Transcript: I'm just trying to write a note for something I'm thinking about, which is it would be really interesting to be able to have the handy program be able to take snippets of context at a time. So I could just copy one file, copy something from another file in VS Code, and then run inference directly on those two pieces of code. But treating them separately, of course. Yeah, I'm just kind of wondering how I can broadly pick at things on my screen and be like, hey, this thing, that thing, and this thing, all do inference at the same time, please, with them all in the context window. So basically letting me manage the context window instead of it just kind of arbitrarily going and figuring out what it thinks the context window should be.
83.42% similar
The author is contemplating a project that involves collecting artifacts and placing them into a hypermedia context on a web page, allowing for complex interactions and transformations of the data. This project is envisioned to have a similar interface to a zine but with added complexity, akin to an application called Burrito. The goal is to enable querying, selection, and relationship-building between pieces of information, with the potential for both individual and combined transformations. The author is particularly interested in manipulating data through various operations and sees potential for integrating this interface into AR or VR environments due to their visual and spatial preferences. The core concept is a "canvas methodology" that can be applied across different digital creations, differing from the simpler drag-and-drop zine format. The author also considers the possibility of using large language models to build the application autonomously by providing detailed, context-rich information over time.
In envisioning an ideal way to integrate new log entries, the goal is to place each entry within the larger context of the whole, which may be an iterative process to determine that context. The author contemplates whether incorporating various data sources into a language model like GPT-4 could help it understand the overarching themes of communications, such as text messages. They propose an experimental approach by loading as much context as possible into the model whenever a new input is received, maximizing the token limit to allow the model to contextualize new information based on previous entries. This method, which involves brute forcing context into the AI's understanding, could potentially be a valuable asynchronous step in refining the pipeline for more nuanced contextual analysis.
The text discusses the concept of building context within a project or a system, starting from a blank state. Two methods are mentioned for establishing this context: importing existing elements from global or other pre-defined contexts, and directly dragging and dropping files. The purpose is to organize information into meaningful units that can be easily manipulated. Additionally, it highlights the importance of making data sources explorable, such as tables and SQLite databases, which is a separate but equally crucial aspect.
The author is contemplating solutions to the "agent problem" they are facing, expressing frustration over the repetitive task of rewriting prompts. They consider whether using a spreadsheet might streamline the process, hypothesizing that it could be a better framework for organizing tasks without constant coding. The person believes that an approach involving spreadsheets, possibly in combination with tools like Microsoft's Autogen, could reduce costs and increase efficiency by avoiding the need to insert excessive code into context windows. Although their current results are almost satisfactory, they are determined to find an architecture that streamlines the process while achieving the desired outcomes without overloading the system prompts.
The author plans to create a meta projects page to organize and categorize audio notes into four main projects: Handy (for single-handed computer use), project tracking, a zine creation tool, and Burrito, which serves as the overarching framework for all the projects. Coffee is aiding the progress. Tomorrow, the focus is on advancing the zine project to a usable state, including adding text, images, and possibly integrating the Glyph API. The author also envisions adding multi-page functionality and allowing users to create and link pages, creating a nested context system. The ultimate goal is to provide a URL for others to explore and use the zine tool.
79.50% similar
The speaker is reflecting on their experience with making audio burrito posts, noting that it often requires multiple attempts to get into the correct mindset—similar to drafting written posts. They're grappling with the challenge of monologuing without a clear understanding of the audience, as they are aware that at least John and CJ will hear it, but uncertainty about the wider audience affects their ability to communicate effectively. This creates a 'contextual membrane shakiness' as the speaker finds the lack of audience boundaries difficult to navigate, which they recognize may vary among different people. The speaker concludes by deciding to end the current note and start a new one.
79.19% similar
The author contemplates the process of converting an audio note into a transcript, then summarizing it on their "burrito" page. They express a desire to adjust the summarization voice to better represent themselves on the page. Recognizing that this feature may not have widespread appeal, the author nonetheless sees value in providing users with controls to personalize their "burrito." The concept of allowing users to fine-tune their experience is seen as an intriguing possibility.
76.95% similar
I've always been drawn to the peculiar and unexplored, which makes me wonder if I can pepper my writing with a bit of the offbeat—things that don't quite fit the mold. Question is, can I make it work? Ditching the third-person narrative and opting for a chat with you in the first person could make my stories feel more intimate, more like we're in this together. And hey, isn't that what storytelling's all about? Let's find out.