[burningChrome]

>> My main problem is defining "ultra-processed," such that I could take action on this in my life.

I asked this question Claude, ChatGPT and Gemini. None of them could give me a good explanation either. One of the themes that did come through was the agents seemed to land on the idea that if something is made purposefully to make you want more of it (in a sense crave it), then its "ultra processed" which is interesting.

I then asked them what the difference between Chipolte and Taco Bell was then. It said some of the ingredients that Chipolte uses are still designed with specific flavorings and salts which would then be considered ultra processed because the point is not to make it healthier. Its to make you want it over other things that would be considered healthy.

It was an interesting conversation, and in the end, I came to the same conclusion, its impossible to tell these days where the threshold is for something to go from processed to ultra processed is.

[zparky]

Thank you for posting your thoughts on a relevant conversation with LLMs, instead of pasting 8 paragraphs of its output.

[robocat]

So ultraprocessed food is addictive food.

It is hard for individuals to fight back against toxic obesogenic[1] advertisers.

Even when we communally create a helpful concept or word, it eventually gets undermined and debased by sociopathic advertising (now with extra helping hand from AI).

Aside 1: there's an art to narrowing down and selecting relevant information from AI generated responses. Plus the art of writing/rewriting good prompts. I fear many people will fail to learn to prompt intelligently (Many people haven't even learned to use search operators over decades!)

Irrelevant Aside 2: AI generated spelling mistakes are fun to look for in AI comments faking humanity. Humans make different mistakes than AI generates - "Chipolte" feels especially human!

[1] prompt: "What are some strong words meaning the opposite of benign (especially as relating to antisocial food advertising)". Aside 3: the response ended "isn’t just harm—it’s harm with intent or harm that multiplies. Accidental harm calls for correction. Engineered harm calls for reform.". A moralising or pandering AI feels systematically dangerous.