these principles alongside your organization’s policies and have the right assurance in place

  • Impact of generative AI in software development

    Impact of Generative AI in Software Development
    The Software Delivery Notebook
    Impact of generative AI in software development
    Loading
    /

    Dives into the new DORA report, titled Impact of Generative AI in Software Development, which looks at the outcomes of using AI for code, docs, and other software delivery tasks.

    The report looks at benefits and problems at the individual and team levels, uncovering some surprises along the way like the vacuum hypothesis and the five key perspectives on AI.

    Here’s another one of those head-scratching moments. Despite all these positive indicators in code and in process the researching surprisingly links AI adoption to negative impacts on overall software delivery performance.

  • AI code quality trends

    AI code quality trends
    The Software Delivery Notebook
    AI code quality trends
    Loading
    /

    Digging into the 2025 AI Copilot Code Quality report from GitClear and Alloy, which looked at 211,000,000 lines of code and made projections for 2025.

    Find out how AI is increasing the speed of change, and the knock on effects of optimizing for short-term speed. Or, more poetically: “Oh, what a tangled web we weave when AI agents we use for speed.”

    Good developers focus on building systems that are not just functional, but also elegant and efficient. They refactor their code, meaning they constantly look for ways to improve the structure and make it more reusable.