Close reading, an ancient analytical technique for deep textual analysis, can be enhanced using Large Language Models. LLMs enable readers to ask clarifying questions, explore tangential topics, and personalize content while reading. The SolveIt platform combines elements of ChatGPT, Jupyter, and other tools to facilitate this process through incremental steps and immediate feedback. Demonstrations show reading a book and analyzing academic papers with LLMs, using techniques like context management, spaced repetition with Anki flashcards, and generating chapter summaries. While the tooling is still evolving and requires some technical setup, grounding and web search capabilities help mitigate hallucination concerns.
Table of contents
The Ancient Art of Close ReadingTwo Examples of Reading with an LLMBenefits of Close Reading with an LLMThe SolveIt ProcessObstacles to Reading with an LLMA Work in ProgressSort: