|Tags||5Cs, 5Cs Learning and Knowledge Management, AI, Category Theory, Digital Reproduction, Fiction, GPT, Generative Pre-trained Transformer, Holon, Kathryn Cramer, Narrative Coherence, Personality, Probability, Tinderbox, Tokens, Writing|
|Video URL||Tinderbox Meetup - Saturday, June 24, 2023: A Discussion with Kathryn Cramer - YouTube|
|Chat||chat.txt (10.0 KB)|
In this Tinderbox meetup, we have a glorious ride, al be it brief, with Kathryn Cramer on the evolution of AI, GPTs, her research on “military privatization,” GPT models and how to create “scalable and reliable foresight models”, “understanding personality in the age of digital reproduction”, and the impact of all these topics on thought, society, and more. And, of course, the role that Tinderbox may play as a UI to AI/GPT.
Kathryn shares with us. Starting with the end, Kathryn in looking to see how all this will influence our efforts on learning and making knowledge, i.e., “What comes after the book?” Perhaps “a memory palace solar that can’t see the edges.” We discussed:
- GPT, “generative pre-trained transformer”
- The root of GPT and the nature of a token, i.e., a single letter, a word, a concept, or a person’s name and how these “token” are influencing what is thought to be “known” today.
- Each word is a probability
- Temperature is the relative probability of the next word, 0 temperature is deterministic, while as some temperature allows for probability
- Personality is really represented in bias
- The probabilities are making weights between nodes, probabilistic trees
- Tokens can end up having multiple meanings
- Holon, the most basic unit is a token “I” and “T” are token and “It” code be a token.
- In Tinderbox, a “node” is a note; in AI it is a “token.”
- Kathryn “We are undergoing a transition where our “private” personalities can be commodified without our knowledge, understanding, permission, or compensation. We all could become Henrietta Lax.”
- Kathryn “The ghost of Donald Trump is influencing open AI given the number of nodes are related to him and the tokenizing of tokens. “
- Kathryn “We are not all equal in the eyes of AI; we need to understand personality.”
- Kathryn “Commercial fiction will go the way of poetry in 10 years, as AI will crack the commercial code for the ‘midlist,’ what will actually sell.”
- Kathryn “GPT4 can help you think better (like a calculator or slide rule).”
- Kathryn “We are becoming intelligent in different ways.”
- Kathryn “Internet roadkill: you have to check if citations exist.”
- Kathryn “What is GPT good for? It is a process tool, not an end-product tool. It is a labor-saving device. It is good to tell you what conventional wisdom is? You can see what is obvious.”
- Mark Bernstein “GPTs are great at getting things done but we don’t really understand what they’re doing.
- Art Currim “With great power comes great stupidity.”
- Detlef Beyer “Tinderbox is a great interface to AI… good way to find your way how to use AI”
- Narrative Coherence
- Category Theory
- GPT authors creating are manipulating the system, how does the “average” person survive
- Is everything going to be suffused with AI, how can we retain human purity?
- In GPT4 use “tabulate” to tabulate column of keywords/answers
- Sudowrite, a GTP3 for generating fiction
- The Cuckoo’s Egg: Tracking a Spy Through the Maze of Computer Espionage
- Wolfram Mathematica
- List of eponyms, An eponym is a person (real or fictitious) from whom something is said to take its name. The word is back-formed from “eponymous”, from the Greek “eponymos” meaning “giving name”.
- Holon, concept of token, is something that is simultaneously a whole in and of itself, as well as a part of a larger whole.
- Common Crawl
- Machine Hallucination
- Token Mapping – this is the process of assigning the created token value to its original value. To enable permitted look-ups of the original value using the token as the index, a secure cross-reference database must be constructed.
- ‘Too many notes, dear Mozart, too many notes’ is what Emperor Joseph II supposedly said after the first performance of the Entfuhrung aus dem Serail in Vienna’s old Burgtheater. Mozart’s reply was: ‘Just as many as necessary, Your Majesty.
- CLIP: The Most Influential AI Model From OpenAI — And How To Use It* 701 Translator
- Georgetown–IBM experiment
- ALPAC (Automatic Language Processing Advisory Committee) *And the all important George Miller “Ambiguous Words”
- Immortal Life of Henrietta Lacks
- Kathryn’s Thesis: A Thousand Future: A Search for Scenario Space
- What Is ChatGPT Doing … and Why Does It Work?
- Kathryn’s page at The Encyclopedia Of Science Fiction
- Project Athena, weave computing resources into the educational process, challenges of collaboration in an academic institution
Please help with the development of future sessions by answering the three questions below.
- What were your top 2~3 key takeaways from this lesson?
- What do you want to learn next? Learn more about?
- What exercises would help reinforce your learning?