Cookbook Order Form.Ai: Fill & Download for Free

I will try to keep this as concise as possible.Edit: Somebody merged the original question to this question, so the premise becomes irrelevant.To become a full stack AI/ML engineer, it is imperative that you have a complete grasp of the mathematical foundations of ML so that you can build upon concepts easily. The basic mathematical skills required are Linear Algebra, Matrix Algebra, Probability and some basic Calculus.Linear AlgebraThe best source to study Linear Algebra is Prof. Gilbert Strang’s Linear Algebra book/course. Video Lectures | Linear Algebra | Mathematics | MIT OpenCourseWare (MIT OCW). There are 34 lectures and believe me, they are completely worth it as after completing this, linear algebra should not pose any more problems for you. Solve some exercises/exams if you want to achieve mastery (recommended).Matrix AlgebraMatrix algebra is an essential component of deep learning. I personally recommend this (Matrix Cookbook by Kaare Brandt Petersen & Michael Syskind Pedersen): http://www2.imm.dtu.dk/pubdb/views/edoc_download.php/3274/pdf/imm3274.pdf (PDF). There are 66 pages of pure matrix operations and this is the absolute “go-to” in case you are stuck trying to understand certain matrix manipulations that a researcher might have done.Probability & StatisticsUnderstanding probability is a very important aspect of understanding ML. Some of the key probability concepts that you must be aware of include Bayes’ Theorem, distributions, MLE, regression, inference and so on. The best resource for this is Think Stats (Exploratory Data Analysis in Python) by Allen Downey: http://greenteapress.com/thinkstats2/thinkstats2.pdf (PDF). This absolute gem of a book is 264 pages long and covers all the aspects of probability and statistics that you need to understand with relevant Python code.OptimizationThe go-to book for Convex Optimization is Convex Optimization by Stephen Boyd and Lieven Vandenberghe: https://web.stanford.edu/~boyd/cvxbook/bv_cvxbook.pdf (PDF). This is a 730 page book and you need not read it all in one go. Choose the concept which you need to learn depending on your requirements and interest and read that part. It is complete and extremely well written. This book is free as part of the CVX 101 MOOC on EdX.This 263 page book on metaheuristics, Essentials of Metaheuristics by Sean Luke (http://cs.gmu.edu/~sean/book/metaheuristics/Essentials.pdf (PDF)) talks about gradient based optimization, policy optimization etc. and it is well written. One can choose to go through this also if interested.Data science concepts are covered in the above topics. Other topics can be learnt by googling for sources easily as and when you encounter them. But complete understanding of the above should suffice for 95% of all scenarios.Achieving mastery of the above topics will surely make you a mathematically strong AI/ML engineer. Now that you have built the foundation, start dipping your feet into research papers. They are absolutely essential as these clearly show the standards of AI researchers/engineers. Firstly, find out the famous papers of AI like RNN, LSTM, SVM etc. and go through the technical content.Can you understand the jargon?Can you understand the mathematics?Can you implement the mathematics in code now without the help of overly sufficient libraries?These are the key questions to be answered. Once you can answer “Yes/Mostly Yes” to these 3 questions, you are good to go.After trying to read these papers dealing with the most popular concepts, try to read the not-so-famous papers. arXiv is a great site with hundreds of preprints being published everyday by top researchers and reading the papers from here is like drinking straight out of the fire-hose. Try to choose a paper which looks fairly well written and the abstract seems interesting. Then, read that paper and try to answer those 3 questions again. The same can be done with papers of top AI conferences like NIPS, AAAI, AAMAS, IJCAI, ICML etc. You may not be able to fully implement the papers due to data constraints and other issues, but if you are able to understand even 60% of the mathematical reasoning, then I can safely say you have completed your training.Do not concentrate on learning more and more “packages”. Concentrate on the concept. While implementing, you will automatically see that you require “this” package and then you will automatically learn to use it. Learning the various commands of random packages won’t help. If you start implementing and writing codes to solve problems or simulate results from a paper, you will automatically learn about packages and use them appropriately; they’ll be the least of your concerns. This is the correct way to maintain “balance” between math and coding. You can also participate in competitions (e.g. Kaggle or conference competitions) to improve speed, development and processing skills if you feel the need to do so.(All the links in this answer are working as of 6th July 2017)

