AI GLOSSARY NOTES


The working title now is something around the word Amulet - 
dense amulet; cognitive amulet; synthetic amulet; computational amulet; amulet of intelligence.
Pluralized?
dense amulets; cognitive amulets; synthetic amulets; computational amulets; amulets of intelligence, distributed amulets.
synthetic amulets - my favorite

In the intro I explain how AI is a dense technological placeholder like nuclear energy in 50s or DNA in 90s that was central for cultural imaginaries of the time. As soon as Ilan finishes the edits we will share it with you.
Yeah a great intro! Still working on it.

• Importance of font supporting rich formatting (Italics, superscript)
• Necessity of different hierarchical regions for footnotes, and full reference lists close to the text
• A certain compaertmentalization of terms / art. Art project titles are not glossary terms.
• Having said that, all biographies (writers and artists) will be grouped together in the end of the book.
• The book is a small format (12x18), we need to keep reading comfortable, and I would vouch for the largest possible typeface weight.
• 100+ terms + 50 artworks, + bios, this might be a consequential size book with >300 pages. I am all for underlining this, with a slight heavy paper and a generous text distribution.
• We need to find the right iconography for the "filigranes/transparency" graphics. Digitization is only a very generic backbone of computation and it doesn't fullfil the purpose here. We have been thinking of something as extreme as Hildegard Von Bingen's lingua ignota, sorry this is Middle-Ages not Renaissance. Language vs Script vs Speech can be approached through fictional or alien languages. As  I liked the lingua ignota the most, I initially had associations with Codex Seraphinianus or Klingon, which would be most certainly protected by the author's rights. So if you wish I think it might be fruitful to explore this first proposal through a dimension that would be much more open to interpretation of artificiality, not narrowing it down to machines. Typefaces like new 'Wired' by Nadine Wetzel or slightly old now but very loud when released 'Icarus' by Sophia Brinkgerd and 'Sword' by Kazuhiro Aihara might be a relevant reference. Adding here that lingua ignota is certainly not under copyright protection, so it can be something we coudl look at.
• I would stay away from Google, AWS or any other FAMGA visual/graphic/iconic languages. I totally agree, we cannot proceed with any technocratic visual indentity, this is too narrow even for conventional AI understanding, and as you can see from my previous comment, we are trying to develop an unconventional understanding of artificial intelligence. 
• Also although I think that the diagrammatic illustration of a NN is beautiful, we also need to take distances from a too litteral understanding of AI. Intelligence is a multi-faceted largely unknown object and flattening it in representations is exactly what this book has been struggling to go against. No perceptrons, no convolutions, no fully connected layers are going to be helpful here. Yes, potential workaround would be the animal automaton, for instance, like this anticolonial mechanical tiger https://en.wikipedia.org/wiki/Tipu%27s_Tiger#History or Shennon's Ultimate Machine which was essentially suicidal https://gigazine.net/gsc_news/en/20120226-shannon-ultimate-machine/. You might also find relevant the aesthetic of Soviet Cybernetic books, for instance my favourite is this cover http://www.libex.ru/dimg/d296.jpg from 1967. I like this one from 1982 https://b.itemimg.com/i/239823121.1.jpg?1 or this from 1967 https://www.livelib.ru/book/1001219218-nachala-kibernetiki-aleksandr-lerner. You can also see that the first book on cybernetics did not have any images of computers on its cover https://www.ozon.ru/product/signal-25392154/ (the style is off obv, this is just context for two previous covers). If we go into contemporary references they would be more of an alien/slime/swarm nature than machine.
• The book is highly distributed project in terms of nationalities, experiences and professional capacities.This may be something to leverage in the design.
• re: the format are you sure that a small book like ours can have so many lines (>45) per page? My calculations is that a Baskerville font might need to go down up to 8.5 points to cover the surface with 45 lines. For a (soon+)40y like me, it's a challenge to read. :)
• Yes! Dictionnaries and glossaries are not entirely concomittant. Dictionnaries can afford to have small typeface because what you need to get ouf of the text is usually a small chunk of information but a glossary is like a short magazine article (fait divers) = a different attention economy.
• From the second proposal I like the small bottom-page illustrations with the etchings (culs-de-lampes).
• Regarding the two columns: I am wondering if this format is not entirely compatible with a more "typical" text distribution flow of glossaries but I am certainly not a connoisseur here.
• CANCELLED: Title font is not Baskerville ||||| I regret Baskerville's lack of OpenType features. For a font that is indeed very popular in lexicography the "fi" of "artificial" is a big time mess, that can hardly be fixed with increased kerning.