Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

I used AI-powered calorie counting apps, and they were even worse than expected

lifehacker.com

Post reply on HN
Print view View original

71–80 of 302 posts

Re: I used AI-powered calorie counting apps, and they were even worse than expected

#71

Post by bitdivision »

bandrami wrote: post #70

AI aside we don't really know how many calories a given person will get from eating a given piece of food. We measure food's calories by burning it, but we don't get nearly all that energy from dissolving it (just consider the fact that solid human waste burns, for instance). At best we've established some (fairly broad) minima and maxima.

Presumably the recommended daily calorie intake figures (and the calculators for these) do to some extent take this into account? i.e. I would imagine they're based on experimentally tested values rather than theoretical energy expenditure.

Re: I used AI-powered calorie counting apps, and they were even worse than expected

#72

Post by blibble »

p1dda wrote: post #63

I welcome every article that exposes just how bad "AI" is.

it still also can't count the legs on a dog

https://vlmsarebiased.github.io/

another decade and we'll be able to get the Not Hotdog app from Silicon Valley

Re: I used AI-powered calorie counting apps, and they were even worse than expected

#73

Post by tough »

wahnfrieden wrote: post #69
tough wrote: post #65

Earlier quoted context omitted.

what's worse, this could be 100% done in local llms since its such a small batchable job

What model can run on a phone with accuracy closer to 4o than Gemini 1.5?

idk tbh asking an llm to estimate calories seems like such a -bad idea- on the first place, that doesnt seem that having a better model would improve results much.

Re: I used AI-powered calorie counting apps, and they were even worse than expected

#75

Post by matkoniecz »

bandrami wrote: post #70

AI aside we don't really know how many calories a given person will get from eating a given piece of food. We measure food's calories by burning it, but we don't get nearly all that energy from dissolving it (just consider the fact that solid human waste burns, for instance). At best we've established some (fairly broad) minima and maxima.

"We measure food's calories by burning it" - are you sure that how it is done?

Humans (unlike say cows) do not digest parts of what is available in plants, which makes some vegetables filling and ultra-low caloric.

"we don't really know how many calories a given person will get from eating a given piece of food"

exact numbers of unknown, but difference is not so great, except truly extreme cases (and vast majority of fat people is not fat due to them)

Re: I used AI-powered calorie counting apps, and they were even worse than expected

#76

Post by adrian_b »

squillion wrote: post #50
tayo42 wrote: post #47

Earlier quoted context omitted.

Oh I guess, I try to go towards overestimating calories taken in and underestimate my calories spent. It does work for me, but being vigilant about it is tedious so I'm open to hearing if people have better ways.

Take a look at these books: - David Ludwig, Always Hungry? - Mark Hyman, The Blood Sugar Solution The first one is very accessible, the second one very posh. But the underlying approach is the same: no calorie counting, just good food in the right proportions.

Good food in the right proportions is necessary, but not sufficient. The total amount of food is at least as important.

I eat only good food in the right proportions. However, it would be enough to double the amount of the food that I eat at one meal, for the next day to see a few hundred grams of additional weight.

I must plan the amount of food to be eaten before starting to eat. Otherwise, I could eat effortlessly not only the double of the amount that I have planned, but even the triple amount or more, with a corresponding increase in the weight gain.

Perhaps there are people who might stop automatically from eating, before ingesting too much, but I am not one of them and looking around I have never met one of those people.

For myself and for most people that I have seen (with the extremely rare exceptions of those people who remain thin despite claiming to eat as much as they can, and who may actually have impaired food digestion or absorption) hoping to stop naturally before overeating does not work. The only thing that works is deciding how much to eat before starting to eat, then never eating more than that. For planning how much to eat, calorie counting works fine.

Re: I used AI-powered calorie counting apps, and they were even worse than expected

#77

Post by matkoniecz »

perching_aix wrote: post #66

How many times a day and how many unique food items does one need to eat to continually remain so hopelessly unfamiliar with their caloric intake that they need an app to track it? What role can AI possibly have in this, when all you need to do is either turn around the packaging or do a web search for the dish you cook to get a number?

> they need an app to track it?

"track" is important here

How packaging or a web search will help with exactly remembering how much was eaten in past?

Re: I used AI-powered calorie counting apps, and they were even worse than expected

#78

Post by swiftcoder »

thierrydamiba wrote: post #26
m_ke wrote: post #21

Earlier quoted context omitted.

I've heard from someone who knows that they're scamming people like crazy. Supposedly they also setup a bunch of LLCs to hire influencers then never paid them.

I think a claim like that requires proof. You’re accusing them of fraud.

Is fraud not a reasonable assumption, when an app that fundamentally does not do what it claims to, nonetheless has legions of glowing reviews?

In the best scenario we are in a TornadoGuard (https://xkcd.com/937/) situation. More likely the developers are paying for reviews

Re: I used AI-powered calorie counting apps, and they were even worse than expected

#79

Post by automatic6131 »

jonplackett wrote: post #55

I always just assumed these were terrible. There just isn’t enough info in a photo, even for the world’s leading calorie counting expert, to give a sensible estimate of the calories content. Ans even more so for macros. Add to that the fact vision is still pretty crap and you have a complete joke that I can’t believe anyone is paying for.

I bet @onionweigher could - if all you ate were onions

Re: I used AI-powered calorie counting apps, and they were even worse than expected

#80

Post by jclardy »

jonplackett wrote: post #55

I always just assumed these were terrible. There just isn’t enough info in a photo, even for the world’s leading calorie counting expert, to give a sensible estimate of the calories content. Ans even more so for macros. Add to that the fact vision is still pretty crap and you have a complete joke that I can’t believe anyone is paying for.

I've used the AI features in Lose It and was pretty impressed, not for the caloric estimate, but when given a picture of a breakfast burrito it accurately split it out into the component parts, from there I could manually adjust amounts easily for food I make without having to manually search out every ingredient manually. As an additional tool it is great.

The issue of course is these new apps built with a sole focus of AI images for tracking. With a photo of restaurant food you can't see a sauce and the 5-10g of additional sugar content, can't get accurate guesses of what is in breads/pastas and figuring out general volumes of foods is not possible unless you have some kind of standard size reference in the frame.

Post reply on HN