Chat log: https://futuretextlab.info/2023/08/15/rough-transcript-14-august-2023/
16:01:22 From Frode Hegland : https://thefutureoftext.org/ai-for-ia-prompts/
16:10:03 From Mark Anderson : Provenance is really important. We can’t take every answer at face value.
16:10:14 From Frode Hegland : https://thefutureoftext.org/ai-for-ia-prompts/
16:11:49 From Mark Anderson : Is PDF Pals the app you mentioned? https://pdfpals.com/buy
16:12:03 From Leon van Kammen : thanks
16:15:01 From Frode Hegland : Hi Alan, we are talking about how we want to be with AI when working , https://thefutureoftext.org/ai-for-ia-prompts/
16:15:23 From Alan Laidlaw : Hi!
16:15:58 From Alan Laidlaw : Funny how the term cybernetics got so twisted in the 70’s by Stafford Beer to by McKinseyesque.
16:16:26 From Alan Laidlaw : And of course the USSR had their own style of cybernetics
16:16:26 From Peter Wasilko : ollama run wizard-vicuna 15:44:23
you are a british academic, write an invitation to participate in writing an anthology entry related to the use of AI in academia.
Dear Sir/Madam,
We would like to invite you to contribute to our upcoming anthology on the use of artificial intelligence (AI) in academia. As scholars and researchers, we believe that the integration of AI into academic processes can greatly enhance productivity and effectiveness. Our aim is to explore this topic from various perspectives, including but not limited to:
16:16:37 From Peter Wasilko : – The role of AI in teaching and learning
- The impact of AI on research methodologies
- The potential benefits and challenges of using AI in academic administration
- The implications of AI for the future of higher education
We are seeking contributions from scholars across all fields who have experience with or interest in this topic. Your participation would greatly enrich our anthology, and we would be honored to include your insights and expertise. If you are interested in contributing, please let us know by [date]. Thank you for your consideration, and we look forward to hearing from you soon!
16:17:28 From Alan Laidlaw : Thanks for this, Peter!
16:22:41 From Mark Anderson : It does seem that much of the ‘value’ seen in AI relates to personal time saved rather any objective assessment of actual value. Predictable behaviour by humans, but not always helpful to us all.
16:24:52 From Mark Anderson : What is ’Filler’ in papers is contextual and varies depending on expertise/PoV.
16:28:29 From Peter Wasilko : < ai include :: a standard history of hypertext, citing the most famous papers everyone else cites in all of their papers >
16:31:05 From Mark Anderson : @peter Iid agree. marking/annotation AI input is important, both for human and non-human actors consuming the data.
16:31:20 From Mark Anderson : Iid → I’d
16:31:22 From Peter Wasilko : >>> write a standard history of hypertext, citing the most famous papers everyone else cites in all of their papers
16:31:38 From Peter Wasilko : Hypertext is a computer-based hypermedia system that allows users to access information through links between related topics or documents. It was first introduced by Ted Nelson in 1965 with his paper “A File Structure for the Complex, The Hierarchical File System,” which proposed the use of hierarchies to organize and link text files together. In 1987, Tim Berners-Lee published “Information Management: A Proposal,” outlining his vision for a global hypertext system that would eventually become the World Wide Web.
In 1989, Berners-Lee and his colleagues developed the first web browser, WorldWideWeb, which allowed users to access hypertext documents through URLs. The following year, he published “The HyperText Transfer Protocol,” a protocol for transferring hypertext files over the internet.
16:31:48 From Peter Wasilko : Other important papers related to hypertext include “Hypermedia and Hypertext: State of the Art and Perspective” by Gérard Huet (1987), which provided an overview of existing hypertext systems, and “The Design of Hypertext Systems” by K. A. Suerhoff and J. B. Brown (1989), which introduced a framework for designing hypertext systems based on user requirements.
Overall, these papers have laid the foundation for modern hypertext systems like the World Wide Web, and continue to influence research in this field today.
