yes, it's an article about scrambling text using fonts in order to prevent LLM scraping.

It's a cute idea. I'd be more at ease if the author had described it as a neat toy that shouldn't be seriously used for anything, but they don't seem to be thinking that way, only describing the accessibility problems as a "drawback".

hm-m, that sounds familiar… ah. Yeah. I've heard of this trick before.

Russia's Central Election Commission (ЦИК) pulled the same stunt with its website during the election, when that website published intermediate results, under the pretense of "preventing the system from crashing from unprecedented attacks"; I had to look up the date when exactly, seems to be 2021, which feels about right:
forbes.ru/tekhnologii/440479-p…

How was this supposed prevent any attacks and what *kinds* of attacks even is beyond me to this day. LLMs did not exist back then. Or in the very least weren't anywhere near as capable.

Памфилова объяснила кодирование результатов выборов на сайте ЦИК

Центризбирком закодировал результаты выборов на своем сайте, чтобы «система не вылетела» из-за обрушившихся на нее «беспрецедентных атак», заявила Элла Памфилова. Ранее эксперты обратили внимание, что копировать данные с сайта ЦИК стало невозможно, п

^{Тимур Батыров (Forbes.ru)}

In #Safari on an #iPhone going to the link I see... something that resembles a blog, but tapping the "reader view" I get the following, proving the point of the headline. If we can't read it, an #AI #Bot can't scrape it.

<begin text>
# /// script # requires-python = ">=3.12" # dependencies = [ # "bs4", # "fonttools", # ] # /// import random import string from typing import Dict from bs4 import BeautifulSoup from fontTools.ttLib import TTFont def scramble_font(seed: int = 1234) -> Dict[str, str]: random.seed(seed) font = TTFont("src/fonts/Mulish-Regular.ttf") # Pick a Unicode cmap (Windows BMP preferred) cmap_table = None for table in font["cmap"].tables: if table.isUnicode() and table.platformID == 3: break cmap_table = table cmap = cmap_table.cmap # Filter codepoints for a-z and A-Z codepoints = [cp for cp in cmap.keys() if chr(cp) in string.ascii_letters] glyphs = [cmap[cp] for cp in codepoints] shuffled_glyphs = glyphs[:] random.shuffle(shuffled_glyphs) # Create new mapping scrambled_cmap = dict(zip(codepoints, shuffled_glyphs, strict=True)) cmap_table.cmap = scrambled_cmap translation_mapping = {} for original_cp, original_glyph in zip(codepoints, glyphs, strict=True): for new_cp, new_glyph in scrambled_cmap.items(): if new_glyph == original_glyph: translation_mapping[chr(original_cp)] = chr(new_cp) break font.save("src/fonts/Mulish-Regular-scrambled.ttf") return translation_mapping def scramble_html( input: str, translation_mapping: Dict[str, str], ) -> str: def apply_cipher(text): repl = "".join(translation_mapping.get(c, c) for c in text) return repl # Read HTML file soup = BeautifulSoup(input, "html.parser") # Find all main elements main_elements = soup.find_all("main") skip_tags = {"code", "h1", "h2"} # Apply cipher only to text within main for main in main_elements: for elem in main.find_all(string=True): if elem.parent.name not in skip_tags: elem.replace_with(apply_cipher(elem)) return str(soup)
<end text>