Developer builds OTP verification screen in 30 minutes using one AI prompt at droidcon Kenya

·1 views

At droidcon Kenya 2026, twenty teams were challenged to build an OTP verification screen from a single AI prompt within a 30-minute window. Submissions were compiled and tested by a CI pipeline, then scored out of 25 by an AI grader. The author built a Flutter version that initially scored 23/25 and reached 25/25 after splitting the screen into smaller components and shortening the prompt. The required interface included six digit boxes with several states, keyboard and paste behaviour, a resend countdown, and dark mode support. The tests covered the requirements most easily faked, such as copy, input handling, and the resend timer.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Why React's Simplicity Hides Friction in Production Apps

Before React arrived in 2013, many frontend apps relied on jQuery, where developers manually located elements and synced data, often producing fragile code. React instead treats the interface as a function of state, composing small components with one-way data flow and letting the library handle DOM updates. Its lightweight, predictable model drew a large community and now dominates web-framework rankings and job listings. The article argues that its 'it's just JavaScript' promise conceals trade-offs in production-scale work. It illustrates this with JSX pitfalls, such as a falsy zero rendering on screen through the && operator, and with the unique key requirement for list rendering.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Myriad Immortal Sect Revises Quest Slot Rules After Players Get Stuck

A three-quest slot cap in the game Myriad Immortal Sect left players unable to accept the closing quest for a realm, since leftover side errands filled every slot. Developers first added a narrow exception allowing one main-road quest a fourth slot, then widened it to cover unclaimed main quests and their prerequisites. Capability quests, such as the one unlocking flying, were later brought into the system, which caused further conflicts at realm boundaries. The result is two separate lanes on top of the three ordinary slots, allowing up to five quests in progress at once. Developers also changed how new quests are placed in the log, which affects what the map tracker points players toward, and corrected the Quest Hall labels that had wrongly marked a required quest as optional.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Yonto Player explains its Android TV stack: Kotlin, Compose, media3 and libVLC

The developer of Yonto Player, an Android TV app started on 15 September 2026, chose its technology stack by working backward from the functions the app needed to perform. Kotlin and Jetpack Compose, with the tv-material library, were selected because they suit Android TV and make remote-driven focus manageable. Media3 and ExoPlayer handle most files by using the device's own hardware decoders, along with HLS and DASH streams. For formats the hardware cannot read, the app falls back to a second engine, libVLC, built in-house under the LGPL, with a similar fallback path for certain audio codecs. The author acknowledges the trade-offs, including added testing burden, a larger download size and the maintenance of a native build script.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Altman Says Data Centers Use Scarce DRAM Better Than Idle Home Computers

OpenAI CEO Sam Altman remarked on a Cheeky Pint podcast in late October that DRAM memory is a scarce global resource and that data centers allocate it more efficiently than personal computers, where it often sits idle. His comment about local AI inference being a poor use of limited supply went viral and sparked memes, especially in India, with many users reading it as an attempt to take RAM away from individuals. Altman later said the remark was taken out of context and that he wants the world to have plenty of memory, but that supply is currently constrained. Elon Musk amplified the meme-driven reaction, while commentators noted that Altman's allocation argument holds on efficiency grounds even as local-model users prioritize privacy and ownership. The debate has highlighted a trade-off between cloud efficiency and local control amid a memory shortage.

0 comments Read more at DEV Community