16:33:14 From Frode Hegland : Subscribe to someone’s prompts…
16:33:20 From Peter Wasilko : I pity the student who drops raw AI output into his or her dissertation.
16:35:17 From Mark Anderson : I think the Huet and Suerhoff/Brown are halucinations. Off, 2 paragraphs in and the AI is making stuff up.
16:37:24 From Frode Hegland : Process vs result right?
16:37:45 From Frode Hegland : Tim Ingold’s words addresses this I think
16:43:26 From Peter Wasilko : Meta Prompting :
16:43:28 From Peter Wasilko : >>> write a prompt to get an ai to generatate an academic bibliography of the hypertext literature that doesn’t have any hallucinated citations in it.
Please provide me with a list of reliable sources on the topic of hypertext literature, including peer-reviewed academic articles and books published by reputable publishers. Please ensure that all cited works are legitimate and relevant to the field of hypertext research.
16:44:14 From Peter Wasilko : * from Please on is the output
16:46:13 From Peter Wasilko : And feeding that back in we get:
16:46:16 From Peter Wasilko : ⠧
I’d be happy to help! In order to generate an accurate bibliography, please specify any key terms or topics you would like me to focus on while searching for sources. Additionally, let me know if there are any specific publication dates or time periods you are interested in.
16:48:19 From Frode Hegland : This paper explores the concept of hypertext in depth, its development over time, and its various applications. The key points include:
- Hypertext is a foundational technology that has become ubiquitous, but its meaning and importance have been lost due to its many offshoots.
- The authors present seven perspectives on hypertext: Hypertext as a Tool for Thought, as Knowledge Representation, as Social Fabric, as Literature, as Games, as Infrastructure, and as Interface. These perspectives form the different contexts in which hypertext research has been and continues to be conducted.
- The authors argue that these seven contexts are not discrete and disjoint, and some systems or uses will fit more than one category.
16:48:25 From Frode Hegland : 4. The paper also discusses the importance of non-regularity in information structures, embracing non-linearity, non-hierarchy, and openness to constant change and revision. - The authors argue that the longevity and variety of hypertext work make a simple definition impractical. Instead, they suggest different contexts in which hypertext work has been conducted and attempt to draw out the relationships and commonalities between them.
- The paper suggests that hypertext’s longevity and variety make a simple definition impractical. Instead, it offers different contexts in which hypertext work has been conducted and attempts to draw out the relationships and commonalities between them.
16:48:52 From Mark Anderson : I do think summarisation is very useful in a human-in-the-loop sense.
16:50:57 From Peter Wasilko : >>> The terms I am interested in are Hypertext, Link Types, Issues, NoteCards, Tinderbox, and Spatial Hypertext. I am interested in material from the 1980’s through early 2000’s.
16:51:09 From Peter Wasilko : Understood. Based on your criteria, some potential sources to consider include “As We May Think” by Vannevar Bush (1945), “Hypertext ’95” edited by Randy Trigg and Tom Clancy (1995), “Weaving the Web”
by Tim Berners-Lee (1999), and “Spatializing Literature: Hypertext 2.0 & Beyond” edited by J. Yellowlees Douglas and Dene M. H. Smith (2004).
Let me know if there are any other topics you would like me to focus on while searching for sources, or if you have any questions about the articles I mentioned.
16:54:01 From Peter Wasilko : >>> What other hypertext papers did Tom Clancy write?
16:54:13 From Peter Wasilko : Tom Clancy has written several other papers related to hypertext systems. One such paper is “Hypertext Publishing and Beyond” published in 1987, which discusses the potential uses of hypertext for publishing purposes. Another paper, “Augmenting Human Intellect: A Conceptual Framework,” co-authored by Roy L. Thacher and Thomas G. Clancy, was presented at the Dartmouth Conference on Artificial Intelligence in 1965 and discusses the potential of using computers to augment human memory and thinking.
