Typescript 2023-07-05

Text extraction. See Typescript Archive

10:20:44

10:20:44 From jsmiller (my audio keeps starting at 6% which is "is anyone talking?" - fixed now) <style> ... </style>

10:21:45 From jsmiller (verbose style element not needed)

10:23:48 From jsmiller SVG is a friendly format for programmers.

10:25:50 From Brian \n is an escape code for newline in many computer language contexts.

10:27:15 From jsmiller unpack a structure <---> better rules for unpacking

10:29:03 From jsmiller "an Omnigraffle document unpacks to a wiki SVG like this [rules]" prompted recipe building? "like this?" "more to that side", "more to the other side", "following this example"

10:29:44 From Brian declarative kind of is like that...

10:31:30 From jsmiller (native graph) -> (exported SVG) -[via a customized process to]-> (wiki SVG of desired format)

10:33:41 From jsmiller "generate the prime numbers" "...digits of pi" (why a generator expression doesn't always have a happy version as an expansion to an array)

10:39:41

10:39:41 From jsmiller DOM visitor instead of CSS selector an element node has contents and attributes a text node has no attributes?

10:40:48 From jsmiller between or within elements SVG DOM is part of the browser document DOM which may mix SVG and HTML

10:42:42 From Eric Dobbs https://developer.mozilla.org/en-US/docs/Web/API/Node/nodeType

10:46:04

10:46:04 From Brian This kind of seems like the wrong API if you only care about 'aliens?' and "why yes" parts...

10:48:44

10:48:44 From Marc Pierson Miro has comment affordance.

10:49:44 From Brian Maybe innerText rather than textContent...but I haven't done any of this kind of programming...so I'm the peanut gallery for this conversation...lol.

10:50:50 From Paul Rodwell Crockford considers generators bad, as what the code does is different from how it looks. He suggests a better way is to use a function (factory) that returns a function (generator)

10:51:16 From Brian https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/yield*

10:51:27 From Marc Pierson What are “generators”?

10:51:44 From Brian Sounds like it does the iterator and yields on each iteration.

10:52:01 From jsmiller Generators are functions that you can ask for "next one?" sometimes there is always a next one, sometimes you run out of next ones

10:53:02 From Brian For example, you could ask for next digit of pi. A generator keeps some state so it knows what "next one" means when you ask it.

10:53:05 From Marc Pierson So generators would need some sort of stop rule?

10:53:16 From jsmiller yes, either inside the generator or outside

10:53:27 From Marc Pierson Thank you!

10:53:28 From jsmiller "give me the next one... ten... etc."

10:53:41 From Brian Generator could be finite. like an array of [1,2,3] or 1..3 etc has 3 elements.

10:53:45 From jsmiller also a generator can answer "no more!"

10:54:48 From Brian But importantly, they can be infinite too, so they need some consideration, like there might not be a stop condition.

10:55:16 From jsmiller yes! which is why even though generators can simplify code as expressed, they can complicate behavior

10:58:14

10:58:14 From jsmiller This is your interpreter! "turtle this text!"

10:58:23 From Brian A take on 'literate programming'

11:02:03

11:02:03 From jsmiller set an RGBA style color?

11:04:11 From Brian This is almost the 21st century sitting around a campfire...except we all are watching turtle graphics...It's great.

11:07:06

11:07:06 From Brian Similar to https://www.edwardtufte.com/tufte/posters. When charts' spatial components relate to quantities, the charts become way more powerful. It really brings charts to life and makes for impactful story telling.

11:08:43 From jsmiller oh right like Napoleon's march into Russian and retreat. "a narrative into a story graph picture" (with floatover as to what the graph elements correspond to in the story)

11:09:44 From Brian dallE sort of does that for paintings. https://openai.com/research/dall-e

11:10:31 From jsmiller orange: caution; red: alarm Can one tell The Tortoise and the Hare with this?

11:11:21 From Marc Pierson http://marc.tries.fed.wiki/view/sailing-with-rob-b

11:11:42 From jsmiller "follow the bouncing ball!" "The Wreck of the Edmund Fitzgerald" ?

11:13:02 From jsmiller "The turtle vocabulary for this narrative is: [...]"

