Kitchen-Sune: A Pandemic-Born Community Cookbook Built by Front-End Foxes

·1 views

Kitchen-Sune is an open-source international recipe web app created by members of Front-End Foxes, a nonprofit, when the organization shifted from in-person workshops to an online boot camp during the COVID-19 pandemic. Built using VuePress, the platform collects comfort food recipes from contributors across the globe, including dishes from Ukraine, Kenya, and Nigeria. The project also served as a practical training tool, teaching boot camp enrollees how to use GitHub and submit pull requests. The developer revisited the project as a submission for a Frontend Challenge, highlighting its clean, accessible design and its role in fostering an inclusive developer community. The repository remains open for contributions, with anyone able to add their own recipe via a pull request on GitHub.

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)

Nassim Taleb Challenges Methodology of Government-Funded Alcohol Research

Nassim Taleb, statistician and author, has published a critique questioning the scientific rigor of a government-sponsored study on alcohol consumption. Writing on his Substack, Taleb argues the research does not hold up under closer statistical scrutiny. His post attracted discussion on Hacker News, where it received points and reader comments. The critique fits into Taleb's broader pattern of challenging flawed statistical reasoning in published research. No specific government or study was identified in the available source material.

0 comments Read more at Hacker News

ProgrammingDEV Community ·

How to Build Tenant-Aware Speech-to-Text Uploads for Multi-Tenant SaaS Products

A software architecture guide recommends that small fintech teams building voice-note transcription tools start with a single synchronous file-upload adapter supporting MP3 and WAV formats. The approach emphasizes logging every upload to a per-tenant ledger before sending audio to a transcription API, ensuring cost visibility across clients. Developers are advised to record key metadata such as tenant ID, file size, media type, and processing region both before and after each transcription request. The guide warns against relying on request counts alone, noting that usage patterns vary widely between tenants and can make billing and abuse detection unreliable. It also stresses keeping provider-specific details hidden behind an adapter layer so the application remains portable and maintainable as infrastructure evolves.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

DeepSeek Introduces Time-Based Token Pricing With Higher Baseline Rates

DeepSeek has rolled out a new peak and off-peak billing structure for its API, effective August 16, 2026, offering off-peak rates at 50% below peak prices. However, the new off-peak prices are still significantly higher than the previous flat rates — DeepSeek V4 Pro input tokens are now roughly 1.5 times more expensive off-peak and up to 3 times more at peak compared to the old pricing. The steepest increase affects cache hits, with V4 Pro cache-hit rates rising up to 6 times off-peak and 12 times during peak hours, a change that heavily impacts long-running agentic workloads. Peak hours are defined as 01:00–04:00 UTC and 06:00–10:00 UTC, which broadly align with China's business day, while the remaining 17 hours qualify as off-peak. Developers using DeepSeek in batch or agent-heavy applications are being advised to schedule intensive workloads during off-peak windows to partially offset the overall price increase.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

11 Real Failures a Non-Developer Hit While Self-Hosting an Agent Gateway on Mac

A small agent organization documented every failure point when its non-developer operator attempted to self-host an agent gateway on a Mac in a single session. The 11 breakdowns included permission errors, misconfigured npm prefixes, missing PATH entries, and typos caused by retyping instructions from images rather than running executable files. A recurring theme was that fixing one issue directly created the next, forming chains that official documentation consistently failed to cover end-to-end. One security-relevant failure involved a Telegram bot search surfacing impersonators, risking token theft if the wrong account was used. The account concludes that non-developers do not fail because steps are hard, but because documentation stops just before the next problem begins.

0 comments Read more at DEV Community