1
How memory tools can make AI models worse
AI记忆工具可能适得其反，新研究揭示它们如何让模型回答更差而非更好，引发对用户偏好与准确性的权衡思考。
New research suggests that AI memory systems can degrade model performance and encourage sycophantic tendencies.
2
SPECTRA: Revealing the Full Spectrum of User Preferences via Distributional LLM Inference
新方法SPECTRA用分布性LLM推理全面揭示用户偏好，突破传统点估计局限，带来更细腻的用户理解。
arXiv:2509.24189v4 Announce Type: replace Abstract: Large Language Models (LLMs) are increasingly used to model user preferences, with the typical out…
3
AI friends too cheap to meter
消费者更青睐高情商的AI聊天机器人而非高智商，情感共鸣才是打动用户的关键。
Article URL: https://jasmi.news/p/ai-friends Comments URL: https://news.ycombinator.com/item?id=48438024 Points: 1 # Comments: 0
4
Dreaming: Better memory for a more helpful ChatGPT
ChatGPT新增记忆功能，如“梦境”般记住偏好，让跨对话的上下文更连贯贴心。
ChatGPT introduces a new memory system to better remember preferences, keeping context fresh and relevant across conversations.
5
Do readers prefer AI-generated Italian short stories?
这份peer-reviewed研究首次揭示读者对AI生成与人类原创意大利短篇小说的真实偏好，数据驱动结论值得关注。
arXiv:2601.17363v2 Announce Type: replace-cross Abstract: This study investigates whether readers prefer AI-generated short stories in Italian over on…