11:13:27 From Peter Dimitrios "It's turtles all the way down..."

11:13:41 From jsmiller sea turtles

11:15:15 From Brian Some story writers are "gardeners" where they start writing and see what the character does. Some writers use structure and models and figure out what one of the scenes is going to be, then outline the plot and character progressions. Along those lines, if you animate the turtle graphics, are you "gardening" or is there a specific emotional response or immersion response you are intending with animation?

11:15:28 From Peter Dimitrios This feels a lot like what happens during 'prompt engineering' - leading the darn AI by the nose

11:16:44 From Peter Dimitrios If you squint a bit the AI transformer as 'advanced Grep' for the data it has ingested

11:17:21 From Brian squinting more than a bit I think. ;)

11:18:49 From jsmiller advanced grep with color themes a prompt is a query

11:19:23 From Brian And biased commentary, which I don't want from grep.

11:19:41 From jsmiller Men Explain Things To Me

11:20:18 From Peter Dimitrios Mansplaining - AIsplaining

11:20:25 From jsmiller :)

11:20:27 From Marc Pierson http://sailing.marc.dojo.fed.wiki/view/welcome-visitors/view/cld-v2

11:20:54 From jsmiller turtle graph fireworks

11:22:57 From jsmiller "an artist ringing changes on their own work" is another approach where generative art can be productive

11:23:25 From Paul Rodwell https://www.theregister.com/2023/07/05/openai_pauses_bing_search/

11:23:53 From Peter Dimitrios FYI for eric: Python has a built-in Logo Turtle, and it’s great for reinforcement learning (Tom Grek 2018-10-14): https://towardsdatascience.com/python-has-a-built-in-logo-turtle-and-its-great-for-reinforcement-learning-6b0dca4acb73

11:24:07 From jsmiller concept-shaped graph databases Hofstadter anticipated a bunch of the LLM stuff the "meaning" thing as an observed strong association "Hofstadter jukebox"

11:25:12 From jsmiller as a search term concept-shaped words "here is a logical proposition" / "here are examples"

11:26:46 From Paul Rodwell https://player.fm/series/getting2alpha/doug-hofstadter-reflections-on-ai

11:27:22 From Marc Pierson https://beta.plectica.com/maps/ODWY4VHBN/edit/7CNV04Y6M

11:28:01 From Paul Rodwell https://gamethinking.io/podcast/s9e2-douglas-hofstadter/

11:28:39 From Ward Cunningham http://found.ward.fed.wiki/hofstadters-surprise.html

11:32:54

11:32:54 From Brian https://www.youtube.com/watch?v=2viJzPDDU8Q Brian Lowrey on Mindscape podcast and explores how one's social self is intertwined and shaped by community and how communities are defined by other communities.

11:34:24 From Paul Rodwell It’s all HTML - with custom elements

11:36:34 From Marc Pierson https://beta.plectica.com/maps/56EAUVNMZ/export

11:38:37 From Marc Pierson https://en.wikipedia.org/wiki/Derek_Cabrera

11:41:22

11:41:22 From jsmiller aha, Cornell, Santa Fe Institute (Derek Cabrera) "how we know things is equally important to what we know"

11:42:57 From jsmiller (I am actually playing hooky from my creative writing class now)

11:43:28 From Robert Sterbal I'm happy with the formatting (of AI)

11:46:46

11:46:46 From Robert Sterbal https://podcasts.apple.com/us/podcast/promises-and-peril-small-modular-nuclear-reactors/id1582667621?i=1000619257755

11:46:55 From Brian https://ggml.ai/ and https://github.com/ggerganov/ggml are chomping at openai's heels...

11:47:04 From Robert Sterbal Nuclear power discussion

11:47:15 From Peter Dimitrios Pete Wood

11:49:09 From Brian Langton's ant is a cellular automata which makes for an interesting visualization/animation. https://www.youtube.com/watch?v=1X-gtr4pEBU

11:49:52 From jsmiller https://en.wikipedia.org/wiki/Langton%27s_ant

11:52:36

11:52:36 From Eric Dobbs I’ve gotta go prep for my next meeting. Fun to see you all again.