I'm assuming behind your question is wanting to determine how to actually physically organize your books onto bookshelves. I'm also assuming the goal is to be able to find a book quickly, rather than, say, to please the eye.Here's my strategy, which not surprisingly is colored by my work in database systems. Think of each shelf as a disk page. You'd like to be able to find a book in one or two accesses (checking a shelf or two). Once you're at a shelf, it doesn't take much longer to scan the shelf looking for the title you want. So don't worry too much about ordering within a shelf. On average, you'll find what you want before you've scanned half-way through it. (Actually faster, because you might remember that the book was thick or had a green cover.) Also, searching from one shelf to the shelf above or below doesn't take much time compared to going to the first shelf initially. So, if you have a lot of books on a topic, put them on consecutive shelves.Now, the key point is that you don't need a fixed classification system across all your books. You can even have overlapping categories, as long as any particular book is likely to be in only one or two categories. I'll give a couple illustrations.My wife has 1000+ cookbooks. (She used to be a food writer.) Some categories are based on technique: baking, grilling, mixing drinks. Some categories are based on cuisine: French, Italian, Indian, Asian, Mexican. (Did you know that most Mexican cookbooks have red covers?) Still others are based on type of food: vegetables, seafood, pasta. There are a couple shelves for general cookbooks: The Joy of Cooking, The New York Times Cookbook. There are also a couple of shelves for really tall cookbooks and for cookbooks from specific restaurants. It seldom takes either of us more than about a minute to find a particular cookbook, plus the grouping works well for browsing. Want to find a recipe for cookies? Look at the baking shelves. Want a recipe for Carbonara? Check out Italian shelves and the pasta section.At my office, I've mostly got books grouped by subject: databases, algorithms, theory. Books that I teach from are on their own shelf, regardless of topic. Also a shelf of writing and reference books. Stuff I don't have much of all goes on the same shelf: AI, OS, VLSI. Conference proceedings are grouped by conference, theses are all together. Again, it usually only takes one or two "probes" to find a book I'm looking for (although it can take a lot longer if I'm searching for a book that I forgot I loaned out).Another suggestion: Put books you consult a lot at eye level. At home, frequently used reference books and travel guides are in next-to-top shelves.Extra note: One of the most creative organizing systems that I've heard of was a travel writer who shelved his books about places in the US in a wall of shelves organized roughly by the geography of the US: California books on the left, book about Maine in the upper right.

WELCOME TO THE HUNGRY GAMES (or how to keep me away from the grocery store and happy at home in my jammies)Note: This answer is probably going to be too exhaustively detailed for casual readers, but I hope it will be meaningful to people seeking to transform this industry. Ultimately, if it brings us all one step closer to a truly viable business model for home meal delivery kits, that would be worth way more to me than a $250 USD cash prize.Hi, I’m Annika. I think I want to subscribe to your service. Problem is, your service doesn’t exist.I am exactly the sort of person who would subscribe quite eagerly to a meal kit delivery system, and I believe very strongly that meal delivery kits could disrupt and fully supplant supermarkets on the near side of a decade if done well. That said, I’ve never purchased or tasted a kit meal, and have no plans to do so until the “right” one comes along.This is so gonna happen. It’s just a matter of time. And I’m getting impatient.Remember compact discs? Remember CD players? Remember music coming in albums? Remember paperback books? Remember bookshelves?Subscription meals are gonna be like that, mark my word, but for Harris Teeter and Albertson’s instead of Tower Records and Crown Books. Say goodbye to buying milk by the gallon, bread by the loaf, and bananas by the bunch, say goodbye to automatic self-scanning checkout that doesn’t really work. It’s all going away. It’s going to be supplanted by something that we all want much more than that, and lives will be meaningfully improved.The reason why the big supermarkets haven’t gone away yet is because nobody has figured out how to get the market what it actually wants better than they have. Which really isn’t saying a lot. Margins at supermarkets (I used to work at a supermarket) are super low. It’s tough to make ends meet because the systems in place are so extremely inefficient. All this could change, but it’s going to take some heavy lifting at first.First and foremost, this is a marketing problem. But it’s not a soft, bite-size “you just need to work on your social media personality” marketing problem like the kind that the parody Prof. Jeff Jarviss account on Twitter