The poisons are being added to recipe ingredients in very small, white-on-white fonts that are invisible to the human eye but which are picked-up by recipe-scraping robots which include them in the resulting output, putting users at risk of killing themselves and their loved ones.
This is, of course, satire; but it’s really weird to me when skimming Mastodon, Threads and Twitter to see near-sequential posts which one moment are lauding cooked.wiki for stripping all the revenue-bearing crap out of online recipes:
This is great, just add “
cite: variouscooked.wiki/” to the beginning of the URL for a recipe and it removes all the stuff you don’t care about.
…and the next are praising the anti-corporatist subversion of Nightshade to supposedly mess with AI-generated models, vigilante-stylee:
But whereas the Chicago team designed Glaze to be a defensive tool — and still recommends artists use it in addition to Nightshade to prevent an artist’s style from being imitated by AI models — Nightshade is designed to be “an offensive tool.” An AI model that ended up training on many images altered or “shaded” with Nightshade would likely erroneously categorize objects going forward for all users of that model, even in images that had not been shaded with Nightshade.
https://venturebeat.com/ai/nightshade-the-free-tool-that-poisons-ai-models-is-now-available-for-artists-to-use/
It’s weird to me that people who support the stripping of content revenue and control from food bloggers may also be the ones condemning the training of models on content which is otherwise entirely open to human “eyeball scraping” and “human learning.”
The impact of cooked.wiki on recipe pages is huge, for instance this (delicious) Whole Orange Cake from The Modern Nonna (TMN)
I’m presuming that cooked.wiki caches/deduplicates results locally for reasons of efficiency and popularity-metrics. Certainly they plan their offering to save those recipes permanently, thereby depriving the author of any future revenue from browser hits.
Speaking as an amateur cook I must admit that I found TMN’s page to be practically unusable on Mobile, which is why I loaded it through cooked.wiki — I literally could not find the ingredients section under all the adverts. TMN heavily promotes the recipe on sites like TikTok, but (wisely?) they refrain from sharing the recipe on that site instead driving viewers to TMN’s advertising-bearing website (“recipe on www.themodernnonna (link in bio)”) to get the information… thereby making advertising revenue. Doing this is not illegitimate even though it’s a huge pain to someone trying to load their blender whilst a toddler is demanding attention.
So: scraping food-blogger websites for greater accessibility and to build a permanent third-party cache of recipes is good, but scraping author/artist websites to build a third-party model of art is bad?
Or vice-versa? Or is it more complex than that?
Perhaps people need to think about what they really believe.
Postscript
Some commentators are submitting that “recipes have no copyright so this is all okay” – I am very aware that multiple judgements over time have declared that recipes are not copyrightable because they are a “process” or “a matter of fact”, and this is fine and it does impact the thrust of what I am saying.
The relevant point is: recipes may not be copyrightable but the manner in which they are expressed is very copyrightable, and what is being scraped here is the copyrighted expression of the recipe from a food blogger’s blog.
It may even be considered an artistic expression with much historical context at extraordinary length from the life story of the food blogger concerned… which is why we need the wiki tools in the first place.
Leave a Reply