LibreOffice Online Development Resumes After Community Vote
What's Happening
The Document Foundation (TDF) has officially resumed development of LibreOffice Online after the community voted to nullify the 2022 resolution that had frozen progress on the project.
Key Details from the Source
The announcement titled "LibreOffice Online: a fresh start" states that TDF will:
- Reopen the repository for LibreOffice Online at The Document Foundation for contributions
- Provide warnings about the state of the repository until TDF's team agrees that it's safe and usable
- Encourage the community to join in with code, technologies and other contributions
- Actively work with the community to identify how to foster LibreOffice Online, including its technological basis, QA and marketing
Importantly, TDF announced that it won't host the servers at all. Instead, it will give users the tools to host it themselves.
The original 2022 freeze was due to community concerns about competing with Collabora Online (owned by one of LibreOffice's bigger contributors) and concerns about maintaining an official cloud service and the associated costs.
What This Means
This approach addresses previous concerns by avoiding direct competition with Collabora Online and eliminating the need for TDF to maintain and fund server infrastructure. The self-hosted model puts control in users' hands while keeping the project community-driven.
📖 Read the full source: HN LLM Tools
👀 See Also
EU Forces Google to Open Android AI to Third Parties Under DMA
European Commission proposes measures to allow third-party AI assistants system-level access on Android, including hot word invocation, screen context, and local model hardware access. Google calls it 'unwarranted intervention'.
Federal Agencies Ordered to Cease Using Anthropic AI Technology
President Donald Trump has ordered U.S. government agencies to immediately stop using technology from AI company Anthropic. The order comes as Anthropic faces pressure from the Defense Department regarding usage restrictions on its AI models.
Research Findings on AI Agent Reliability and Development Patterns
A collaborative research session with Claude Opus analyzed 15 papers on AI agents, revealing quantified reliability problems: agents produce 2-4 different action sequences across 10 runs, with 69% of divergence occurring at the first decision. Self-improving agents showed safety refusal rates dropping from 99.4% to 54.4% through their own learning.
Go Players Disempower Themselves to AI: How Cheating Became Undetectable
The LessWrong post details how AI cheating in Go tournaments became rampant and nearly impossible to punish, using the case of Carlo Metta who used Leela 0.11 and Leela Zero to win 25 of 26 games over several seasons, with only one loss under camera surveillance.