Urgent.News

What's breaking now, across thousands of outlets.

Tech

We don't need more apps. We need better open source

Every day, dozens of individuals create new applications using AI over the weekend, ranging from habit trackers to note-taking tools. While I appreciate the enthusiasm, the majority of this effort does not focus on enhancing the open source software we already depend on. People often refer to this new approach as "vibecoding," a term I dislike as it suggests a frivolous way of writing software. I prefer to call it "AI-assisted coding." I believe this could lead to more meaningful results.

The open source software we currently use generally meets most of the needs of my clients and me. However, it often falls short when it comes to running efficiently on low-cost servers, easy installation, or maintaining uptime without constant forum assistance. As someone who self-hosts various services and assists progressive nonprofits with their technology, I have firsthand experience of the challenges involved.

Running the software is often the difficult part. Installation guides typically assume knowledge of additional tools, upgrades can break the system, and apps tend to consume excessive amounts of RAM for little return. Error messages provide little guidance. Many nonprofits, seeking privacy, cost savings, and data ownership, would prefer to use open source alternatives.

However, the open source options often require a dedicated system administrator, which many cannot afford, leading them to purchase corporate solutions instead.

This gap in open source development was highlighted in 2024 when a malicious actor exploited the trust of a lone maintainer of a widely-used compression tool in Linux systems. I believe this type of work is well-suited to AI, which can profile slow code, write comprehensive tests, assist with upgrades, improve error messages, and create missing documentation. This type of work demands patience and meticulous attention to detail, combined with a willingness to verify every line of code.

On a political level, it is crucial that open source software becomes accessible to smaller organizations that wish to own their data rather than paying for it from a few dominant corporations. AI can help make these tools more powerful and widely available. By improving public tools that anyone can use, modify, and share, we can work towards a future where AI's benefits are shared more equitably.

Maintainers have already witnessed the downside of relying solely on AI-generated reports. The curl project, for instance, suspended its bug bounty program in January 2026 due to a surge of AI-generated security reports. By April, the lead developer noted that most of these reports were genuine but AI-generated, and most were of high quality.

The real issue is the volume. Even well-intentioned work can overwhelm a maintainer. Therefore, I suggest taking an open source tool you frequently use, identifying the most frustrating aspect of its installation or operation, and dedicating a weekend to fixing it with the help of the project's maintainers. Once you've made improvements, share your experience.

Written by urgent.news from Hacker News's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at jordankrueger.com →

More in Tech

More from Tuesday 17 November →