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

81–90 of 302 posts

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

#81

Post by neither_color »

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.

add to that food calorie labels arent required to be 100% accurate, there's an acceptable range like +/- 20% iirc. That's enough to throw off a precise deficit. Calorie counting is more about directionally helping you lower your food consumption but it isnt as accurate as people think.

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

#82

Post by darkwater »

tayo42 wrote: post #44
surfmike wrote: post #36

The founder of parrotpal (another AI calorie tracking app, supporting both text and photos) points out that using photos is one of the least accurate ways for tracking food: https://www.instagram.com/parrot.pal/reel/DAB9NtfM250/ Users also seem to like using photos less than text for food tracking: https://www.instagram.com/p/DBoSeQzM3bQ/ The author of this article seems to be against the concept of calorie counting…

> seems to be against the concept of calorie counting as a whole too What are peoples criticisms of calorie counting? Its the only thing that keeps my weight from creeping up. There are way to many calorie dense foods that can easily sneak in if im thoughtless. (I want my 20 year old life style back :( )

> What are peoples criticisms of calorie counting?

Probably the same criticism that applies to all "methods to do a thing". That people often miss the forest for the trees and obsess on the wrong metric (counting calories while still ingesting preprocessed industrial food and beverages) instead of the right one (losing weight while being healthy at the same time).

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

#83

Post by spacecadet »

Kiyo-Lynn wrote: post #34

I’ve tried a few AI-powered calorie tracking tools too. At first it felt super convenient, just take a photo and get the numbers. But over time I kept seeing mistakes. Sometimes it would even tag a salad as soup. Eventually I went back to the old way. Fewer photos, more mindful eating. Honestly I feel better that way.

Honestly this is it for me. Im actively trying to reduce my reliance on tech, even more so when it comes to gimmick tech. A nightly walk with no phone will make you drop pounds too...

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

#84

Post by jclardy »

jonplackett wrote: post #57

I used MyFitnessPal to lose 15kg a few years ago and I have a theory that the act of calorie counting (ie the effort of working it out and labour of putting it into the app) is a key part of why it works. It becomes just slightly more effort to eat something and it makes you stop and think. Making it easier I would expect to make it also easier to eat more.

I find a similar thing with budgeting apps. Manually tracking requires effort and for you to think twice before every purchase, whereas automatic transaction syncing with your bank means you can just not think about it.

In the case of dieting/calorie counting - I don't think you can get away with not thinking about it, especially with inaccurate estimates.

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

#85

Post by snakeboy »

This feels like the same type of critique people have about code-assistant LLMs. We need to be a lot more realistic about their role as a 'next-gen autocomplete' who can write some boilerplate code, not a replacement for a mid-level software developer.

These apps seem useful to reduce the friction of entering your meal's calorie contents from 30 seconds of ingredient estimating, to 20 seconds (auto-filling main ingredients and a baseline portion estimate). As everyone points out, it's impossible for these to get any better than that (GIGO). But reducing the overhead for calorie-counting seems like a reasonable goal and a valuable service. Of course the marketing departments for these companies is going to say whatever bullshit they want; that's their prerogative...

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

#86

Post by adrian_b »

irrational wrote: post #33
nvahalik wrote: post #30

I work for a company that offers nutrition tracking on an app in the App Store. We are not shipping camera functionality yet. But our concept is to not necessarily guarantee the accuracy of portions but to make lookup easier. We also spent the time to get the AI integrated with a verified database. This made our results far more accurate. We tended to find that without the lookup the calories and macros would be gene…

Does that work for homemade food as well? The vast majority of the food we eat is homemade with recipes that don’t have any sort of nutritional information. I’ve always wished there was a simple way to figure out the calories. Taking a picture would be ideal.

For homemade food it should be easier to make reliable estimates of the calorie content, because you know with certainty all the food ingredients and their amounts.

The food ingredients with the highest calorie content, like various kinds of seeds or nuts or flour or meal or oil or fat or sugar or dried fruits or dairy products, come usually with calorie estimates from their vendors.

For other ingredients, like various kinds of meat or of fresh vegetables or fruits, there are online databases with typical nutritional information, like the USDA database. Some of that information can even be found in the corresponding Wikipedia pages.

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

#87

Post by zeta0134 »

neither_color wrote: post #81
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.

add to that food calorie labels arent required to be 100% accurate, there's an acceptable range like +/- 20% iirc. That's enough to throw off a precise deficit. Calorie counting is more about directionally helping you lower your food consumption but it isnt as accurate as people think.

Honestly the main thing was learning that perfectly satisfying food I cook at home tends to land around 600-800 cal for a meal, where eating out almost anywhere is often twice that, even if my beverage is water. Just being aware how how heavy those otherwise delightful meals are gives me pause. Granted these figures are all a side effect of my personal eating habits and preferences, but that's rather the point of the exercise isn't it? There's a lot of individual variation, and the individual is probably best served by a personalized approach.

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

#88

Post by yoz-y »

jonplackett wrote: post #57

I used MyFitnessPal to lose 15kg a few years ago and I have a theory that the act of calorie counting (ie the effort of working it out and labour of putting it into the app) is a key part of why it works. It becomes just slightly more effort to eat something and it makes you stop and think. Making it easier I would expect to make it also easier to eat more.

Yep. Iirc any kind of diet works as long as it is restrictive. Even if the restriction is “do not eat anything that has blue on the package”.

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

#89

Post by mkfs »

Fat is the most calorie-dense macro at 9 per gram (protein and carb are both 4), so your accuracy is going to be heavily skewed by how accurately you can identify the fat content of the meal. The problem is that it's often in forms like oils that don't seem well-suited to AI image recognition.

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

#90

Post by ap99 »

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 might know the inner workings of the "machine" but we can control inputs (food) and measure outputs (weight).

Let's say I eat 3 hamburgers per day and gain ~1kg per month.

If I reduce that to 2 hamburgers per day I maintain my weight.

And when I reduce that to 1 hamburger per day I lose ~1kg per month.

That is simple. As for how I feel while eating 1 hamburger that's another question.

I can try to improve the machinery of my body by e.g exercising, eating healthier food, adjusting my macros, taking supplements, etc.

So now I can eat 1.5 hamburgers per day and I feel better and still lose 1kg per month.

Post reply on HN