Multizone | All our technotes
If we've published it, you can find it here.
The AWS Diagram-as-code tool provides a practical way to manage AWS diagrams. With each update, it has become easier to model complex architectures such as Virtual Private Clouds (VPCs), PrivateLink connections, and service integrations. A recent workflow for a client demonstrates how to use Gemini CLI to represent a VPC connected to AWS Licence Manager via AWS PrivateLink.
TL:DR – The AWS Diagram-as-code tool is rather picky about the structure it uses for its YAML configuration file. Using Gemini CLI allows you to experiment with diagrams and iterate using english language prompts until you get what you need. A serious timesaver.
- Details
My mum is 86 years old. She has an old 2014 MacBook Air and an iPad Mini 6th Generation. She doesn't use the MacBook so much anymore but she loves her iPad. It is her third now. She uses them pretty hard! Almost 24/7. Her faithful iPad mini (6th generation) (Wi-Fi + Cellular) started to behave badly last week when it failed to take charge anymore. I took it home and verified there was a problem with it. Fortunately I was able to back it up. I thought about binning it and buying another one, but decided to take it to the Apple Store to see what could be done.
TL:DR – Apple replaced it with a new same model iPad for the charge for a battery repair. Only Apple behave this way, at retail outlets worldwide. They stand by their products and help customers if at all possible.
- Details
With the unveiling of Liquid Glass at WWDC 2025, Apple has undoubtedly dominated the conversation surrounding its forthcoming operating systems. But before you dive headfirst into the reality distortion field created by the fluidity of translucent widgets, it’s essential to scrutinise the broader landscape of innovation across Apple's platforms. Liquid Glass isn’t the sole revolution taking place; rather, it is a most visible tip of the iceberg of a multitude of innovations meant to redefine user experience. Lets explores not just the impact of Liquid Glass, but how its introduction is a signpost for other exciting developments across Apple’s ecosystem for the next decade and what all of this means for users, developers, and competitors alike.
TL:DR – Liquid Glass is a significant update in Apple's design ethos, but alongside it, the company is propelling forward with advancements in AI, infrastructure, and developer tools that will have a profound impact on software development, design frameworks, and user experiences. More innovations are on the horizon and many, like Liquid Glass will be shared across different Apple's different platforms, reshaping how we interact with their technology. For those not in software development, it may be wise to keep your device on the current release until the upgrade is released out of beta. Buckle up, it’s about to get interesting.
- Details
Read more: Beyond the glimmer of Liquid Glass is a sparkle concealing a revolution
In product management, where the simplistic notion of developing "great software" has devolved into a relentless race for market share, the idea of using Artificial Intelligence as a tool to enhance productivity might cause a raised British eyebrow given the hype in the market. Yet, the increasingly complex demands of stakeholders and the relentless urgency of customer expectations make it imperative for product managers to embrace AI if they wish to not only survive but thrive.
TL:DR – Artificial Intelligence (AI) can provide product managers with tools to improve productivity through better decision-making, efficient prioritisation, and enhanced collaboration. By leveraging AI-driven frameworks and tools, product managers can streamline workflows, reduce the cost of the product creation and ideation process, ensure product alignment with customer needs, and ultimately deliver value more efficiently. Ignoring AI in the product creation process is, frankly, a dereliction of duty. The choice then is straightforward: embrace AI-driven solutions for product creation and management or risk becoming an anachronism in your own field.
- Details
Read more: Improving product management productivity with AI
Ah, Microsoft Lists. The latest noble attempt to create a flexible, user-friendly tool for collating information and managing tasks. Yet, despite years of iteration and a raft of AI-powered promises, these Lists continue to morph into a tangle of SharePointy confusion and frustration. As organisations deepen their Microsoft 365 commitments in 2026, the question remains stubbornly relevant: does the brittle nature of Microsoft Lists inadvertently stifle adoption? The answer, regrettably, reveals much more about Microsoft's approach to software than its ardent advocates would care to admit.
TL;DR – Microsoft Lists hold genuine potential for improving productivity, yet their fragility continues to hamper user adoption. A catalogue of issues — from conditional formatting that shatters on contact to an ecosystem now lurching through forced AI integration — contributes to a poor user experience. The mobile apps are gone entirely as of late 2025. Copilot is now the front door. Whether that makes things better or just differently broken remains an open question.
{ToC}
The SharePoint Effect
To create a simple list in 2026, you must still first reckon with the broader ecosystem of SharePoint. A platform long hailed for document sharing and collaboration, SharePoint has nevertheless evolved into a labyrinth riddled with complexity. Central to its inadequacies are decades of accumulated technical debt — the kind that doesn't disappear just because Microsoft has rebranded the front end and bolted on a Copilot button.
The Lists Experience: Grief or Gain?
