Papel is an iOS app offering a TikTok-style full-screen feed for academic papers, with on-device AI chat (Apple Intelligence/MLX), quizzes, and researcher social features.
Key Takeaways
Discovery feed uses vector-similarity recommendations ranked by interests, trending topics, freshness, and community engagement; switchable to a Latest mode.
AI chat runs a RAG pipeline over full PDF text entirely on-device via Apple Intelligence or local MLX models, no data leaves the device.
Gamification layer generates 3-question quizzes per paper with XP, streaks, and academic ranks from Undergrad to Nobel Laureate.
Community features include likes, comments, direct messages, and academic profile badges tied to rank progression.
Hacker News Comment Review
Core skepticism centers on the cold-start problem: without critical mass of interactions, a social layer on papers has no pull, as alphaxiv.org already demonstrated with sparse engagement despite existing infrastructure.
Multiple commenters question format fit: papers demand sustained reading on desktop or tablet, not mobile swipes, making the TikTok framing more a liability than a hook for the researcher demographic.
The “TikTok” and “AI” branding were flagged as polarizing terms that may repel the exact audience Papel needs, with the underlying recommendation engine seen as the actual differentiator worth leading with.
Notable Comments
@taikon: argues mobile reading habits cap out at abstracts, suggesting concise bullet-point abstract summaries as a more realistic wedge feature.
@DonaldPShimoda: warns AI summaries should not be default or replace author abstracts, citing ACM’s reversal after researcher backlash over similar integration.
