| layout | title | permalink |
|---|---|---|
page |
Discussions |
/Discussions/ |
The discussions we are refering here to are basically our own little discussions about the LARGER Big Discussions of our day.
Through Annotify.APP we are engaged daily in the process of thinkering with and discussing the development, extension, customizations and refactoring of LLM models. We anticipate and hope that someone will, or already has, developed a social media appliance or plug-in for annotating and discussing lists of machine intelligence tools and models.
We develop, curate, customize, extend lists OR listify collections of AWESOME repositories, papers, papers with code, models, book abstracts and books [those that are part of the public domain as well as those that widely available through Scribd] and collections of published work and awesome code repositories.
Listification is about learning, categorizing, processing, grokking ... it's our process for exploring connections and developing HYBRID intelligence. Listifying, just like LLM AI assistants, is a fundamentally recursive activity.
Through Annotify we are engaged in discussing the development, extension, customizations and refactoring of LLM models. We are also discussing the development of a social media appliance or plug-in for annotating and discussing lists of machine intelligence tools and models.
In effect our simple little lists, perhaps something starting with a *.csv or *.json file exported from Excel or some data API, become executable ... lists with added annotations are presented stylistically in a human readable knowledge graph analtyic format ... the listified knowledge graphs have limited immediate informative value, but it is discussions and annotations to the graph which can used as an API input for generating the next customized list or customized LLM model.
Adding intelligence through curation by collecting, annotating and/or indexing human-readable lists of machine intelligence tools ... but the user-interface is about the general nature of human intelligence and cognition ... sure, we want to lists to computer-executable, but we also want to lists to be human-readable and human-annotatable.
We are especially interested in lists of pre-print archives and repositories of code, data, papers with code and machine learning models ... these machine learning models will typicably be from Python libraries excuted in Jupyter notebook or something that behaves similarly in an environment like CoLab or perhaps in a more configurable environment such as Google CloudVertexAI].
The process of Listification is a SOCIAL process, but this form of social interaction is like an intense graduate seminar ... it's NOT for passive observers who watch teevee or Youtube videos ... it's an ACTIVE discussion which is ONLY for those who are willing to read and write and argue and defend their opinions without polluting the discustion with WEAK, ad hominem attacks.
Reading something does not really give on much insight. The insight comes later after mulling over what was read, sharing an opinion on the reading and, ideally, getting into an argument and defending one's opinion.
Thus our ANNOTIFY.app will necessarily be a social media appliance or plug-in ... without the ability to at least share an annotation, we are stuck with only re-reading our own annotations ... which is no small thing, but the important part comes from writing something that one knows others might argue over.
EVEN when it's private your annotations are written to be read by a later version of you, the version of you who is hopefully smarter and more cultivated, the version of you that has cooled off emotionally, the version of you that sees the world differently than the you of right now. The point here is that annotations are necessarily social ... even when they are private or restricted to small audience of very similar people.