Privacy policy added
A bilingual privacy policy page is now available — find the link in the footer. Briefing card excerpts have also been cleaned up, fixing occasional stray separators.
WeChat entry & richer newsletters
A WeChat Official Account QR code has been added to the nav bar and footer for easy following. Newsletter emails now include a notable papers section so you can get the full picture without visiting the site. Chinese and English subscriptions are now separate channels — subscribe to the one you prefer, or both.
Collapsible overview & cover lightbox
The overview section on article pages now starts collapsed at two lines, keeping the page tidier. Click to expand when you want the full picture. Cover images now open in a lightbox when clicked.
Cleaner navigation & better English writing
The top nav bar has been streamlined to just methodology, search, and language toggle — less clutter, easier to find what matters. A link to the author's homepage has been added to the footer. The site also sports a new favicon.
English briefings are now sharper — tighter phrasing, less filler, better rhythm throughout.
Rename, cover images & font upgrade
The site has been renamed to "AI Research Brief" to better reflect its focus. Each briefing now features bilingual cover images for richer social sharing. Chinese headings use a new serif typeface for a more polished reading experience. Mobile layout has been overhauled — nav wraps gracefully, card info no longer overlaps, and heading sizes adapt to smaller screens.
The briefing writing process has been completely revamped — each briefing now goes through dedicated stages for topic selection, writing, compilation, review, and revision, resulting in noticeably better content quality and consistency.
Email subscriptions are live
Subscribers now receive daily briefings directly in their inbox — bilingual (Chinese + English) with links to the full articles on the site.
Fewer gaps in coverage
Days with fewer papers now still get a briefing — no more missing days.
Localized sharing & auto language detection
Share buttons are now tailored by language: WeChat, Weibo, and X for Chinese readers; X and LinkedIn for English readers. On mobile, you get your device's native share sheet. The WeChat QR code now appears as a small popover instead of a full-screen overlay.
The site auto-detects your browser language and redirects to the matching version. The paper sources page now shows what percentage of papers made the cut, and cover image dates have been corrected.
Topic pages & visual refresh
New topic pages let you browse past briefings by theme — multimodal, agents, reasoning models, and more. Each briefing now gets an AI-generated cover image that shows up when you share links on social media.
The homepage has been redesigned with a new tagline, rewritten FAQ and methodology page, and a more prominent subscribe box. Article navigation now shows related briefings instead of plain prev/next links, making it easier to discover content you care about.
English body text now uses Source Serif 4, and CJK-Latin spacing is handled automatically by the browser for a cleaner look.
Site search is here
You can now search across all past briefings by keyword using the new site search.
The Featured and Overview sections now show the same number of items for a more consistent layout. Scroll animations on article content have been removed — content now appears instantly.
English version, homepage upgrade & paper sources
The site is now fully bilingual — every briefing is available in both Chinese and English. URLs have been changed to a keyword-based format that's easier to read and better for search engines.
Homepage cards now show the actual briefing content for each day instead of a generic "AI Daily | Date" title. Each briefing now includes a "Paper Sources" link at the bottom, where you can view all papers covered in that issue (titles, arXiv links, scores, and scoring rationale).
Site launch
The site is now live at ai-brief.liziran.com — daily AI briefings from a practitioner's perspective.
RSS and email subscriptions are available via the footer.
First briefing
The first AI briefing is published, covering reasoning model safety, deterministic reasoning, and diffusion language models.