Oversight Board/Meeting Minutes-2025-12-17

Dec 17 2025 Meeting

Attending

• Devin Ulibarri (SL Inc. board)
• Samson Goddy (SL In. board)
• Sumit Srivastava (SL Inc. board)
• Safwan Sayeed (community member)
• suryaansh (community member)
• MostlyK (community member)

Absent

• Walter Bender (SL Inc. board)
• Claudia Urrea (SL Inc. board)
• Sebastian Silva (SL Inc. board)

Agenda

• Election report by myself and Safwan
• Any end-of-the-year work
• AWS and infrastructure
• Let's talk about our meeting time
• Anything else

Election report

• See https://www.sugarlabs.org/news/community-news/fall-board-elections-2025-how-to-participate
• It looks like we received 15 applicants to vote in the election.
• However, we have any candidates at this time.
• Candidates are suppose to add themselves here: https://wiki.sugarlabs.org/go/Oversight_Board/2026-2027-candidates

Any end-of-the-year work

• The Sweet Spot newsletter was sent out: https://www.sugarlabs.org/news/community-news/2025-12-16-sweet-spot-004
• We welcome any reflections on work done this year. You can propose an article to the website by creating a post in https://github.com/sugarlabs/www-v2

AWS and infrastructure

• The rationale for the credits Devin requested was to test GSoC projects -- specifically GSoC/DMP/SSoC -- because the concern was that the contributors needed to be able to test their work in order to best understand their work.
• bedrock access gateway would still support all the latest models without any: https://github.com/sum2it/bedrock-access-gateway
• ai.sugarlabs.org gives out access to contributors on using the Sugar-AI API key which will infer the model on our AWS server, it acts like any other LLM service provider but here we are the one hosting LLM on our AWS provided GPU.
• There is a sugar-ai activity which can query using this service and acts like an activity, most contributors this Summer worked with Gemini End-Points and added Sugar Endpoints for testing too, so we know it works.
• If we get an LLM service provider like Google, Anthropic, OpenAI, we will be happy to swift our end-points to access from their service and Sugar-AI keeps running, so anything tested on Sugar-AI remains functional and we have 2 amazing things.

"Let's talk about our meeting time"

• We agree to (1) skip 12/24 and 1/7 and (2) communicate async for anything urgent until we meet again 1/7

Other

• Special acknowledgement to Krish for presenting at GNOME Asia: https://www.youtube.com/live/WZ63lQ-DsOA?t=4h5m18s
• Source for slide deck: https://github.com/MostlyKIGuess/gnome-asia-talk