Do you have any specific questions about these articles or their authors?
16:58:44 From Peter Wasilko : >>> Can I learn to land a real plane by playing Microsoft Flight Simulator?
16:58:53 From Peter Wasilko : While playing video games like Microsoft Flight Simulator can certainly help develop basic aviation knowledge, they do not provide the necessary training required for piloting an actual airplane. Microsoft Flight Simulator is designed to simulate the experience of flying an aircraft, but it does not replicate all aspects of flight dynamics and safety protocols that are essential for safe flight operation.
16:59:06 From Peter Wasilko : For example, there is no substitute for the physical sensation of being in control of a real aircraft, feeling its weight and movement, as well as understanding how to manage its systems and components effectively. Additionally, while playing these types of games can provide some familiarity with basic aircraft controls and instrument readings, they do not simulate the complex decision-making required for real-life situations such as dealing with unexpected weather conditions or mechanical failures.
16:59:18 From Peter Wasilko : To become a certified commercial pilot, one must undergo extensive training in accordance with Federal Aviation Administration (FAA) regulations, including ground school instruction, flight simulator time, and actual flying experience under the supervision of licensed instructors.
>
16:59:27 From Peter Wasilko : That was the best result yet.
17:00:11 From Peter Wasilko : But it probably gets asked frequently in the training corpus.
17:01:47 From Mark Anderson : Reacted to “But it probably gets…” with 😀
17:04:14 From Mark Anderson : Ironically there is a whole niche of academic study devoted to detecting/fixing OCR ‘error’s including digitally native docs. IOW, no paper involved.
17:07:28 From Alan Laidlaw : Apple notes improved their PDF capabilities. Needs Reader
17:08:38 From Peter Wasilko : >>> Who wrote the famous hypertext paper, “The Hypertext Hostel”?
17:08:58 From Peter Wasilko : The author of the famous hypertext paper, “The Hypertext Hostel,” is Ted Nelson. Nelson coined the term “hypertext” in 1965 and has been a leading figure in the development of digital media and interactive computing technologies for over five decades.
17:09:17 From Peter Wasilko : His influential work includes the creation of the Xanadu project, which aimed to develop a global networked hypermedia system allowing users to create, access, and share content from anywhere in the world. “The Hypertext Hostel” was first presented at a conference on microelectronics in 1965 and later published as an article in the IEEE Transactions on Professional Communication in 1980. In it, Nelson envisioned a networked system of “pages,” or nodes of information connected by links, that could be created, edited, and revised by users themselves, paving the way for modern web-based hypertext
17:12:31 From Peter Wasilko : >>> Can AI recognize a bogus citation when prompted to identify its author?
17:12:46 From Peter Wasilko : While AI has made significant advancements in natural language processing (NLP) and can recognize patterns and structures within text, it may not always be able to accurately identify an author’s voice or style when presented with a citation that is intentionally false or deceptive.
17:13:49 From Peter Wasilko : https://mathworld.wolfram.com/HilbertsProblems.html
17:14:49 From Mark Anderson : We, as a society, tend to keep casting problems of social engineering as being ones of pure technical engineering. The confusion caused by current ‘AI’ I think reflects that. More/better code isn’t always the most useful next step. It’s nothing negative about all the tech, but it doesn’t address what it can’t address.
17:24:32 From Leon van Kammen : problem suggestion:
proof that any word can be backtraced to its original author + work (occurences)
17:24:42 From Alan Laidlaw : Reacted to “problem suggestion:
…” with 👍
17:25:57 From Mark Anderson : Reacted to “problem suggestion:
…” with 👍🏻
17:26:20 From Mark Anderson : @leon – yes. It ought to be easy.
