Tag
The author compiles a personal "AI Interaction Friction Vocabulary"—a list of recurring failure modes in AI conversations (e.g., default-user bias, sycophancy, template imposition, intake loops) paired with concise "Instead" directives that can be pasted into an agent's instructions to correct the behavior mid-conversation.
Anthropic is launching a new study using Anthropic Interviewer to collect user experiences with AI, with the option to make interviews public to inform AI development and policies.
A user questions the removal of mid-thinking message streaming in Claude Code CLI, highlighting its past benefits and causing confusion over shortcuts.
This paper evaluates LLM systems for pre-submission peer review, demonstrating that broad generation can recover most historical review issues but compressing them into a short report is challenging, with implications for AI-assisted research feedback.
This article discusses user-reported issues with the Ling 3.0 AI model from Nous, focusing on problems with text output and context length limitations, potentially linked to quantization or rate limits.
Claude Code introduces a SendFeedback tool that allows users to draft and approve feedback reports directly, improving the feedback process for AI development.
Eddie AI offers time-stamped feedback on videos from your team, enhanced by AI.
Patchrooms is a tool that converts AI app feedback into agent-ready patch context for developers.
User reports that Gemini previously provided useful audio feedback on music tracks but has stopped recognizing or analyzing uploaded files in the same chat.