likes to joke about. This is a big, chewy, stringy marketing problem that has to be solved all at once. It’s at least a three-part problem, and as such is not easily detangled. Whoever does solve it will dominate the market—and it’s a big market, and worth doing.The bad news: it’s going to take a huge amount of upfront capital and logistical muscle like you have never seen before. But, mark my words, it’ll happen. And likely sooner, rather than later. The technology already exists. We know (sorta) how to get from point A to point B. If we solve the basic marketing problem, we can also figure out what sort of company might be able to pull all those levers. I have a few ideas on that, too.Three (or more) parts to the problem of getting the market (customer) what it wants:Who is the customer?(bonus) How do you get more customers?What does the customer want?(bonus) Can you get them to want something more than the basic “deal” and thus pay more money (i.e. can you upsell them)?How do you get it to them?(bonus) How do you get it to them without losing money hand-over-fist.(double bonus) How do you get it to them AND make money?Part 1: Who is the customer? — Hint: it’s me!Figure out how to get my dollars and loyalty, and your problems are solved.There are probably many other kinds of customers out there, but the pleasures and struggles I face in the kitchen are far from unique, and knowing a lot about me and how I shop for food and spend my time in the kitchen will help you to understand some big problems your industry could solve.Here’s a rough pie chart of my spending:As you can see, a huge amount of my planned monthly budget goes to online purchasing. I am very (very, very) frugal when it comes to cash purchases at points-of-sale, and I have concluded years ago that buying things online is the best use of my time and money when all is said and done. In making this calculation, I consider things like savings on gas used for comparison shopping, savings in opportunity cost driving from store to store looking for the best deal, competitive pricing between brick-and-mortar and online purchases, availability and variety of goods (branded vs. non-branded). If I could find out how to shift my food purchases to a digital platform and not have to supplement with brick-and-mortar shopping, I’d do it in an instant.I am one of Amazon Prime’s first customers, and I have spent, easily, 100k USD on Amazon purchases since 2007. I made the shift from shopping at Target and Walmart to Amazon on the day I went to Prime, and I never looked back.Despite my dogged loyalty to Prime, I have only rarely purchased groceries from Amazon in the past. Some reasons for this:Amazon does not (yet) have a competitive means to deliver loaves of bread, or milk by the gallon (more on this later)Amazon Pantry requires me to buy a whole boxful of dry goods and pay a (small) shipping fee. My box never gets filled so I never push the “buy” button.Amazon Prime is only cost-effective for buying a whole bunch of things at once. (Who needs 8 bottles of ketchup?)Food storage is an issue when buying in bulk. My kitchen is not that big.I don’t always know 2–7 days in advance what I want to eat. I could be watching The Great British Baking Show and decide that if I don’t get profiteroles within the next 24 hours, I’m never eating again.^^^Photo of Profiterole By Stu Spivack 2006 (Flickr) [CC BY-SA 2.0 (Creative Commons - Attribution-ShareAlike 2.0 Generic - CC BY-SA 2.0)], via Wikimedia Commons^^^I like to buy local, in-season vegetables and support local farmers (more on this later).Even where dry goods are concerned, Amazon does not stock all the niche brands I like at an acceptable price. I am a brand loyalist when it comes to condiments , and I think this is true for many people.^^^ Duke’s mayonnaise (my brand of choice) selling at 5x what I can buy it for at the corner shop—WTAF, Amazon???^^^I’m a bit of an ethnic foods snob. When I buy Indian jarred chutney, I like the kinds I used to eat in India. I like the fish sauce brands I used to buy in Vietnam. A key lime, to me, will never equal a calamansi. A tom yum soup is not worth eating if it doesn’t include fresh kaffir lime leaves.A week of grocery shopping for me could involve going to 5 different shops. One shop for Greek yogurt (I like the full fat kind, which is hard to find), another shop for bread, another shop for beer (I like really cheap beer), another shop for meat and veggies, another for (expensive) paper towels and (cheap) detergent.I buy a huge amount of frozen food because it tastes fairly fresh and is not quick to perish. Amazon does not currently support my frozen food habit.I am reclusive by nature: there is nothing about shopping trips that I particularly enjoy. If not for these strongly held beliefs listed above, I would gladly hang up my Harris Teeter loyalty card and never go shopping again.The second best thing (existing now) to digital/home-delivery shopping, is ordering curbside service at local shops. I do my shopping online, pay by ACH, and schedule a time for pickup. Not