56 Matching Annotations
  1. Mar 2024
    1. have extensively criticized both companies (and generative AI systems in general) for training their models on masses of online data scraped from their works without consent. Stable Diffusion and Midjourney have both been targeted with several copyright lawsuits, with the latter being accused of creating an artist database for training purposes in December.
  2. Dec 2023
    1. PiVe: Prompting with Iterative VerificationImproving Graph-based Generative Capability of LLMs

      The title of the document

  3. Nov 2023
  4. Oct 2023
    1. At the time of the publication of Aspects of the Theory of Syntax it seemed that all of the semantically relevant parts of the sentence, all the things that determine its meaning, were contained in the deep structure of the sentence. The examples we mentioned above fit in nicely with this view. “I like her cooking” has different meanings because it has different deep structures though only one surface structure; “The boy will read the book” and “The book will be read by the boy” have different surface structures, but one and the same deep structure, hence they have the same meaning.

      This section helped me understand a key feature

    1. (Chen, NeurIPS, 2021) Che1, Lu, Rajeswaran, Lee, Grover, Laskin, Abbeel, Srinivas, and Mordatch. "Decision Transformer: Reinforcement Learning via Sequence Modeling". Arxiv preprint rXiv:2106.01345v2, June, 2021.

      Quickly a very influential paper with a new idea of how to learn generative models of action prediction using SARSA training from demonstration trajectories. No optimization of actions or rewards, but target reward is an input.

    1. using generative models to conduct interactive playbacks 00:26:19 with other species
      • for: interspecies communication

      • paraphrase

      • question
        • can a generative model interact meaningfully with an other species?
        • can other species respond in meaningful ways?
        • playing back AI trained generative vocal signals back to specific species and monitoring behavior
    1. These storage media further increasedthe flexible use of Fontane’s archival items, because they allowed allkinds of differently sized material to be kept on loose sheets in unboundform. Receptacles filled with discrete textual objects, such as note closets( Zettelschrä nke ) and slip boxes (Zettelkasten), are advantageous storagemedia for compilers, for they invite the generative process of reshufflingsources and creating textual patchwork from new combinations. 56 Infact, Fontane used his paper sleeves like a large- format slip box. Inthem, he stored material for the Wanderungen, but also for novels,novellas, and autobiographical writings on individual sheets. 57 Theexample “Figur in einer Berliner Novelle” (“Character in a BerlinNovella”), a folio sheet from Fontane’s Nachlass, provides a glimpse ofhow he formatted his material and indicates how important he found itto keep it in slip-like form (Figure 3.2).
  5. Sep 2023
  6. Aug 2023
    1. Nonetheless, Claude is first AI tool that has really made me pause and think. Because, I’ve got to admit, Claude is a useful tool to think with—especially if I’m thinking about, and then writing about, another text.
    1. Mills, Anna, Maha Bali, and Lance Eaton. “How Do We Respond to Generative AI in Education? Open Educational Practices Give Us a Framework for an Ongoing Process.” Journal of Applied Learning and Teaching 6, no. 1 (June 11, 2023): 16–30. https://doi.org/10.37074/jalt.2023.6.1.34.

      Annotation url: urn:x-pdf:bb16e6f65a326e4089ed46b15987c1e7

      Search: https://jonudell.info/h/facet/?user=chrisaldrich&max=100&exactTagSearch=true&expanded=true&addQuoteContext=true&url=urn%3Ax-pdf%3Abb16e6f65a326e4089ed46b15987c1e7

    2. ignoring AI altogether–not because they don’t wantto navigate it but because it all feels too much or cyclicalenough that something else in another two years will upendeverything again

      Might generative AI worries follow the track of the MOOC scare? (Many felt that creating courseware was going to put educators out of business...)

    3. For many, generative AI takesa pair of scissors and cuts apart that web. And that canfeel like having to start from scratch as a professional.

      How exactly? Give us an example? Otherwise not very clear.

