Outreachy report: August 2025

Summary

Program activities

• Our June 2025 cohort will soon end with a 96.15% internship success rate — we fostered 25 successful internships across 12 mentoring organizations.
• We have announced our December 2025 cohort. We have published our call for mentoring organizations and we have opened initial applications today. Our blog posts includes announcements about (1) our generative AI policies (2) changes to program activities.

Open Mentorship Handbook

• I decided to adopt a documentation-as-code approach for the Open Mentorship Handbook. In my research, I realized that the majority of platforms and tools only officially support GitHub, GitLab and/or Bitbucket repositories. I started experimenting with Read the Docs, but Forgejo isn’t officially supported — using unsupported platforms heightens the risk of running into breaking changes.

Description of activities

Program activities

Our June 2025 cohort will soon end with a 96.15% internship success rate — we fostered 25 successful internships across 12 mentoring organizations. We started our June 2025 cohort with 26 interns:

• 1 intern failed their midpoint evaluation. This internship was terminated back in July.
• All remaining 25 interns have passed their midpoint evaluation. All 25 second stipend payment authorizations were fully processed by the program.
• 24 out of 25 remaining interns will wrap up their internships at the end of August.
• 1 intern will finish their internship in September due to a 3-week extension.

Another important datapoint: we rejected two extension requests for two different interns. Both mentors requested an extension to completely wrap up their internship projects, but we don’t approve extension requests of that nature.

We’ll host our very last text-based intern chat on Zulip tomorrow (August 26, 2025). Text-based scripted intern chats are one of the activities we’ll phase out in the December 2025 cohort. They will be replaced by three scheduled but unscripted social sessions on BigBlueButton. This change addresses (1) our observations about the decrease in engagement and interest in scripted text-based chats in recent cohorts (2) feedback we received from coordinators and mentors of mentoring organizations during our community listening sessions.

We have announced our December 2025 cohort. We have published our call for mentoring organizations and initial applications. The key dates for this cohort are:

The publication of the calls for mentoring organizations and applicants includes sections about our policy on generative AI and changes to program activities.

We’re asking applicants to refrain from using generative AI to write their initial applications. We want to encourage applicants to use their authentic voice to describe their lived experiences when answering essay questions. The use of LLMs in initial applications the past 3 years has impacted the quality of essays we’ve received in our last 6 cohorts — they’ve become more generic and impersonal. We believe this is having a significant impact on the amount of applications accepted each cohort — so we’re asking applicants to invest some personal time writing their initial applications.

Additionally, we’ve added a new field to project descriptions so mentoring organizations can disclose their own policies concerning the use of LLMs and other tools during the contribution period and the internship itself. This addition addresses feedback we’ve collected in the last 6 cohorts and in our community listening sessions.

Lastly, we’re announcing a couple of significant changes to the program:

• We’ll no longer host text-based intern chats on Zulip — they will be replaced by three unscripted social sessions hosted on BigBlueButton.
• We’re retiring our blog prompts and bi-weekly assignments. Interns will now be required to write three public reports on the progress of their projects — a starting report, a midpoint report and a final report.
• We’re phasing out the facilitation of informal chats with members of the free and open source community.

Open Mentorship Handbook

Sage Sharp and I attended a meeting of the FLOSS Mentoring initiative. Their discussions are a major source of inspiration and resources for us: they’re discussing mentorship relationships in open source from several different perspectives (educators, researchers, practitioners).

I spent some time reading discussions taking place on their platforms and artifacts of sessions they’ve hosted at different conferences. My initial impression is that there’s a gap in perspectives from the Global Majority, and that we (as in “Outreachy as a program”) have developed significant expertise in policies and practices that address common struggles from marginalized populations (e.g. lower self-confidence, unstable or difficult access to electricity and/or internet connectivity, financial control risks).

On the more technical side of things, I had a meeting with Karen Sandler where I told her I was inclined to use Read the Docs to deploy and maintain the Open Mentorship Handbook. However, I was surprised to learn from Codeberg’s official documentation that Forgejo integration isn’t officially supported by Read the Docs — and Software Freedom Conservancy has been using Forgejo for its own repositories.

Codeberg’s documentation offers a way to integrate both platforms, but going an unofficial route means raising the risks of facing breaking changes in the middle of the project. Therefore, a frugal and more trusted path for the deployment of the Open Mentorship Handbook may entail using a platform such as GitLab.

The discussions about technical decisions were interrupted by a couple of unplanned but more critical discussions about program activities. More concrete answers to the questions I asked in the previous report are all gated on those technical decisions. My plan for September to make a choice for once and for all.