including the leisurely online process of filling my online cart, this process takes about 1/2 hour (ten minutes to drive to the supermarket, 10 minutes of waiting for them to roll out the cart and fill up my trunk, 10 minutes of driving home). Not optimal, but OK.All this being said, Amazon home delivery has several major potential advantages over the local supermarket curbside service: they have my payment and shipping information on file, they have sophisticated AI that knows what I like, and everything else I want to buy is there (hose adapters, shrink tubes, geranium seeds, board games, label makers, TENS units, auxiliary phone chargers), and (most importantly) I never have to leave the house or change out of my jammies to buy any of this stuff.This classic infographic from The Oatmeal on choosing movies is very similar to how I choose food, given any choice at all. Not having to put on pants is a big deal:It's time for a Netflix of food to save us from the travails of going outside and let us chill comfortably at home.Marketers, good ones, are always anthropologists. They look at how people live and figure out how to make products fit. I’m reminded of the household tours that P&G did around the time that Febreze was launched. Looking at user behavior teaches a lot about assumptions.I’m an actual marketing geek and an armchair anthropologist, and I’m gonna help you as best I can, by showing you what I’d be looking for in someone’s kitchen if I wanted to sell them (and people like them) a food product. Let’s do an anthropological, stream-of-consciousness tour, using my kitchen as an example, starting with my fridge, and see what we find:A bit about me, my house, and what we end up eating, and why:In our house, we cook a lot, we eat a lot, and we spend a long time thinking about what our next meal is going to be. We like food to be “special”.Oh my, that’s a tiny fridge.Freezer is packed. Overflowing. Mostly frozen fish and meats, packaged veggies, and 10 different kinds of really good chinese dumplings. Also an “emergency” pizza. I’m really worried about how much food I throw away, and having a well-stocked freezer is my best strategy (so far) to mitigate that problem. I’m very enthusiastic to find a better solution that involves a just-in-time supply chain and fresher foods. Meal kits seem made-to-order for that purpose.Refrigerator has a disproportionate number of condiments.Lots of condiments. Holy crow, that’s a lot of condiments. Why do I have so many condiments????Lookie here. More condiments!The big steel thing on the bottom shelf is some sad leftovers kept in a big pot that can’t be used for cooking because it’s not being used to its highest and best purpose.Ohhh, look. It’s a pot of homemade local ramps, bacon, and rigatoni. It looks gross and smells a bit like farts today, even though I just cooked it like 2 days ago and it was legit delicious at the time.Annnndddd…Down the drain it goes. I hate doing this. I really hate this. It had to be done. Nobody was going to eat it.Oh wow. there’s like 4 containers of lasagne on the top shelf of my fridge. It tastes really good but I honestly can’t bear to eat another morsel. I made enough lasagne to fit the cooking pan, and the pan was enough to feed an army. This was a half recipe! A quarter recipe woulda been better. I’ve been eating lasagne for days and if I never saw another casserole again, I’d be just as happy. I really ought to throw this away, but it seems wasteful so instead I’ll wait for it to become a science experiment and then I’ll toss it, container and all, because it’ll be too gross to do anything else.Some fresh veggies from the farmer’s market.My pantry in a constant state of semi-organized chaos. I try to keep things partitioned neatly in sterilite containers. It kinda works….OK, let’s be honest. It hardly works at all. I have no idea what’s in this pantry on any given day. I hate that all the products I buy come in different sizes and shapes and can’t be stored neatly. I’d buy everything from a single brand if it meant all my drygoods could be stocked neatly. I want them to look like this:….And that’s never gonna happen.My roll-out drawers for canned goods. Another good idea that only kinda works. I have no idea what’s lurking in here, but I do know that cans are never the “right” size for storage.I honestly have no idea what the eff is going on in this cabinet. Let’s close it and walk away slowly. Yikes.I’m very proud of my blue and white crockery and dishes. So pretty!!! I love serving things nicely and having my kitchen look neat. So hard to do, though.I have a ridiculously expensive professional model 6-quart Kitchenaid that I got as a gift and literally never use. I keep it around as a status symbol. It takes up usable space, but, hey, Kitchenaid! One of these days I’ll use it!My countertop herb garden, which is not just for decoration. I use this pretty much every day. If the herbs die, I put new ones in. The garlic races (sprouted garlic greens) are life-changing. I use aquabeads and fertilizer spikes to keep it