Why building another to-do list app can still work
Task managers are the classic example of a market everyone warns you away from: a native default, well-funded competitors, and countless half-finished side projects. Austin Blake went the other direction, shelving five other projects to concentrate on Stuff, a task manager for people who care about how the software feels. In a conversation with host Charlie Chapman, he described the build workflow, the trial change, and the review delays that shaped the Mac release. This guide turns those specifics into decisions you can check against your own launch.
The crowded-market objection, and what a commitment actually requires
The standard advice is simple: don't build a to-do app. Austin heard it directly, including blunt messages asking why he would waste time on another one. His answer was not that the market is empty, but that additional competitors can still find room if the builder commits and keeps investing time.
That reframes the decision. The question is not whether a category is crowded but whether you can sustain a differentiated product long enough to matter. Stuff's angle was design and feel rather than feature count, and the effort was concentrated: five other projects were dropped so the work could be exclusive.
Before committing, ask yourself:
- Can you name the specific quality your app will be judged on, beyond a longer feature list?
- Are you willing to pause other projects for it?
- Can you absorb critical feedback without abandoning the plan?
A crowded category raises the bar for differentiation; it does not remove the possibility of a viable product.
A multi-agent planning loop for unfamiliar platforms
Austin had never shipped a Mac app. He began with a conventional route, a Swift macOS book, then moved to an AI-assisted workflow in which Claude was not simply asked to produce code. Instead, plans were written for each piece of work and passed between agents so they could critique one another's plans, with the developer reviewing all code before it was merged.
The structure matters more than the tool. One agent drafts a technical plan, another challenges it, and a third writes code against the revised plan. That loop gave a first-pass prototype without requiring hours of learning every AppKit and macOS-specific API up front, which shortened the path to release.
A practical version of the loop:
- Write the plan as a reviewable document, not a prompt.
- Have a second pass attack assumptions, edge cases, and platform conventions.
- Generate code only from the revised plan.
- Read every line before it enters the codebase.
The human review step is the control. Without it, the loop produces confident output with no one accountable for correctness.
Desktop expectations that a mobile port will miss
Moving an iPad app to the Mac is not a checkbox in Xcode. Larger iPad screens do not prepare you for desktop habits, and two interaction patterns stood out during the build.
The first is selection. On iPhone, tapping a task opens it. On a Mac, people expect to move with arrow keys, highlight a row without opening it, and press Enter to expand details. That intermediate highlighted-but-closed state had to be added.
The second is undo. Completing a task and reversing it with Command-Z is a baseline desktop expectation, even though nobody shakes a phone to undo something.
Window behavior is the third difference. Mobile design optimizes for full-screen layouts; desktop users often shrink productivity windows into small panels parked in a corner. Stuff was built to shrink far enough that monthly goals could sit in a tiny box on screen, an affordance that suits task apps particularly well.
Use this as a porting checklist: selection state, keyboard navigation, undo, and small-window layout. Each is a place where a straight port feels wrong on desktop.
Cutting a 30-day trial to seven days
Stuff launched with a one-month free trial, which seemed generous. The result was heavy drop-off: with 30 days available, users had no reason to form a daily habit, and many forgot the app before the trial ended.
Working from observation rather than a formal experiment framework, Austin shortened the trial to seven days and saw conversions rise. The reasoning is about attention, not pressure tactics: a shorter window forces an active decision while initial interest is still fresh, instead of letting the trial quietly expire in the background.
Treat this as a hypothesis to test in your own product, not a universal rule. The reported result came from a gut-driven change without complex A/B tooling, so it does not isolate trial length from seasonality, marketing timing, or other shifts. If you want a defensible answer, change one variable at a time and watch trial-to-paid conversion alongside early engagement.
A simple decision frame:
- If trial users go quiet in the final weeks, the window may be too long for your usage rhythm.
- If support requests spike right after expiry, the window may be too short to reach first value.
- Define what "first value" looks like before you change the length.
Budget review time for the platform, not just the features
The Mac submission was sent a week before the planned launch, which felt like a comfortable buffer. It was not. The app was rejected repeatedly, and the rejections concerned features that had already been live and approved on iOS for years. The back-and-forth pushed the release three days past the target and disrupted a coordinated marketing push.
The lesson is that a new platform can reopen questions about functionality you consider settled. Reviewers evaluate the build in front of them, and prior approval on another platform is an argument to make, not a guarantee.
Austin's revised guidance was to allow roughly three weeks. Build that into the schedule as working time for responses, not dead time:
- Submit early enough that a rejection does not move your launch date.
- Keep a written record of prior approvals for features you expect to be questioned.
- Hold marketing announcements until the build is actually approved.
Background behind the episode
The full conversation also covers Austin's route from a film major at BYU to contract work at Apple, the Apple Newsroom feature that highlighted Stuff's on-device foundation models, and his move to a developer advocate role. Those details sit in the episode itself; the sections above focus on the decisions a builder can reuse.
| Decision area | What was done | Reported outcome | What to check in your app |
|---|---|---|---|
| Category choice | Dropped five other projects to focus on one task manager | Positioned as a design-focused alternative in a crowded market | Name the quality you compete on and whether you can sustain the focus |
| Build workflow | Plans passed between AI agents for critique, with all code reviewed by the developer | First-pass prototype without hours spent learning each macOS API | Keep a human review gate on every generated change |
| Trial length | Reduced from one month to seven days | Conversions rose, based on the developer's own observation | Test one variable at a time and track trial-to-paid conversion |
| Review buffer | Submitted one week before launch | Repeated rejections over features already approved on iOS; launch slipped three days | Allow roughly three weeks and hold announcements until approval |
Figures and outcomes are as described by the guest in a June 2026 episode and reflect one developer's experience, not a controlled study or a benchmark for other apps.
Frequently asked questions
Is it worth building an app in a market that already has many competitors?
It can be, if you can commit to a specific quality and keep investing in it. The developer behind Stuff treated the crowded category as a reason to differentiate on design and feel, and he concentrated his effort by dropping other projects rather than splitting attention.
How were AI agents used to build the Mac app?
Plans were written for each piece of work and passed between agents so they could review one another's plans, then code was generated from the revised plan. The developer reviewed all code before it went in, which kept a human accountable for what shipped.
What breaks when you port an iPhone or iPad app to the Mac?
Interaction expectations change. Desktop users want to highlight a task with arrow keys without opening it, press Enter to see details, and undo a completed task with Command-Z. They also expect windows to shrink into small panels rather than fill the screen.
Should I shorten my free trial?
Consider it if users go quiet before the trial ends. In this case a one-month trial produced heavy drop-off and a seven-day trial was followed by higher conversions, but that change was made on judgment without formal A/B testing, so test it in your own product before adopting the number.
How long should I allow for App Store review of a new platform build?
Roughly three weeks is a reasonable buffer if the release matters to a marketing schedule. A one-week buffer was not enough here, because features already approved on iOS were questioned again for the Mac build and the launch slipped three days.