Future Of SoftwareLightning talk20min
The Future of Open Source Has New Faces
The talk shows how open source is becoming more accessible, using the speaker’s path from coding school to merged OpenTelemetry contributions. It covers the work behind a first PR, changing maintainer roles, and why barriers to entry are falling for new contributors without traditional backgrounds.
talk.summaryAiDisclaimer
Mohammed Abdessetar ElyagoubiSofrecom Services Maroc
Most of the software running the world depends on open source maintained by a small group of people. That group is changing, and I'm one example of how.
Two years ago I was at YouCode, a coding school in Morocco. Since then I've had four patches merged into OpenTelemetry Java, a CNCF project in production worldwide. Three of them are different ways a telemetry library can lie to you: a DoS vector in the HTTP sender, silent data loss in async metrics, and config silently ignored. No CS degree. The skills came from building and reading other people's code.
In 20 minutes I'll walk through what the first one took: reading an unfamiliar codebase, isolating the bug, writing the fix, and defending it in review until a maintainer merged it. Twelve days. The fourth took two.
My argument: the on-ramp changed, the bar didn't. Open knowledge, communities, and AI tooling speed up navigating a large codebase, but none of that writes a fix that survives maintainer review. Engineering judgment does. What fell is the credential barrier, so the next wave of contributors won't fit the old profile, and maintainers shift from writing code to vetting and mentoring it.
Takeaways: a realistic path to a first merged PR in a large codebase, what maintainers check before trusting an unknown contributor, and how the contributor and maintainer relationship changes as the pool widens.
Who this is for: contributors who've been told they're not ready, and maintainers deciding how much to trust them.
Two years ago I was at YouCode, a coding school in Morocco. Since then I've had four patches merged into OpenTelemetry Java, a CNCF project in production worldwide. Three of them are different ways a telemetry library can lie to you: a DoS vector in the HTTP sender, silent data loss in async metrics, and config silently ignored. No CS degree. The skills came from building and reading other people's code.
In 20 minutes I'll walk through what the first one took: reading an unfamiliar codebase, isolating the bug, writing the fix, and defending it in review until a maintainer merged it. Twelve days. The fourth took two.
My argument: the on-ramp changed, the bar didn't. Open knowledge, communities, and AI tooling speed up navigating a large codebase, but none of that writes a fix that survives maintainer review. Engineering judgment does. What fell is the credential barrier, so the next wave of contributors won't fit the old profile, and maintainers shift from writing code to vetting and mentoring it.
Takeaways: a realistic path to a first merged PR in a large codebase, what maintainers check before trusting an unknown contributor, and how the contributor and maintainer relationship changes as the pool widens.
Who this is for: contributors who've been told they're not ready, and maintainers deciding how much to trust them.
Mohammed Abdessetar Elyagoubi
I'm a Software Engineer at Sofrecom Services Maroc (Orange Group), where I build microservices-based backend systems on PaaS. I was a student at YouCode, Université Mohammed VI Polytechnique. I actively contribute to open source. I've had multiple pull requests merged into the OpenTelemetry Java SDK, including a bug fix preventing silent data loss in async metrics. I build across the stack: from automated Darija-language media pipelines and AI agents to electronics embedded projects. I organize hackathons and workshops in the Moroccan tech community, and I'm passionate about making developer knowledge accessible in Morocco and beyond