Reminds me of playing https://ncase.me/trust/

I know you can handcraft these by having the text overlay over an exact X,Y on an image but that seems overblown for just these maps than having the text rastered into the image loaded.

So, is there some tool like an SVG creator that helps create these with the responsiveness and everything handled?

"ai2html" sounds like the source is Adobe Illustrator and probably refers to this tool: https://github.com/newsdev/ai2html

Not sure what "freebird" is, some internal software. I can't find much about it with Google search.

    <!-- data processed in process/freebird/process-graphic.js -->
    <!-- asset wrapper : start -->
    <!-- ASSET : START -->
    <div class="g-asset g-graphic" style="max-width: 945px">
    <!-- Intro elements -->
    <!-- inner graphic element -->
    <!-- print .html file -->
    <!-- Generated by ai2html v0.100.0 - 2020-03-22 00:19 -->
    <!-- ai file: maps_la.ai -->
    <!-- preview: 2020-03-19-coronavirus-us-cities-shutdown -->
    <!-- scoop: coronavirus-usa-traffic -->

As for freebird, my search led me to this. https://github.com/freebirdjs/freebird

Seems to be the NYT has an array of pollution sensor IOT across the cities and they are using Freebird to control them. Just my guess.

The Baby University series is pretty interesting.

The titles go like Bayesian Probability for Babies, Neural Networks for Babies and Rocket Science for Babies.

But the answer is that 32 is not better than 16, which is not better than 8, in this specific case. The bit count here is about memory efficiency. There are developers who think/thought that UTF-32 would improve random access into strings because it's a fixed-sized encoding of codepoints, but in fact it does not because there are glyphs that require multiple codepoints.

All who pass must abandon the idea of random access into Unicode strings!

Once you give up on Unicode string random access, you can focus on other reasons to like one UTF over another, and then you quickly realize that UTF-8 is the best by far.

For example there's this misconception that UTF-8 takes significantly (e.g., 50%) more space than UTF-16 when handling CJK, but that's not obviously true -- I've seen statistical analyses showing that UTF-8 is roughly comparable to UTF-16 in this regard.

Ultimately UTF-8 is much easier to retrofit into older systems that use ASCII, it's much easier to use in libraries and programs that were designed for ASCII, and doesn't have the fundamental 21-bit codespace limit that UTF-16 has.

  Also, it isn't difficult to imagine a future where people are criticized for correctly identifying deep fakes.

[1] "Troll" as in "derailing tactic used by people who don't know or care if it's really fake or not" simply to be extra-clear. In politics, throwing out chaff and using derailing tactics to make it effectively impossible to have certain discussions in open fora can be a good way to prevent inconvenient ideas from being spread, in addition to the usual tactics around sowing uncertainty and confusion among the enemy.