6 Comments
User's avatar
Eric Jain's avatar

Even if food was labelled with complete accuracy, not everyone absorbs nutrients equally efficiently at all times, and how food is prepared can make a big difference, too!

I see labels less as data, and more as an educational tool...

Expand full comment
Richard Sprague's avatar

That's a good point. I'm surprised at the number of people who don't know very basic facts about what they eat (e.g. does X or Y have more calories?) and just paying attention to labels over time is worthwhile, whether you track your food or not.

Expand full comment
Conor's avatar

I tried using chatGPT to design a meal plan for grams of carbs, protein, and fat, but when I checked the math it didn’t add up. Might want to check your diet for a day with Cronometer and see if the LLM can do math.

Expand full comment
Richard Sprague's avatar

I had a similar complaint a year or two ago with the earlier models but now I’m finding it pretty accurate.

Expand full comment
Conor's avatar

Good to know!

Genetic Lifehacks (great content) recently posted about using the major LLMs to analyze SNPs and concluded “ The way the AI information is given is so confident and smooth that I wanted to believe it all - even the stuff that I knew was completely wrong.”

https://substack.com/@geneticlifehacks/p-163710587

Expand full comment
Richard Sprague's avatar

Interesting, thanks! I uploaded my 23andme SNPs to Claude and it gives confident-sounding results but I'm skeptical. Perplexity on the other hand seems to always back its conclusions with accurate source.

Expand full comment