Bluesky (Organization): news timeline & CVEs

Automattic has a new board after failed attempt to put CEO on leave

Automattic CEO Matt Mullenweg named a new board after directors failed to sideline him.

Automattic CEO Matt Mullenweg installed a new board weeks after prior directors voted to place him on leave and he retook control in about 33 hours, citing an 84 percent voting stake. New directors include Hugh Howey, Amy Chan, and former IRL co-founders Henry Khachatryan and Krutal Desai. Toni Schneider, Ann Dunwoody, and Sue Decker left the board, and the CFO and chief legal officer were removed. The dispute overlaps Automattic's ongoing trademark litigation with WP Engine.

Hacker News · securitySep 26, 15:40 UTC · 19h agoOther HN 23↑ · 18 comments

YouTube will let you build your own algorithm with AI

YouTube is launching AI-powered custom feeds that let users build their own recommendation algorithms using natural language prompts.

YouTube announced 'custom feeds,' a new feature allowing users to describe and create personalized recommendation feeds using natural language prompts. Powered by Google's Gemini AI, the feeds will be pinned to the user's home page. The move aligns with a trend across social platforms to offer AI-powered feed curation, with rollout starting next month.

TechCrunch · AIupdated Sep 23, 16:37 UTC · 3d agofirst Sep 23, 14:30 UTC · 3d agoAI industry 4 sources

Can Foundation Models Moderate Online Content? Evaluating Instruction- vs. Example-Driven Policy Operationalization

ModerationBench shows foundation models can nearly triple Bluesky's moderation F1 (0.60 vs 0.22), with instruction- and example-driven guidance performing comparably.

Researchers built ModerationBench, a new benchmark of 4,000 manually annotated in-the-wild posts from Bluesky, to test whether foundation models can reliably operationalize content moderation policies. They systematically compare instruction-driven guidance (reasoning from policy precepts) with example-driven guidance (generalizing from precedents) for Vision-Language Models. Both paradigms achieve comparable peak effectiveness, and foundation models nearly triple the F1 of Bluesky's deployed moderation system on Random Posts (0.60 vs 0.22).

arXiv cs.AI / cs.LG / cs.CLSep 9, 16:30 UTC · 17d agoAI research1

Bluesky

Timeline3 articles

Appears with

Automattic has a new board after failed attempt to put CEO on leave

YouTube will let you build your own algorithm with AI

Can Foundation Models Moderate Online Content? Evaluating Instruction- vs. Example-Driven Policy Operationalization