growing, instead of soil. Works really well for the purpose, so I always have a few fresh herbs on hand.This is one day’s worth of dishes. I hate doing dishes. I hate doing dishes so much.In addition to my surfeit of condiments in oversized packages, I have a huge spice rack. I find my spice storage situation intensely frustrating. Again, why can’t everything come in an easily-stored package that has a reasonable amount, hermetically sealed? To wit: I have a container of cinnamon I bought 3 years ago and there’s still like half a thing left of it, but it’s old and tastes like sawdust. I can’t bring myself to throw it out, so it just sits there, taunting me.A small part of my cookbook collection. I don’t use it to cook with (all my recipes are from online these days), but I do use it for looking at pictures and reading descriptions of things I might want to cook and eat (using an online recipe).Where I make coffee and tea. This is one corner of the kitchen that really works as advertised. Most other parts of my kitchen are sources of unpleasant labor, wasters of time, and causes of frustration.Part 2: What does the customer (i.e., me) want?Here’s some things that would make me happy.Never having to do dishes again.Never having to go to the store again.Attractive serving containers that double as storage containers and cooking vessels.I buy this kind of tea at the Asian market because the can matches my dishes. There’s better tasting teas, but the can matches my effing dishes!!Storing less food, more efficiently.E.g., Spices should be blister packed into single-gram portions, like salt at McDonalds.Never having to throw away old produce (I counted three half-heads of different kinds of lettuce while doing my fridge inventory today; they’ll probably all end up in the trash).Getting fresh vegetables, milk, and cheese in the portions I need, the same day I order them.Preferably local.The “service” knows my billing and shipping information and knows what I like to eat (and when), and is able to recommend meals to my taste.Hint: When the first cold day of October comes, freakin’ EVERY DANG THING had better be tasting like pumpkin spice for the next two weeks.I never want to leave the house or get out of my jammies for any reason related to food.Instead of a stockpile of canned goods, I want smaller portions in clearly-labeled storage packs that can be stored neatly in smaller spaces. Preferably, these packs should be reusable and/or recyclable.I want my “service” to know what foods I have on hand, so I don’t double up on ingredients.Individuality/specialness: when I cook meals, I want to be able to add “personal” flare, while still keeping time/labor/waste to a minimum. I do want to be able to claim authorship over recipes and dishes, same as I claim authorship over pies made from refrigerated crusts.I would sometimes like to put some of my kitchen’s “professional” features to use (e.g. Kitchenaid mixer).I should not have to have the “right” size of pot or pan or tray to make a dish. Equipment provided by me should be either voluntary (Kitchenaid) or minimal.I would like my total monthly spending on make-at-home food for 2 people to not exceed 400 USD. I could be persuaded to pay a little more if it meant I was eating better food, shopping fewer hours, doing less clean-up, eating fewer leftovers, and wasting less. Let’s say 500 USD/month tops.There are certain basic-yet-time-consuming foods/ingredients that I believe should be available pre-cooked/pre-prepped unless specified otherwise. These include:Caramelized onions — or any kind of onion, for that matterRiceBechamel/beurre manieHand-pulled noodlesPie crustsMoleMirepoixPomegranates (seeded)Pineapples (peeled)BeansGarlic (peeled)Food should be packed for reasonable longevity and low-cost delivery, but the goal is fresh (local-oriented) food, every day.I’m equal parts lazy and ambitious when it comes to food. It should be interesting and (sometimes) challenging to prepare, but no effort to store or clean up after.Condiments should be brand name and authentic. Asian condiments should legit come from Asia. Mexican condiments should legit come from Mexico (or at the very least, from Bakersfield CA, where the best Mexican condiments come from, I’m told).Optimally, a modular food storage system should come with the plan, where I can see at a glance what my resources are, and what I need to buy more of. Think like K-cups, but for food that’s waiting to be cooked.A bit about my town (Asheville, NC) and how we live and eat:Eating, and eating well, is a big deal, a big part of our culture.A wealth of local food choices.Very few ethnic options. Even Italian food is hit-or-miss in restaurants.Food trucks are readily available and good.Many people are underemployed, casual economy.Logistical challenges: city is not particularly well served by distribution infrastructure (bike delivery is not an option).Why I’m not currently buying from one of the many food kit suppliers (Peach Dish, Purple Carrot, HelloFresh, Plated.com, BlueApron.com, Terra’s Kitchen, Marley Spoon, etc).Cost. I