Here in 2020 I am writing a review of the Dieta app. This app is designed to use artificial intelligence (AI) to take negative health symptoms and correlate them with food input, in an attempt to determine a cause. I think the app is intended for people with IBS. However, I attempted to use the app to form a correlation with my daytime sleepiness and occasional heartburn.

I used the app for approximately a month or maybe as little as 20 days, before I lost the will to do it anymore.

My Results

Behold! The AI gods have spoken! Here is the summary from my Dashboard:

My safest food: bacon
My worst trigger food: bacon

Well there you go. I should both avoid and eat as much bacon as possible! I haven’t been doing either of these things! It’s no wonder I’m not an elite athlete. It is moderation that is killing me.

The Big Picture

Okay. What I think is going on is the AI thinks that bacon is good for one symptom and bad for another, but the dashboard doesn’t make it easy to understand which is which. I tried to look at the Recommendations screen but it displays an incorrect error stating that I have not filled out the GI Survey. I have, and I have contacted support twice with no response.

I think the data is probably in here, but I am not exactly sure how to read it. I also don’t know how wide the window is for a reaction. I get the impression they are looking for food intolerances in a 24 hour period, but they really should extend that to 96 hours.

I’ve looked at the dashboard several times during my usage of the app and I have noticed a trend that my safest foods seem to be straight up meat and salt. I don’t really consume these foods in isolation so it is fascinating that it’s separated these from everything else. Is it recommending me a carnivore diet? Or is this just some weird anomaly with how frequent foods are more likely to be marked as safe? So much is unclear.

Summary

I like what these guys are doing. But because they never responded to my support requests, I think 2020 hasn’t been a good year for Dieta.

I think they have some work to do to make the dashboard understandable. Instead of telling me to simultaneously eat and avoid bacon, they could tell me reasons why for each—relate it to the symptoms. They could also tell me with what degree of certainty they are making these recommendations. They could fix the Recommendations page in the dashboard so it actually works instead of displaying an error. They could also fix the asynchronous requests in the app, so it doesn’t perform and then wait for twenty different data lookups while I’m typing each letter of a single food. (You can separate foods by commas by the way, to bypasses this horrible time-consuming bug.) They could offer to connect to Cronometer so users don’t have to fuss so much with the Dieta app. (It might take more time to make a diagnosis, but it might be easier to onboard new users.)

In the future I think this app could be a real game changer. It’s practically there already. The data is in the dashboard; it is just hard to read. I think it still might actually work for some people in its current state. As for me, I’m still stuck in the world of quantum bacon.