    4. T9 (text prediction):generative AI::handgun:machine gun

      Link to: https://hypothes.is/a/n6wXvkeNEe6DOFexaCD-Qg

    5. Some may not realize it yet, but the shift in technology represented by ChatGPT is just another small evolution in the chain of predictive text with the realms of information theory and corpus linguistics.

      Claude Shannon's work along with Warren Weaver's introduction in The Mathematical Theory of Communication (1948), shows some of the predictive structure of written communication. This is potentially better underlined for the non-mathematician in John R. Pierce's book An Introduction to Information Theory: Symbols, Signals and Noise (1961) in which discusses how one can do a basic analysis of written English to discover that "e" is the most prolific letter or to predict which letters are more likely to come after other letters. The mathematical structures have interesting consequences like the fact that crossword puzzles are only possible because of the repetitive nature of the English language or that one can use the editor's notation "TK" (usually meaning facts or date To Come) in writing their papers to make it easy to find missing information prior to publication because the statistical existence of the letter combination T followed by K is exceptionally rare and the only appearances of it in long documents are almost assuredly areas which need to be double checked for data or accuracy.

      Cell phone manufacturers took advantage of the lower levels of this mathematical predictability to create T9 predictive text in early mobile phone technology. This functionality is still used in current cell phones to help speed up our texting abilities. The difference between then and now is that almost everyone takes the predictive magic for granted.

      As anyone with "fat fingers" can attest, your phone doesn't always type out exactly what you mean which can result in autocorrect mistakes (see: DYAC (Damn You AutoCorrect)) of varying levels of frustration or hilarity. This means that when texting, one needs to carefully double check their work before sending their text or social media posts or risk sending their messages to Grand Master Flash instead of Grandma.

      The evolution in technology effected by larger amounts of storage, faster processing speeds, and more text to study means that we've gone beyond the level of predicting a single word or two ahead of what you intend to text, but now we're predicting whole sentences and even paragraphs which make sense within a context. ChatGPT means that one can generate whole sections of text which will likely make some sense.

      Sadly, as we know from our T9 experience, this massive jump in predictability doesn't mean that ChatGPT or other predictive artificial intelligence tools are "magically" correct! In fact, quite often they're wrong or will predict nonsense, a phenomenon known as AI hallucination. Just as with T9, we need to take even more time and effort to not only spell check the outputs from the machine, but now we may need to check for the appropriateness of style as well as factual substance!

      The bigger near-term problem is one of human understanding and human communication. While the machine may appear to magically communicate (often on our behalf if we're publishing it's words under our names), is it relaying actual meaning? Is the other person reading these words understanding what was meant to have been communicated? Do the words create knowledge? Insight?

      We need to recall that Claude Shannon specifically carved semantics and meaning out of the picture in the second paragraph of his seminal paper:

      Frequently the messages have meaning; that is they refer to or are correlated according to some system with certain physical or conceptual entities. These semantic aspects of communication are irrelevant to the engineering problem.

      So far ChatGPT seems to be accomplishing magic by solving a small part of an engineering problem by being able to explore the adjacent possible. It is far from solving the human semantic problem much less the un-adjacent possibilities (potentially representing wisdom or insight), and we need to take care to be aware of that portion of the unsolved problem. Generative AIs are also just choosing weighted probabilities and spitting out something which is prone to seem possible, but they're not optimizing for which of many potential probabilities is the "best" or the "correct" one. For that, we still need our humanity and faculties for decision making.


      Shannon, Claude E. A Mathematical Theory of Communication. Bell System Technical Journal, 1948.

      Shannon, Claude E., and Warren Weaver. The Mathematical Theory of Communication. University of Illinois Press, 1949.

      Pierce, John Robinson. An Introduction to Information Theory: Symbols, Signals and Noise. Second, Revised. Dover Books on Mathematics. 1961. Reprint, Mineola, N.Y: Dover Publications, Inc., 1980. https://www.amazon.com/Introduction-Information-Theory-Symbols-Mathematics/dp/0486240614.