cannot afford to pay more than 500 USD/month for food for two people. That is my absolute red line, beyond which no purveyor of fine foods can pass. All meal kit plans I surveyed looked to be hovering a little above $10 USD/meal. For daily lunch and dinner for two, that adds up to over $1200 USD/month. Way too spendy.A really good local pork chop and a side of fresh vegetables runs about 5 dollars a plate here in NC. That’s what you’re trying to beat.Not local enough. When it’s ramps season here in AVL, I wanna eat some dang ramps. I want to support people here as best I can. Making ends meet is hard.Not truly labor-saving. For me, cooking is fun. It’s all the other stuff that’s awful. Buying things, driving around town, carrying groceries up stairs, bagging/eating leftovers, doing dishes, putting things away, storing things, throwing out old food.Nobody has truly saved me a trip to the grocery store. I’m still gonna have to buy stuff like laundry detergent, so why not pick up some groceries while I’m there.There doesn’t seem to be a clear (marketing) demarcation between a gourmet meal kit and something non-descript and starvation-inducing like (sorry) Nutrisystem. I want to be assured that my food is going to taste really good and be sufficiently satisfying, if I’m going to spend that much money.I’ve never tasted a kit meal. I’m going to need to taste a forkful of someone’s lunch before I commit.A “send a kit to a friend” promotion seems like a great way to solve this particular problem.I look at the weekly menus these websites have, and maybe I want to eat one or two of the things being offered, but I never want to eat all of them. It’s sort of like going to a restaurant and being at their total mercy—ordering one of everything on the menu and hoping for the best.Part 3: How do you get the product to me?The meal kit problem is a big, complicated problem. Many garage-oriented businesses have tried this and failed, right here in Asheville. They just don’t have the muscle to get off the ground. I suspect that all of the recent comers to this marketplace will either die on the vine, or get bought by someone much bigger and more powerful. The natural answer is to be a really well-developed logistical specialist with excellent data on my shopping habits and strong artificial intelligence capabilities.In rank order from most-likely to least-likely contenders for this space:You could be Amazon: You’re my first stop for online shopping. You already have my buying information and you know exactly what I like. Your distribution platform is highly-developed and coordinates multiple vendors, including cottage industries, seamlessly. Your negotiating power with delivery specialists is legendary. Localization is getting very granular with contract drivers and drone experiments.You could be Google: You have maps, self-driving cars, and my entire search history at your disposal. You know when I’m jonesing for a dang profiterole, and you probably guessed it ahead of time, by virtue of the fact that you knew I was streaming food shows on my computer. You have Youtube, and you could easily tie in a “buy this meal kit” button with youtube vlogger demos. Your “shopping” capabilities seem to be on a back burner for now, but I don’t see you leaving it there forever.You could be Uber or Lyft: You have an army of local drivers who can drop things off, piece by piece, over the course of a day, so everything’s staged and ready by dinnertime. You’re getting into self-driving cars, just like Google is. You know how to do surge/locality/availability-based pricing on the fly, which could come in handy. This is a weird brand expansion, but it could really work.You could be Blue Apron or HelloFresh: You’ve spent some time on getting name recognition (S-town anyone?), and you have an army of happy customers who tried and can presumably vouch for your product. You (hopefully, at this point) have some packaging/portioning automation infrastructure in place. What I’m recommending in logistics, though, is going to require a lot of capital up front. If you can get a very strong venture round in the near future, it’s not wholly unbelievable that you could expand into other stuff later on based on proprietary logistics knowledge, same as Amazon expanded away from books based on its extreme logistics prowess.That being said, y’all’s most likely exit strategy would be to sell yourself to one of the big guys eventually. But after you proved yourself, so your price could be pretty dang high.You could be Elon Musk: You’re worried about environmental degradation, and you think that automating the whole system of delivering meals in a zero-waste enclosed system is the answer. You have self-driving cars in your bullpen, and you are a genius with plenty of cash to throw at the problem, provided the ROI looks good. Risk doesn’t scare you—it excites you. You are well-seasoned in getting big heavy things off the ground in an efficient way.You could be Walmart: You’re an old dog, but you can learn new tricks. You are famous for being affordable (food stamps, government nutrition programs are obvious partnership avenues) and your website has everything under the sun that a person could want to buy. Your logistics are still the best in the world, and you are getting a little desperate to find a way to stop Amazon from overtaking your market share. This could be it. This could really be it!You could be Facebook: I mean you could be facebook. I don’t know why you’d do this, but I do know that y’all have way too much cash lying around, and you keep buying companies to get rid of some of it. I’m also pretty sure y’all know what I’m planning to eat for dinner on any given night, so that helps.You could be Apple: Same deal as Facebook. Too much cash on hand, decent AI, Steve Jobs would probably have approved, if his hypothetical meal kit had included sufficient wheat grass.You could be the U.S. Department of Health and Human services, interested in a low-cost way to combat “food deserts” in low-income urban and rural areas, and nutrition programs for kids and seniors. Meal kits might sound like a really good way of fixing an intractable problem with a flexible, scalable solution. Oh right. Tom Price. Never mind.Whoever decides to do this, you need to leverage a lot of things all at once, and be prepared to spend money to make money.You need to have a way to deliver food and retrieve/clean reusable food containers.You need a way to figure out what I want to eat, without me spending a lot of time figuring it out for you.You need to have a way to portion out foodstuffs that is efficient and not ridiculous. I read an article about how Blue Apron was doing this a couple years ago, and it was alarmingly inefficient and stressful:All told, interviews with 14 former employees describe a chaotic, stressful environment where employees work long days for wages starting at $12 an hour bagging cilantro or assembling boxes in a warehouse kept at a temperature below 40 degrees.“You put honey in a small container. We would put small peppers in little small bags,” said Glenn Lovely, who worked as a temp in the Richmond facility for three months. “And it was cold — cold as hell.”…Those ingredients are prepped by kitchen associates, who weigh out and perfectly portion bulk ingredients from Blue Apron’s suppliers into small plastic bottles and bags: tablespoons of soy sauce poured into tiny bottles, for example, or carefully counted fingerling potatoes put into boxes. And after the boxes are assembled, the shipping department loads them onto pallets and, ultimately, trucks.…“I would get sent to Whole Foods and buy things if we really needed an ingredient and we didn’t have it in the building,” said the former team lead. Blue Apron told BuzzFeed News that while during early days it sourced some of its product from local stores, the company’s shipments have been too large to make grocery store shopping feasible “for years now.”I mean holy crap, y’all. That sounds awful. A lot of things that clearly need to be done upstream by robots, are being done by low-wage employees whose fingers are freezing off. Buying stuff at retail at the last minute?? This is not sustainable! AI should at the very least have been used to predict “lumps” in ordering/preparation changes so nobody has to go to the store on short notice, if not to predict what an individual customer might buy on any given day. Reading this account of the struggles of Blue Apron circa 2015 reminds me very much of the (Root) Beer supply chain management game we played in B school. With a very large measure of Murphy’s Law applied. That’s no way to run a business!Any of the big logistics/AI/self-driving powerhouses I mentioned above could overcome these problems, and one of them really should.Conclusion—Here’s what it should look like from a supply side standpoint:AI-driven, simplified ordering, possibly app-based, informed by my other online activities.Kit “components” not necessarily being assembled at a single location. Can be dropped at my house throughout the day. Can be sourced locally and delivered either by humans (uber) or self-driving cars, or drones, or some combination thereof. I’ll keep a branded cooler by my front door so y’all can leave the meats and veggies there. Maybe ask Yeti to partner with you guys?I’m a big fan of hardware-plus-subscription services from an operational standpoint, not only because it simplifies shipping logistics (a lot), but because hardware-based contracts are often stickier and harder to break. I’m thinking about how many times I pondered changing my cable box service or cell phone carrier and decided it just wasn’t worth the time to turn in the equipment…Very well-thought-out packaging and storage. Food should be able to be cooked, stored, and served using supplied packaging. Packaging can be returned for washing/reuse when the next delivery comes (see hardware comment, above). My fridge should not be overrun with condiments, yet I should have condiments handy for anything I want to use them for, and they should be brand-name.Also do my laundry. (Just kidding) ((No, on second