- From: Paola Di Maio <paoladimaio10@gmail.com>
- Date: Fri, 27 Jun 2025 17:35:46 +0800
- To: Milton Ponson <rwiciamsd@gmail.com>
- Cc: W3C AIKR CG <public-aikr@w3.org>
- Message-ID: <CAMXe=SosMO1qhkdfug2qiqGJuzztE1Nagy_ipsNRw-Yt0MZe9A@mail.gmail.com>
Milton The video the ' way of coding' contains interesting points, it opens up the discussion on AI into so many dimensions, I like to see AI through this lens and the book shows how he started off a traditional analysis of a domain using a non technical approach, then produced a book, then used some new tech to transform into something else I am not sure everything he says is true. I am watching multiple videos on the same topic simultaneously, It is tru that AI processes non textual knowledge such as visualizations e, of course, thanks for picking on this point in my previous email in fact AI can be good at that. I was referring to logic/reasoning and other constructs that benefit from having a natural language counterpart- text to execute various functions Imagine trying to query an AI about an image, or an audio, and the output would be using a video or audio? It seems that humans communicate with each other, and now with machines using natural language, and that knowledge captured and stored using natural language can be processed and used by systems with or without visuals. A reliable system still needs to have formal representations an index. an ontology, to map it against. a natural language interface It is interesting times I am sure LLM s are here to stay - despite their limitations PDM On Fri, Jun 27, 2025 at 3:54 PM Milton Ponson <rwiciamsd@gmail.com> wrote: > The Way of Coding is just one more multi-disciplinary take on how to > translate "knowledge" into machine readable format. > > Philosophy, music, visual arts and performing arts all capture > "universals", that is identifiable audible or visual ideas, concepts and > experiences (including emotions) that transcend language and culture, and > in the same vein so do culinary and epicurean experiences. > > And here is where we also see the problem of capturing knowledge. When we > strictly limit ourselves to what we can capture as data, information and > textual knowledge we arrive exactly at the starting point indicated by me > earlier, elaborated upon by ChatGPT and converted into StratML by Owen. > > Audiovisual input can also be processed by artificial intelligence, but it > transcends textual input and output, a prime example being the ability of > AI to produce whalesong enabling humans to communicate with cetaceans, as > has been done recently. > > We are also able to capture, classify and manipulate movements, brainwave > activity and electric signals in the nervous system, e.g. in bionics, > advanced prosthetics, robotics and mechanotronics. > > These obviously are also not text based. > > This is why we must limit ourselves to semiotics, symbol set, pictogram, > (petro)glyph, alphabets and other identifiable visual inputs for knowledge > and knowledge representation. > > We may now also intuitively make the distinctions between artificial > intelligence, artificial general intelligence and artificial > superintelligence. > > The first is the capability of processing identifiable visual inputs. > The second includes all audiovisual inputs as well and movements, > brainwave activity and nervous system signals of humans and other species. > Superintelligence includes the capability of processing the universals in > a way that transcends the anthropocentric limitations, that is the > cognitive, cultural and emotional makeup of humans. > This is why generative artificial intelligence based on large language > models will soon reach a dead end. > > As AIKR CG we can limit ourselves to the visual inputs that can loosely be > described as text. > And here is where logic, set theory, mathematical modeling and > computability set the stage for rigorous formalization, including for > philosophical concepts ( reality, mental constructs and ideas, perception, > cognition, reason etc.). > > I have been making notes and I am now in the process of putting these > into some articles. > > This will be a work in progress. > > Milton Ponson > Rainbow Warriors Core Foundation > CIAMSD Institute-ICT4D Program > +2977459312 > PO Box 1154, Oranjestad > Aruba, Dutch Caribbean > > On Thu, Jun 26, 2025, 20:59 Paola Di Maio <paola.dimaio@gmail.com> wrote: > >> One of the reasons we are developing vocabularies *lists of terms, bags >> of words >> is because these days text can be processed easily and intelligently by >> machine >> and is powering computation >> >> The entire language model driven AI, is based on text *text input, text >> based libraries. text output >> >> Diagrams, formal and informal visualizations are useful but ultimately, >> whatever the diagram >> represents, needs to be represented as text for LLMs too be processed >> >> We need to be able to construct natural language expressions and leverage >> natural language >> *symbolic >> >> I find the vibe coding conversation interesting and reflective of the >> state of the art >> >> https://www.youtube.com/watch?v=7s9C92Pkcc0 >> >> >> >>
Received on Friday, 27 June 2025 09:36:28 UTC