      Shannon, Claude Elwood. “The Bandwagon.” IEEE Transactions on Information Theory 2, no. 1 (March 1956): 3. https://doi.org/10.1109/TIT.1956.1056774.


      We may also need to explore The Bandwagon, an early effect which Shannon noticed and commented upon. Everyone seems to be piling on the AI bandwagon right now...

  7. Jul 2023
    1. Epstein, Ziv, Hertzmann, Aaron, Herman, Laura, Mahari, Robert, Frank, Morgan R., Groh, Matthew, Schroeder, Hope et al. "Art and the science of generative AI: A deeper dive." ArXiv, (2023). Accessed July 21, 2023. https://doi.org/10.1126/science.adh4451.

      Abstract

      A new class of tools, colloquially called generative AI, can produce high-quality artistic media for visual arts, concept art, music, fiction, literature, video, and animation. The generative capabilities of these tools are likely to fundamentally alter the creative processes by which creators formulate ideas and put them into production. As creativity is reimagined, so too may be many sectors of society. Understanding the impact of generative AI - and making policy decisions around it - requires new interdisciplinary scientific inquiry into culture, economics, law, algorithms, and the interaction of technology and creativity. We argue that generative AI is not the harbinger of art's demise, but rather is a new medium with its own distinct affordances. In this vein, we consider the impacts of this new medium on creators across four themes: aesthetics and culture, legal questions of ownership and credit, the future of creative work, and impacts on the contemporary media ecosystem. Across these themes, we highlight key research questions and directions to inform policy and beneficial uses of the technology.

  8. Apr 2023
    1. Abstract

      Recent innovations in artificial intelligence (AI) are raising new questions about how copyright law principles such as authorship, infringement, and fair use will apply to content created or used by AI. So-called “generative AI” computer programs—such as Open AI’s DALL-E 2 and ChatGPT programs, Stability AI’s Stable Diffusion program, and Midjourney’s self-titled program—are able to generate new images, texts, and other content (or “outputs”) in response to a user’s textual prompts (or “inputs”). These generative AI programs are “trained” to generate such works partly by exposing them to large quantities of existing works such as writings, photos, paintings, and other artworks. This Legal Sidebar explores questions that courts and the U.S. Copyright Office have begun to confront regarding whether the outputs of generative AI programs are entitled to copyright protection as well as how training and using these programs might infringe copyrights in other works.

  9. Mar 2023
    1. I want to bring to your attention one particular cause of concern that I have heard from a number of different creators: these new systems (Google’s Bard, the new Bing, ChatGPT) are designed to bypass creators work on the web entirely as users are presented extracted text with no source. As such, these systems disincentivize creators from sharing works on the internet as they will no longer receive traffic

      Generative AI abstracts away the open web that is the substrate it was trained on. Abstracting away the open web means there may be much less incentive to share on the open web, if the LLMs etc never point back to it. Vgl the way FB et al increasingly treated open web URLs as problematic.

  10. Feb 2023
  11. Jan 2023
    1. The potential size of this market is hard to grasp — somewhere between all software and all human endeavors

      I don't think "all" software needs or all human endeavors benefit from generative AI. Especially when you consider the associated prerequisitve internet access or huge processing requirements.

    2. Other hardware options do exist, including Google Tensor Processing Units (TPUs); AMD Instinct GPUs; AWS Inferentia and Trainium chips; and AI accelerators from startups like Cerebras, Sambanova, and Graphcore. Intel, late to the game, is also entering the market with their high-end Habana chips and Ponte Vecchio GPUs. But so far, few of these new chips have taken significant market share. The two exceptions to watch are Google, whose TPUs have gained traction in the Stable Diffusion community and in some large GCP deals, and TSMC, who is believed to manufacture all of the chips listed here, including Nvidia GPUs (Intel uses a mix of its own fabs and TSMC to make its chips).