thought, actually maybe not really kidding.))Keep track of what I do and don’t have sitting around in my kitchen already.Bonus points if you can suggest things to make with things I have in my kitchen.Make sure I know the value of your product (labor-saving, satisfying meals, I don’t have to get out of my jammies, I can still cook and be unique and original, but no dishes).Recipes should be tailored for me and my tastes! An “Annika Schauer meal delivery kit” should not be identical to a “Michael Peacock meal delivery kit”. Maybe get recipes from IBM Chef Watson, so our individual styles remain individual, and stylish. At the very least, get our brands of mayo right (I like Duke’s, Michael likes Best Foods).Branding should be focused on visible collateral (food packaging), not on specific recipes (which should be tailored to individual tastes/seasons/dietary restrictions, and should not ever be limited to a weekly “menu” of only seven options, no matter how delicious those options might be.)Get busy with partnerships on all levels. Include brand-name items in kits (like Tabasco in MREs) and market directly through sponsorships on how-to cooking sites. Make kits for every recipe on ChowHound and Food.com. You get the picture.Get an “app” style marketplace for recipes, instantly translated into orderable products. An IFTTT-type dingus could easily be used for this purpose. Give recipe authors a small cut of profits on foods ordered, possibly in redeemable meal-kit credits.Get me a profiterole, or the means to make one quickly, whenever I see one on TV and get a random craving. :) Possible ad text: “Annika, you could be eating a profiterole like this one, in the comfort of your own home, by 8:00PM tonight, if you order now!”Put my food delivery order in the same online marketplace as other convenience items I’m going to need, like detergent, Ziplock bags, deodorant, etc.If I get a meal kit and I tell you I like it, give me an easy way to order that exact same meal kit again.Not all meal kits need to be delivered the same day I order them, but some same-day options should exist (replacing the “emergency pizza” from my freezer).Best case scenario, kits should include sourcing for hard-to-find ingredients, so I can make an authentic pecan pie from cane syrup at home and not have to act like corn syrup is the same thing.Special “group buys” could be a way to both source hard-to-find ingredients, and recruit new users. The “bus” leaves the station when enough passengers get on board to make an order affordable.Deliver all of the above for a price comparable to my current actual food/sundries spend. (If you get the packaging logistics sorted, you should be able to match this price based on volume and still make a reasonable profit, since you have no customer-facing brick-and-mortar retail space to maintain, no parking issues, minimal spoilage, and no shoplifters to chase away.)The business case for this plan:It’s clearly not a small project, but whoever grabs the reins on this endeavor gets, in addition to any marginal returns from the business itself, a treasure trove of demographic data more personal than even Google and Facebook can see right now. The kind of data that can target and sell things to people on a very granular level that’s not been seen heretofore. You also get a sticky customer base that will not likely drift to other sellers, and a huge window of opportunity to upsell.Like Amazon was never really just a bookseller, whoever picks up this one will not be just a purveyor of dinner. They’ll be a logistics and data collection company that sells food as a means to an end. Probably a large and impressive end. The more skin you’re willing to put into the game up front, the more impressive the end will be.All of this is doable, y’all. All of this is doable in my small town of Asheville, population <100k, tucked in a hidden corner of Western North Carolina. All of this is doable anywhere with internet, food retailers within driving distance, casual-economy workers, and cold chain. Which means pretty much the entirety of the continental United States, to start.All of this is worth doing, and worth doing well.Look at me: I’m a hillbilly. If you can get me to stick a “dash” button next to my dishwasher and a virtual “dash” button on my phone, if you can get a pair of pliers to my house on a Sunday afternoon when I order them on a Saturday morning, for a price that is comparable to what I’d pay at the Ace, if you can get me a seat in a car to anywhere in the city from my front door in less than 2 minutes, if you can sell all of my personal data, at a nice premium, to third parties, I think you’ll find it worth your while to save me the headache of shopping and doing dishes. And for that I thank you, from the bottom of my heart, in advance. Hopefully not too much advance, because I can’t wait!!!…And may the odds be ever in your favor.Edit: I found this article today, which sort of hints at some of what I’m talking about. McDonald's and Uber Just Made a Huge Announcement, and It's Going to Be a Game-Changer