I've been using language models to write things since I got into the OpenAI API Beta in March, 2021. I use them to analyze and understand documents (pages/PDFs/

I've been using language models to write things since I got into the OpenAI API Beta in March, 2021. I use them to analyze and understand documents (pages/PDFs/raw text). Several projects I've created from scratch embed and index document collections in a variety of ways to increase the relevance for prompt building. Before I understood embeddings, and how to use them, I build a search index using Solr for the document texts to use for assembly of a prompt. I use language models to write code that I would normally search for on Google (and I used it to parse and write Solr queries when my parsers failed) and I use them to write blog posts about the topics I'm thinking about. I sometime use them to write plans or emails. Using language models makes me faster at what I do and frequently it seems like what I'm doing with them is increasing the quality of my output. Blog posts that I would antagonize about publishing are cleaned up and made ready for submission in a matter of an hour or so. Before, it took me several days to get happy with the results and even then I would continue to refine it after publishing. Language models don't come up with the ideas for me, however. All the things I use that it produces was heavily prompted and discussed with myself. At all times I'm driving the conversation and what is being communicated. It's how the information is communicated that is being augmented now. Clarifying my fragmented thoughts on paper becomes far easier to transform into something another can consume easily. As a related example, over the past 10 years or so, I've been obsessed with thinking about consciousness (or thinking about thinking). After talking to many about this, and reading as much as I can about it that has been written, I've come to the conclusion that many of us think differently. Coming up with a plausible argument that explains this thought in a clearer manner would be one example of how I use language models. After a long interaction with ChatGPT4 on this subject, which included Buddhist thought, Aphantasia, Bicameral Mind (itself controversial) and more, I issued the following request: "Summarize the argument people think differently with a comparison to Aphants not being able to use mental or auditory information recall." Here's the answer it gave: > The comparison implies that just as individuals with aphantasia lack visual mental imagery, others may have variations in their cognitive processes that affect their ability to use or manipulate mental or auditory information. These differences in information processing can result in diverse thinking styles, problem-solving approaches, and ways of perceiving and understanding the world. If we think about the lack of something as an edge case of what is known, then maybe we can "outline" the idea using the "edges" of the concept... If you have Aphantasia writing is very much a way of thinking about thinking about things. Like CSSer, it gives me a way to do critical thinking about a subject, without relying on visual or audio information recall. Language models enhance my natural ability to "only think" by allowing me to write and read more about a subject that interests me. Using critical thinking, I try to understand and verify all the text I'm reading makes sense, which include what I've written as well as what the model has generated. It's only a "stochastic parrot" if you fail to think critically about the output before sharing it with others, would be one way to put it. In Bhikkhu Bodhi's functional model of the mind [1], the six sense bases are building blocks for a unified consciousness. They are often referred to as the "Six Sense Spheres" or "Six Sense Doors." These six sense bases are: Eye, Ear, Nose, Tongue, Body and Mind. Body does things like feel heat, or hunger. Mind is the mental sense base, which encompasses the cognitive and perceptive functions of the mind, enables the perception of mental objects, thoughts, emotions, and ideas. The computed language model furthers this: > In some individuals, the mind has the capacity to produce internal imagery or auditory experiences. This ability allows them to generate mental representations or simulations of visual or auditory stimuli without the presence of external sensory input. These internal experiences can vary in vividness, clarity, and detail from person to person. > Internal imagery refers to the mental generation of visual images within the mind's eye. Individuals with this ability can conjure up mental pictures, scenes, or objects, similar to "seeing" something in their mind without the actual physical stimuli being present. > Similarly, internal audio refers to the experience of hearing sounds, voices, or music within the mind. People who possess this ability can mentally replay or recall auditory stimuli, creating an internal auditory experience that resembles hearing without external sound sources. I am a person with a condition labeled "Aphantasia". For me, writing is a means of thinking and that thinking is now being augmented by language models. If I choose to approach the process with a critical view of the content, then it is most certainly not a waste of my time. How others interact and think with it is going to vary widely however, so any statement I make about how language models affect others will be irrational and wrong, in many cases. And if what I said doesn't make sense to you, maybe this will: > Ultimately, the availability of language models serves as a reminder that every individual possesses a unique view of the world, and their thoughts and expressions may resonate differently with others. Embracing this diversity of perspectives and engaging in respectful and open dialogue can lead to a richer understanding and appreciation of the multifaceted nature of human cognition and communication. [1] https://www.researchgate.net/figure/this-figure-shows-the-basic-functional-model-in-this-model-as-compared-to-Bhikkhu_fig2_315940922 https://www.researchgate.net/figure/this-figure-shows-the-ba...