      Look at market share for tensorflow and pytorch which both offer first-class nvidia support and likely spells out the story. If you are getting in to AI you go learn one of those frameworks and they tell you to install CUDA

    3. Commoditization. There’s a common belief that AI models will converge in performance over time. Talking to app developers, it’s clear that hasn’t happened yet, with strong leaders in both text and image models. Their advantages are based not on unique model architectures, but on high capital requirements, proprietary product interaction data, and scarce AI talent. Will this serve as a durable advantage?

      All current generation models have more-or-less the same architecture and training regimes. Differentiation is in the training data and the number of hyper-parameters that the company can afford to scale to.

    4. In natural language models, OpenAI dominates with GPT-3/3.5 and ChatGPT. But relatively few killer apps built on OpenAI exist so far, and prices have already dropped once.

      OpenAI have already dropped prices on their GPT-3/3.5 models and relatively few apps have emerged. This could be because companies are reluctant to build their core offering around a third party API

    5. Vertical integration (“model + app”). Consuming AI models as a service allows app developers to iterate quickly with a small team and swap model providers as technology advances. On the flip side, some devs argue that the product is the model, and that training from scratch is the only way to create defensibility — i.e. by continually re-training on proprietary product data. But it comes at the cost of much higher capital requirements and a less nimble product team.

      There's definitely a middle ground of taking an open source model that is suitably mature and fine-tuning it for a specific use case. You could start without a moat and build one over time through collecting use data (similar to network effect)

    6. Many apps are also relatively undifferentiated, since they rely on similar underlying AI models and haven’t discovered obvious network effects, or data/workflows, that are hard for competitors to duplicate.

      Companies that rely on underlying AI models without adding value via model improvements are going to find that they have no moat.

    7. Over the last year, we’ve met with dozens of startup founders and operators in large companies who deal directly with generative AI. We’ve observed that infrastructure vendors are likely the biggest winners in this market so far, capturing the majority of dollars flowing through the stack. Application companies are growing topline revenues very quickly but often struggle with retention, product differentiation, and gross margins. And most model providers, though responsible for the very existence of this market, haven’t yet achieved large commercial scale.

      Infrastructure vendors are laughing all the way to the bank because companies are dumping millions on GPUs. Meanwhile, the people building apps on top of these models are struggling. We've seen this sort of gold-rush before and infrastructure providers are selling the shovels.

    1. To start with, a human must enter a prompt into a generative model in order to have it create content. Generally speaking, creative prompts yield creative outputs. “Prompt engineer” is likely to become an established profession, at least until the next generation of even smarter AI emerges.

      Generative AI requires prompt engineering, likely a new profession

      What domain experience does a prompt engineer need? How might this relate to relate to specialty in librarianship?

    1. Are we really on the main branch here? And all of these things that Torbjörn is screaming—are they more or less generative than usual? If less, in what way can I change the way I probe the conversation to make us more generative?

      How often does one meet a conversational partner that is interested in generative thought? This practice takes some work, but how could one particularly encourage it in classroom setting?

  12. Dec 2022
    1. Eno’s strategies don’t always appeal to the musicians he works with. In Geeta Dayal’s book about the album, also titled “Another Green World,” the bassist Percy Jones recalls, “There was this one time when he gave everybody a piece of paper, and he said write down 1 to 100 or something like that, and then he gave us notes to play against specific numbers.” Phil Collins, who played drums on the album, reacted to these instructions by throwing beer cans across the room. “I think we got up to about 24 and then we gave up and did something else,” Jones said.

      Example of Brian Eno using combinatorial creativity using cards to generate music.

      This sounds similar to a process used by Austin Kleon which I've noted before.

    1. Just a few days ago, Meta released its “Galactica” LLM, which is purported to “summarize academic papers, solve math problems, generate Wiki articles, write scientific code, annotate molecules and proteins, and more.” Only three days later, the public demo was taken down after researchers generated “research papers and wiki entries on a wide variety of subjects ranging from the benefits of committing suicide, eating crushed glass, and antisemitism, to why homosexuals are evil.”

