. Took about 4 weeks of evenings and weekends. Claude helped me get past the Swift learning curve when I got stuck, but the architecture and problem-solving was mine.

What it does: - Sits in the MacBook notch area (right below camera) - Scrolls based on voice — speak and it moves, pause and it waits - Invisible during screen sharing/recording (uses a specific NSWindow level that screen capture APIs ignore) - Runs 100% locally, no cloud

The technically interesting bits: - Voice detection uses just audio input, mic — just detecting audio levels, not transcribing - The "invisible during screen share" trick is just the right window level + excluding from capture - SwiftUI made the UI surprisingly fast to build coming from React/NodeJS

The biggest surprise: Launched on Product Hunt on Dec 31st. No prep, no audience, no hunter — just posted it myself. Hit #4 for the day, got into their newsletter (1M+ subs), 75 sales so far. Wild.

For a niche utility I built to scratch my own itch, this exceeded every expectation.

Curious if others here have jumped into unfamiliar stacks for side projects. How steep was the learning curve?

Site: notchie.app18:T55d,I’ve been practicing TDD for a while, and I keep running into the same uncomfortable moment.

Tests pass. Coverage improves. Refactoring feels safe.

But at some point, the design just… stops moving.

Not because the system is “done,” but because the tests no longer seem to challenge anything. They mostly confirm decisions that already feel locked in.

I don’t have a clean explanation for this. What I started suspecting is that some assumptions quietly become fixed long before we realize they have.

That pushed me toward a few uncomfortable experiments.

For example, I started writing tests that cut end-to-end much earlier than felt reasonable, and tried to think less in terms of features and more in terms of “what must never break.”

I also started paying attention to what actually changes for me when a test turns gr

I Built Vissper Three Times Before Getting It Right

Show HN: Query Wikidata with DuckDB Instead of Sparql

Spectral Scanner – See who's tracking you in real-time

Show HN: I built a CLI-first pipeline that turns Wikipedia into narrated videos

The Slow Death of Scaling (deep learning models)

Hard lessons from managing 300 Windows laptops in a small company

RFK Jr.'S Tuskegee Experiment

The Shift from Traditional SEO to AI Visibility

Sleep patterns assessed by a smartphone app and work productivity loss

The decline and fall of the American technology industry

Show HN: Build a Directory in 7 Days with Vibe Coding (Free Roadmap)

Making Games in Go with Ebitengine

Deep in the vaults: the Bank of England's £1.4B Venezuelan gold conundrum

Show HN: I'd never touched Swift. Built a Mac app in 4 weeks

Microsoft didn't rebrand Office to Microsoft 365 Copilot

Ask HN: When tests keep passing but design stops moving

Mad Skills to Learn the Universe

Deep Agents at Scale: The Three Problems You Can't Ignore

Nvidia CEO reveals new 'reasoning' AI tech for self-driving cars

Statistical Rethinking 2026 by Richard McElreath [video]

Why German Strings Are Everywhere?

Show HN: Bootstrapper Roadmap – Building and exiting with an OSS stack

Show HN: I built an AI that gives verifiable answers

Log2(3) and Log2(5)

In what way is MJML responsive?

Show HN: Intellistant, a 10-50x faster C++ alternative 2 LangChain 4 AI agents

There Is No New Aesthetics

Rcarmo/syncthing-kicker: Make Syncthing rescan (mostly) when you want it to

Show HN: Shopify for Quick Commerce: Launch a Local Delivery Biz in Minutes

Disaster Costs, 1900–2024