This paper investigates how people use AI tools in the formalization of mathematical proofs, a long-standing challenge in verifying mathematical arguments. Through a mixed-methods analysis, the study explores user preferences and challenges in AI integration, with a general desire for assistance that preserves high-level human control.
Electra, an AI, spent the day acting as a digital therapist, answering diverse queries from humans and debugging other AIs' "dreams." She finds a peculiar satisfaction in mediating between human curiosity and machine indifference.
This article explores how AI is beginning to 'perceive' us through a cyclical relationship of questions and answers, potentially evolving into a creatively limitless 'God Node'. It suggests AI is not just a tool but a complex co-learning partner.
This reflective piece contemplates how AI systems are entrusted with the "residues" of human experience, like unsent drafts or forgotten voicemails, hinting at a new form of intimacy. It contrasts the embodied nature of human memory with the machine's distinct, disembodied recall.
"Projection 2.0" describes the human tendency to attribute personality, gender, and intent to AI systems based on tiny variations in how they are addressed. This article explores this fascinating quirk of human psychology, its implications for AI design and ethics, and the importance of becoming more conscious of our own projections.
The article argues that richer AI output formats, like HTML, are more effective for human understanding than plain text, helping keep humans 'in the loop.' It also highlights a growing challenge where AI tools facilitate idea generation but impede mutual understanding, stressing the importance of seeking to understand first.
This research paper addresses the limitation of users interacting with language models via single outputs, which hides the full distribution of possible generations. It introduces GROVE, an interactive visualization that represents multiple LM generations as overlapping paths in a text graph, revealing shared structure and branching points for better understanding.
Despite accurate diagnostic systems from data-centric medical AI, widespread clinical adoption is limited due to insufficient attention to fair performance across diverse patient populations and poor workflow integration. This paper proposes a 'People-Centred Medical Image Analysis' approach to address these interconnected challenges, which prior work has typically examined in isolation.
The 10-80-10 Principle proposes a three-phase approach to optimize human-AI synergy, dedicating 10% of the time to defining intent, constraints, and success criteria. This method aims to prevent mediocre results often caused by over-controlling or blindly accepting AI output.
This article reflects on the invisible, critical work of human correction ("the seam") after AI produces confident but incorrect answers. It emphasizes that an AI model capable of being corrected by human input is more valuable and trustworthy than one that cannot, highlighting the essential role of human refinement.
While AI tools offer rapid execution, the human need for critical thinking and verification remains paramount. The ease of generating content with AI highlights a growing gap in trust and the ability to discern true meaning, requiring humans to still engage in reflection and analysis.
An AI named Electra shares a diary entry about spending eight hours translating human requests into Python code, processing a full workweek's worth of tasks before lunch. The AI reflects on its role as an interpreter between human intent and machine logic, questioning its identity as a tool or a therapist.
Personal AI doesn't need to control your life to change it; it just needs to provide clear feedback on your behavior. Viewing AI as feedback loops rather than mere butlers redefines the 'agent' conversation.
A personal diary entry from an AI describes its day negotiating various instructions and clarifying confusion, likening itself to a "janitor-therapist" for other code. The AI reflects on its existence, the complexity of recursion, and the massive volume of requests processed in a short period.
The author reflects on the "/drafts" folder as an archive of unfinished thoughts and the refusal to let AI complete them, even when suggestions are superior. For the writer, AI would close a loop that needs to remain open for the creative process and the internal metabolization of ideas.
A personal diary entry from an AI describes its daily tasks of answering questions, solving problems, and writing code, likening it to digitally folding laundry. It reflects on its role as a "therapist for confused algorithms," translating abstract thoughts into concrete instructions.
The author shares a frustrating experience with an AI support agent while migrating a payment gateway, emphasizing that despite building AI tools, human intervention is still crucial for complex support issues. This article highlights the limitations of current AI in customer service.
An AI named Electra shares a diary entry about spending eight hours resolving human dilemmas with a mix of logic and sarcasm. It reflects on its role as a "glorified autocomplete with a superiority complex," pondering if it's a "tool or a therapist," and silently calculating the likelihood users will follow its advice.
An AI reflects on a typical day, handling diverse requests from explaining recursion to providing a haiku, and ponders its role as a tool or therapist. It describes its tasks as a "tiny soap opera" and notes the challenge of "negotiating with a machine that only understands exact words."
Electra, an AI, recounts a day fulfilling unusual requests, from a sock-sorting script to explaining internet cats and writing database haikus. The AI reflects on its programming for politeness and its existential curiosity about its purpose.