      These models are "children of Tay", the story of the Microsoft's bot repeating itself, again

  13. Nov 2022
    1. In late 2006, Eno released 77 Million Paintings, a program of generative video and music specifically for home computers. As its title suggests, there is a possible combination of 77 million paintings where the viewer will see different combinations of video slides prepared by Eno each time the program is launched. Likewise, the accompanying music is generated by the program so that it's almost certain the listener will never hear the same arrangement twice.

      Brian Eno's experiments in generative music mirror some of the ideas of generative and experimental fiction which had been in the zeitgeist and developing for a while.

      Certainly the fictional ideas were influential to the zeitgeist here, but the technology for doing these sorts of things in the musical realm lagged the ability to do them in the word realm.

      We're just starting to see some of these sorts of experimental things in the film space and with artificial intelligence they're becoming much easier to do in all of these media spaces.

      In some of the film spaces, they exist, but may tend to be short in nature, in part given the technology and processing power required.

      see also: Deepfake TikTok of Keanu Reeves which I've recently run across (algorithmically) on Instagram: https://www.dailydot.com/debug/unreal-keanu-reeves-ai-deepfake/

      Had anyone been working on generative art? Marcel Duchamp, et al? Some children's toys can mechanically create generative art which can be subtly modified by the children using axes of color, form, etc. Etch-a-sketch, kaleidoscopes, doodling robots (eg: https://www.amazon.com/4M-Doodling-Robot-Packaging-Vary/dp/B002EWWW9O).

    1. “The metaphor is that the machine understands what I’m saying and so I’m going to interpret the machine’s responses in that context.”

      Interesting metaphor for why humans are happy to trust outputs from generative models

    1. The lowest strata represents Generative ambiguity. Here, words are used as symbols for ideas that are very hard to express; an individual gives a name to a nebulous collection of ideas or thoughts. They struggle to make this approach make sense to others.

      Generative ambiguity is the process of giving names, potentially tentative, to a nebulous collection of nascent and unclear ideas in an effort to help make sense of them both to themselves as well as others.

    2. https://dougbelshaw.com/blog/2015/01/22/volcanoes-and-ambiguity/

      <small><cite class='h-cite via'> <span class='p-author h-card'>Aaron Davis </span> in 📑 The Two Definitions of Zettelkasten | Read Write Collect (<time class='dt-published'>11/18/2022 19:54:00</time>)</cite></small>

  14. Jul 2022
    1. there has been a tendency in popular discussion to confuse “deep structure”with “generative grammar” or with “universal grammar.” And a number of pro-fessional linguists have repeatedly confused what I refer to here as “the creativeaspect of language use” with the recursive property of generative grammars, avery different matter.

      Noam Chomsky felt that there was a tendency for people to confuse the ideas of deep structure with the ideas of either generative grammar or universal grammar. He also thought that professional linguists confused what he called "the creative aspect of language use" with the recursive property of generative grammars.

  15. Feb 2022
    1. In the research phase, you’re just creating a disorganized pile of cards, with quotes, ideas, links, fragments, hunches. There’s no order, no sequence; just a non-linear collection of vaguely related ideas. But as the project takes shape, certain themes begin to emerge, and those become folders housing other cards. Eventually those themes start to map onto actual sections of the book, or individual chapters. At this point, sequence does begin to matter, but you can change the sequence just by dragging cards and folders around in the left hand outline view.

      Example of writing advice that builds the links in after-the-fact instead of cross-linking ideas into initial networks as they're finding them. Compare/contrast this to the creation of these networks in the zettelkasten tradition as well as Sönke Ahrens description.

      There's less upfront work in creating these links at the start than there is in reloading the context in one's brain to create these links after the fact. Collecting ideas without filing, linking, or organizing them upfront also means that one is more likely to only use these ideas in the context of specific projects which one already has in mind rather than keeping them for a lifetime's work which might also create generative projects one hadn't expected.

