2 ms·
I've had very similar ideas, and although I can't quite get the vision of how you're organizing the material since your last pivot, I want to throw in a few ide
by pessimizer 4y ago
I've had very similar ideas, and although I can't quite get the vision of how you're organizing the material since your last pivot, I want to throw in a few ideas (that you may be already doing.)
1) Cooking skills are physical acts of dexterity and timing that can be explained and modeled by text and videos, but not mastered. This is a good place for the use of spaced repetition; every time someone makes a recipe that uses that technique, ask them how they did (and give them the criteria e.g "was the sauce shiny?") Make practicing fading or weak skills, or taking advantage of mastered skills, a criteria for suggesting recipes to the user. i.e. "do you want an Easy, Challenging, or Difficult recipe?" and select based on their current skill level at the techniques involved (and the ingredients at hand.) Choosing Easy would be biased towards recipes utilizing techniques and ingredients that they have mastered, Challenging would be for techniques that they are learning, and Difficult would be for introducing new techniques.
2. Organize learning more around techniques than ingredients. Some techniques only apply to a single ingredient, but most apply to many ingredients. If techniques are verbs, ingredients are conjugations - many irregular, but most regular. Some ingredients require a higher level of a particular technique than others. Figure out what ingredients they like by asking what meals they like, and use those to help suggest future techniques to learn.
3. Many people want to become vegetarian, or vegan, or paleo/Mediterranean/whatever, but don't like the food yet. If a program like this suggests logical progressions to train the palate towards a user's goal, it could help them switch to these diets (clinically picky eaters, or people who have discovered they are pre-diabetic, diabetic, celiac, or that they have fatty liver disease, have a serious problem with this.)
This could be amenable to a machine learning recommendation approach if you ask what meals people enjoyed, you know the techniques they're good at, and you know their diet goals. You could say "buy avocados on your next shopping trip and you can try X, Y, or Z, which people who like what you like also seem to like."
- whimsicalism 4y ago> 1) Cooking skills are physical acts of dexterity and timing that can be explained and modeled by text and videos, but not mastered. Not sure I agree - I actually get a lot of value out of videos for cooking techniques and they can often nip bad habits in the bud. For instance, I recently discovered a new method for matchsticking carrots from a video that i really like.
- pessimizer 4y agoI didn't mean to suggest that people don't learn from videos, I think they mostly learn from video, from Julia Childs to America's Test Kitchen to youtube. I was trying to say that you get good at something by trying to do it, not by watching the video over and over again.
- mizzao 4y ago1) totally agree with the spaced repetition. Can't quite get that in a MVP though ;) 2) this is a really interesting idea, but I'm not sure it's really a conjugation, but just different nodes on the knowledge graph. What do you think is the benefit of a noun-verb relationship? We believe there's some incredible fruit in a serious computer science approach to cooking and food. Here's a preview: https://parsnip.substack.com/p/vision-part-one https://parsnip.substack.com/p/vision-part-one