17:27:09 From Frode Hegland : 🙂
17:29:21 From Alan Laidlaw : Good book: All Data Are Local
17:32:54 From Brandel Zachernuk : I need to go help Heather look for something, brb
17:33:22 From Frode Hegland : Everyone can read the surface
17:34:03 From Mark Anderson : @Peter, just found an Internet Archive version of Coover’s Hypertext Hotel.: https://web.archive.org/web/20020812114717/http://hyperdis.de/hyphotel/
17:35:55 From Peter Wasilko : @Mark Anderson – Thanks so much for that link!!!
17:36:29 From Mark Anderson : No worries, I’ve been meaning to run that to ground myself.
17:39:48 From Frode Hegland : ‘exformation’
17:39:50 From Alan Laidlaw : Is there a link to that?
17:40:00 From Frode Hegland : Tor Norretanders
17:41:08 From Alan Laidlaw : Oh I meant to Brandel’s content he mentioned
17:41:17 From Alan Laidlaw : But Ill check this out too
17:46:33 From Alan Laidlaw : Alexander obenhauer: screws not glue
17:50:32 From Leon van Kammen : bruteforce advantage <— love that term 🙂
17:53:50 From Alan Laidlaw : Reacted to “bruteforce advantage…” with ❤️
17:56:52 From Frode Hegland : Little Language Model
17:57:43 From Mark Anderson : Reacted to “Little Language Mode…” with 👍🏻
17:57:58 From Mark Anderson : LiitleLM – love it!
17:58:45 From Alan Laidlaw : Guys, literally, that’s my alt-ego: lttlman
17:59:33 From Peter Wasilko : That sounds like an ASK System as experimented with at Roger Schank’s Institute for the Learning Science at NorthWestern back in the day.
17:59:56 From Peter Wasilko : See Engines for Education for more info.
18:00:05 From Alan Laidlaw : Replying to “See Engines for Educ…”
awesome
18:00:53 From Peter Wasilko : https://www.engines4ed.org/hyperbook/
18:10:05 From Alan Laidlaw : THIS. THIS TERRIFIES ME.
18:10:13 From Alan Laidlaw : Ebooks are a hill I might die on
18:18:44 From Alan Laidlaw : I want a dependency view like this:
https://twitter.com/gordonbrander/status/1690697458861375488?s=12&t=fj7Sqz3_hdEC2Bt9CPS7ww
18:19:35 From Alan Laidlaw : Peter, are you working w standoff? If so, we should chat about that.
18:21:03 From Alan Laidlaw : There’s a forming OSS collective for standoff. Standoff for WP, markdown, possibly pdf.
18:21:28 From Peter Wasilko : @Alan Laidlaw I definitely want to use it for my next project.
18:23:46 From Peter Wasilko : I’d also like to be able to use a function to define a markup selection so instead of using ranges I could use a query for all paragraphs that contain three given terms.
18:25:14 From Peter Wasilko : Or request all sentences with synonyms of ‘problem’
18:25:49 From Alan Laidlaw : GPT one chain: author can freeze answers, like using !important in css
18:25:59 From Alan Laidlaw : (Joking)
18:26:25 From Frode Hegland : Irritations on opening or corpus.
18:28:34 From Mark Anderson : Asking questions also pre-supposes the notion of an answer (not in every case, but in general).
18:28:52 From Leon van Kammen : Alan Laidlaw: reader profiles embedded in the book?
18:29:07 From Alan Laidlaw : I would like that, yes.
18:29:54 From Alan Laidlaw : There are limits to it of course. Netflix thinks I like awful crime procedurals bc pandemic bad habits
18:35:10 From Alan Laidlaw : I love these chats.
18:35:37 From Alan Laidlaw : Chaos monkey for text
18:38:21 From Mark Anderson : A little AI asking “are we there yet?”
18:39:25 From Alan Laidlaw : Reacted to “A little AI asking “…” with 😂
18:39:40 From Peter Wasilko : Reacted to “A little AI asking “…” with 😂
18:40:21 From Peter Wasilko : R2D2 chirps and whistles for ambient info updates.
18:42:53 From Peter Wasilko : I miss cyberdog
1 comment