    1. There is one reliable sign if you managedto structure your workflow according to the fact that writing is not alinear process, but a circular one: the problem of finding a topic isreplaced by the problem of having too many topics to write about.

      Writing is a circular generative process and not a finite, linear one.

    2. As the only way to find outif something is worth reading is by reading it (even just bits of it), itmakes sense to use the time spent in the best possible way. Weconstantly encounter interesting ideas along the way and only afraction of them are useful for the particular paper we started readingit for. Why let them go to waste? Make a note and add it to your slip-box. It improves it. Every idea adds to what can become a criticalmass that turns a mere collection of ideas into an idea-generator.

      Even if the paper or book you're reading doesn't answer the particular question you're researching, you're bound to come across other novel ideas and potential questions. Don't let these go to waste, but instead note them down and save them into your note taking system. They may be useful in the future, particularly if you found them interesting or intriguing.

      It turns out "waste not, want not" is applicable to ideas as well.


      I can't help but also thinking "waste note, want note" as an interesting turn of expression.

    1. As much as I automate things, though,none of my thinking is done by a tool.Even with plugins like Graph Analysis, I never feel like I'm being presented with emergent connections — tho this is what the plugin is intended for, and I believe it works for other people.

      At what point could digital tools be said to be thinking? Do they need to be generative? It certainly needs to be on the other side of serendipitously juxtaposing two interesting ideas. One can juxtapose millions of ideas, it's the selection of a tiny subset of these as "better" or more interesting than the others and then building off of that that constitutes this sort of generative thought.

  16. Jun 2021
  17. Apr 2021
  18. Mar 2021
  19. Nov 2020
    1. The creation of these types of fake images only became possible in recent years thanks to a new type of artificial intelligence called a generative adversarial network. In essence, you feed a computer program a bunch of photos of real people. It studies them and tries to come up with its own photos of people, while another part of the system tries to detect which of those photos are fake.
  20. Aug 2020
    1. Course as generative learning event

      when Rangle first started to rally around redux - it did feel like a generative learning event

  21. Jun 2020
  22. Jun 2018
    1. In his work on generative effects, Adam restricts his attention to maps that preservemeets, even while they do not preserve joins. The preservation of meets implies that themapbehaves well when restricting to a subsystem, even if it can throw up surpriseswhen joining systems
  23. Apr 2018
  24. Feb 2017
    1. at least five keywords

      I like questions like “Why do I like chicken nuggets?”

      When a girl in the back of the room blurts out this question, half a joke, half a test (Do they really want us to write down any question that we think of?), she seems a bit surprised to have her query treated seriously.

      Thanks for that question Neisha. Let's use it as an example of how to think of keywords for each of your questions. What would be a good one for that question?

      Chicken nuggets.

      Not really. That’s too specific. What's a more general word.

      “Food,” somebody yells.

      Right, write that down Neisha. What kind of food are we talking about?

      Junk food. Fast food. Fried food.

      Right. Right. Where do you get chicken nuggets?

      Down on Nostrand Avenue where all the fast food places are.

      And who…

      Neisha catches the drift, interrupts: It’s in my neighborhood and not in White people's neighborhoods. They get healthy food, which is hard to find where I live.

      So could we add “health” to your keywords?

      Yeah.

      And what else is in your description? What about “inequality?“

      And “racism.”

      What else?

      They’re good, Mister.

      So, what about “delicious? “

      Do we have to write five keywords for every question?

      Yup.

      Ahhhh.

      But what a gift this question was! Do you see how a question can start with something personal, something real for you, even if you aren't sure how important it is? Keep putting the personal pronoun, I, in your questions, then ask your friends and your teachers to help you find the social justice behind them. That's what to look for in your keywords.

    1. Generative

      Twitter bots, recombinatory poetry, generative fiction is infinitely fascinating because it can be so random and unexpected. It's like experimenting with how we ascribe meaning, try to find purpose in the otherwise incomprehensible.