The "new" Lists experience was supposed to revolutionise the way we compile and interact with data. Users envisaged enhanced performance and well-integrated features. What many found instead was a series of persistent bugs, a UI that oscillates between pretty and punishing, and a product roadmap that has now pivoted so hard toward AI that some of the basics feel more neglected than ever.
The most striking change arriving in 2026 is that the blank-canvas approach to list creation has been quietly retired. Manual list creation has been replaced by the Microsoft 365 Copilot SharePoint List Agent, which generates lists from natural language prompts. In theory, this is a genuine improvement — describe what you need and let the agent scaffold it. In practice, the output still requires significant hand-holding, and users who simply wanted a straightforward blank list now have to navigate an AI intermediary that may or may not understand what they actually need.
Lack of Flexibility
The fundamental rigidity that frustrated users from the outset has not been resolved — it has simply been redecorated. You still cannot import a CSV into an existing list. You can import one to create a new list, and now Copilot will assist with column mapping and data analysis during that import, which is a welcome if modest improvement. But if you want to add more data later via another CSV? You are still, in 2026, out of luck. The data model remains stubbornly flat, with no relational structure and no meaningful schema enforcement. Progress is slow when the foundations are shaky.
Issues, Issues Everywhere
Issues abound, evoking frustration when building even the simplest list. Conditional formatting, driven by JSON layouts, continues to be a particular source of misery with configurations that work perfectly one day silently stopping working after a routine update. Custom views remain fragile, breaking with minor edits. Grouped views on large lists still render slowly. These are not edge cases; they are the everyday experience of a significant portion of the user base.
On the positive side, Microsoft has introduced a native "Quick Steps" column type that replaces JSON-based column button formatting with a no-code interface. For users who were previously forced to hand-craft JSON to add a simple action button, this is a meaningful step forward. Static list forms have also evolved to support dynamic conditional branching logic natively, removing one of the more common reasons people reached for PowerApps when a simpler solution would have sufficed. These are genuine improvements — but they arrive against a backdrop of so many unresolved problems that it is difficult to celebrate them without qualification.
Example: Import a Calendar to a List to Make It More Useful
First, Microsoft Lists does not natively support importing .ics (iCalendar) files or subscribing to a calendar feed.
An alternative is a Power Automate flow that monitors the calendar and pushes events into a list, but this requires ongoing expertise and maintenance. The Teams "Workflows" app now surfaces some of this with natural language automation, which reduces the barrier slightly — but it is still a workaround for something that arguably should be built-in.
The most practical approach remains converting to CSV, importing to create a new list, then modifying column types manually. Copilot's AI-assisted mapping does make the import step less painful than it once was.
What would fix this is proper incremental CSV import capability, or native calendar feed integration. Neither exists at the time of writing.
The Mobile App Is Gone
If the mobile experience was previously described as viewing a large poster through a letterbox, Microsoft has resolved that problem in the most decisive way possible: the Microsoft Lists mobile apps for iOS and Android were retired in November 2025. The browser experience, now accessed via office.com/launch/Lists, is the intended path for mobile users. Whether that constitutes an improvement depends entirely on your screen size and patience. For most people working on a phone, it is a significant regression in convenience, even if the underlying experience is technically more consistent.
This retirement is symptomatic of a broader pattern. Rather than investing in making the mobile experience genuinely good, Microsoft has removed it and pointed users at a browser. It is a pragmatic decision dressed up as simplification.
Microsoft Lists evolved in 2026 with retired mobile app, Power BI button, and manual creation replaced by SharePoint List Agent and Copilot.
A Roadmap to What, Exactly?
The Power BI "Visualize the List" button was also retired in December 2025. Users who relied on it must now use "Export to Power BI" or native Power BI Desktop connections instead. This is not necessarily a worse outcome, but it is another example of Microsoft removing something familiar without ensuring the replacement is equally discoverable or intuitive.
The need for Microsoft to resolve persistent issues cannot be overstated. Users are voicing their frustrations or avoiding the technology altogether — and the feedback loop remains as broken as ever. Official support channels are still populated by responses that answer a different question or offer platitudes in lieu of solutions. The community forums are a graveyard of "we've passed this to the product team" responses that lead nowhere visible.
Microsoft has a set of challenges ahead in 2026 to undo the damage to its brand and reputation that has accumulated in the past couple of years.
