Nitter maintainer seeks funding and legal counsel after X Corp. cease-and-desist letters
The Nitter maintainer is seeking funding and legal help to respond to cease-and-desist letters from X Corp. The company sent letters on 24 August and 8 September 2026 demanding a permanent takedown of Nitter instances and the project's repository. The maintainer stated that previous assurances received were untrue, leaving the project without legal counsel. Donations will prioritize legal costs if needed, supporting the development of Nitter or its successor.
nitter.net
·
HN discussion
Fingerprinting as a new tracking standard
The author argues that device fingerprinting is replacing third-party cookies as the primary tracking method. While browsers like Chrome and Edge isolate code execution, they do not prevent fingerprinting, which uses network headers and hardware metrics to uniquely identify users. The article notes that Google's Privacy Sandbox initiatives have failed to gain traction, and ad platforms now offer device fingerprinting as a capability within their products. This shift renders traditional privacy protections insufficient against this persistent identification model.
royfish.dev
·
HN discussion
LineageOS 24 brings Android 17 and new browser-based flashing tools
LineageOS 24, based on Android 17, is now available. The team rebased their changes onto the new codebase faster than anticipated, thanks to prior adaptation work. Security patches through October 2026 have been merged into active branches, while legacy branches received updates through November 2025. The release also introduces new browser-based flashing tools to simplify the installation process for several common device methods. Additionally, the Updater app now streams updates by default, which the team claims will greatly decrease installation time on A/B devices. The project also notes that many core applications have been updated to Material Design 3 Expressive, including a complete redesign of the Updater with better animations and branding.
lineageos.org
·
HN discussion
HashiCorp archives Boundary Community Edition repository
HashiCorp has archived the Boundary Community Edition repository, ending active development for the open-source version. The repository is now read-only and no longer accepts new issues, pull requests, or community contributions. Existing source code and previously released binaries remain available for reference and self-managed use. HashiCorp states that future development will focus on HCP Boundary and Boundary Enterprise to provide enterprise-grade access capabilities. No new releases, maintenance updates, or security patches are currently planned for the Community Edition.
github.com
·
HN discussion
MASS proposes a method for AI models to self-supervise through multi-agent coordination
Researchers introduce Multi-Agent Self-Supervision (MASS), a framework designed to enable homogeneous recursive self-improvement. In this setup, a single language model acts as the task solver, workflow optimizer, and evaluator. The system repeatedly optimizes multi-agent workflows and distills collective intelligence to generate self-supervision signals. The authors argue this approach is necessary because human experts may struggle to assess results on non-verifiable tasks. The paper notes that agreement with external judges rises from 73% to 93% across three cycles, though the samples cover only 8–9 tasks. The authors emphasize that understanding shared blind spots is critical for maintaining safety and oversight in such autonomous loops.
pub.sakana.ai
·
HN discussion
Developer pays users to test README usability
Terence Eden paid volunteers to follow the installation instructions for his ActivityBot project. He recruited participants via Mastodon and conducted video calls where users shared their screens and narrated their actions. Eden took handwritten notes to identify confusing steps, such as incorrect links and ambiguous terminology. He updated the documentation after each session and re-tested it with the next participant. The total cost was approximately €150. Eden argues that real human feedback reveals biases that automated tools or LLMs miss, making the documentation demonstrably easier to follow.
shkspr.mobi
·
HN discussion