Top 5 Pain Points for End Users (2026)
- Fragile custom views — filters and formats break easily with minor edits
- Slow UI rendering — particularly on large lists or with grouped views
- Flaky conditional formatting — JSON layouts still silently fail after updates
- Confusing permissions — hard to know who can see or edit what
- No mobile app — browser-only experience on mobile is a significant step back
Top 5 Pain Points for IT Admins (2026)
- SharePoint dependency — tightly coupled to SharePoint backend quirks
- Brittle data model — no schema enforcement or relational structure
- Copilot integration gaps — List Agent output still requires significant manual correction
- Versioning issues — change tracking and rollback remain unreliable
- Migration and scaling challenges — hard to move or grow lists beyond practical limits
The Feedback Loop Is Still an Echo Chamber
Is Microsoft genuinely listening, or are users simply contributing to an echo chamber filled with platitudes? The rebranding of Microsoft Office to the "Microsoft 365 Copilot app" — and the broader corporate pivot toward AI as the answer to every question — suggests that the strategic priority is AI adoption metrics, not resolving the unglamorous plumbing issues that make tools like Lists genuinely trustworthy. Copilot enterprise rollouts are reportedly resulting in low actual usage after initial licensing, which suggests the market is reaching a similar conclusion.
Is It Really Worth It?
The question remains whether the juice is worth the squeeze. The promise of improved efficiency through Microsoft Lists echoes the software of thirty years ago that promised to re-engineer the corporation. It didn't then. Lists, in its current state, delivers partial on that promise — there are genuine use cases where it works well, particularly for straightforward structured data within a team that already lives in Microsoft 365. But the moment requirements grow even modestly complex, the brittleness becomes the story.
Best Practices for a Less Painful Implementation
For those still willing to invest the effort, the most reliable advice remains the same: keep it simple. Avoid unnecessary customisations. Use out-of-the-box templates where possible, and treat the Copilot List Agent as a starting point rather than a finished product — review everything it generates before sharing it with a wider team. The new native conditional branching in forms and Quick Steps column type genuinely reduce the need to reach for PowerApps or JSON, so lean on those before going anywhere near custom code.
For automation, the Teams Workflows app now provides a more accessible entry point than building full Power Automate flows from scratch. Natural language automation is imperfect, but for simple notification or approval workflows it is meaningfully less painful than it was eighteen months ago.
A Mixed Bag, Still
Ultimately, Microsoft Lists in 2026 is a product in awkward transition. Some of the most frustrating rough edges have been smoothed — native conditional form logic, better import assistance, no-code action buttons. But foundational problems persist: the flat data model, the SharePoint dependency, the fragile views, the permissions confusion. And the removal of the mobile apps, combined with the forced pivot to Copilot-led creation, introduces new friction even as it removes old friction.
The potential was always real. The execution remains, at best, a work in progress. For organisations evaluating Microsoft Lists today, the honest answer is that it works well within a narrow band of use cases and becomes progressively more painful as requirements grow. That has been true since launch. It remains true now. Whether Microsoft's AI-first roadmap eventually closes that gap, or simply papers over it with a more articulate chatbot, is the question worth watching.
- Details
A Journey with Gemini AI towards Modularity and Quality
For seasoned software development professionals , the hype surrounding artificial intelligence often overshadows its practical implications. This article charts a meticulous journey towards refactoring a Node.js application, focusing on how iterative prompt engineering and a modular approach transformed a generic tool into a highly refined, maintainable with 100% test coverage. We'll dissect the truth of integrating AI tools, using our own development process as a case study.
TL:DR – Refactoring a Node.js app to ensure code quality and maintainability can take a great deal of time and effort, but with the help of AI you can achieve impressive results in hours rather than days through an iterative process of prompt engineering. Crucially, this journey transformed a basic app nto a sophisticated one, We'll explore how specific prompt adjustments, HTML sanitization, and a robust testing framework (Jest) were implemented, offering practical insights into building reliable applications with the help of AI.
- Details
In software development, effective collaboration is essential for producing high-quality code in a timely manner. Pair programming, a practice wherein two developers work together at a single workstation, has proven to enhance both the quality of code and the learning experience. With the advent of advanced generative AI tools, notably the Gemini CLI, developers now have at their disposal a new method of collaboration that raises questions about the nature of pair programming. Is using Gemini CLI to assist in writing software the same as engaging in traditional pair programming? This article seeks to explore this query and examine the implications of generative AI in the programming space, particularly regarding collaboration dynamics and output quality.
TL:DR – The prospect of using AI, like Gemini CLI, in software development introduces new methods of collaboration but does not fully substitute for the human elements inherent in pair programming. While generative AI can enhance productivity and output quality when used effectively, the essence of pair programming involves the interpersonal dynamics of two developers working together, and the whole being greater than just the sum of the parts which is something AI, at its current level of sophistication, cannot replicate. Balancing AI assistance with human collaboration is crucial for optimal software development outcomes. That doesnt mean Gemini CLI isnt incredibly helpful, just that you need to use it for what it is!
- Details
Read more: Is it pair programming when you use Gemini CLI to help you write software?
