Three Reflections on Modernist Cooking: Convenience Stores, the Limits of AI, and Blenders

This has been, by my math, the longest stretch I have gone without posting something here. And trust me, I have good reason: I have been launching a business! I am now the owner of Opossum House Accessibility, which is my vehicle for providing bespoke accessibility consulting services for public and private clients. Launching a business is hard, but has been supremely fun. Subscribe to the newsletter here – I plan to write something in the future about how food blogging gave me skills I applied for launch.

Besides that, I have also been traveling. My husband (love that word!) and I went to Japan and South Korea on our honeymoon in May and June, and we have also been traveling on the Eastern Seaboard of the United States. 

Anyway, enough prattle from what I am doing. You are here to read about food!The travels and business startup process have given me a lot to “chew on,” metaphorically and literally, about modern food and what modernist food is and is not. So, I want to share three reflections on modernist cooking.

Japan and South Korea do modernist food really well. 

I am very much not the first person to write about the wonders of convenience stores in Japan and Korea. In these chains, you can get simple, reasonably healthy, and traditionally-rooted dishes for very cheap. Favorites include onigiri (Japan) or samgak gimbap (Korea) – rice balls with fillings, various noodle salads, and filled buns. While we definitely had “nicer” meals too, the stores were helpful for snacks or after a long day of sightseeing. 

On both this trip and a past trip to these countries in 2019, I found myself thinking about how these stores exemplify what Rachel Laudan calls for her in her seminal article about culinary modernism: that we should advocate for cheap, high-quality processed food for everyone – not to undo processing. (I have written about this at length on these pages.) While 7-Eleven is making some moves towards this in the United States, I think these Japanese and Korean stores give us in North America a lot to think about. These stores also made me wonder about how these tastes have then affected Japanese and Korean cuisine more broadly. Do the wares of konbini in Japan and pyeonuijeom in South Korea change what people seek to make when they are at home? This is something that, despite the language barrier, I want to learn more about. 

Generative artificial intelligence (AI) is overrated – especially in the kitchen. 

When you start a business nowadays, people want to talk to you about generative AI. This is especially true in accessibility, where a lot of people try to apply generative AI badly. I was already a bit suspicious of claims about generative AI, but decided to give it a little bit of a shot because it seemed money was there. After talking to Generative AI enthusiasts and my own research, I am now more skeptical than before. (As a result, I chose to ignore the misguided advice from more than one person to spend an hour a day fooling around on ChatGPT or Claude.)

Because people mean lots of things when they say “AI,” I will be more specific. I am talking about predictive large language models that generate content; for example, ChatGPT, Claude, or Gemini. These are prediction-based tools that use “the most likely” thing to generate something. When you state it like that, the weird claims about generative AI sound a lot less plausible.

(I acknowledge there are some limited use cases that work well. For example: the designer I hired for my logo used AI models to replicate the beautiful logo she designed across business cards and letter heads. A friend of mine used a large language model to help people find the correct lawyer for their needs. Image description AI tools, which can come from multiple forms of AI, have been helpful for my blind friends. But the first is a talented artist using a tool to save time in applying her skills, and the second has many sources and is not usually reliant on large language models.)

I will link some great AI skepticism below. Anyway, I was originally going to write about how generative AI is actually a problem when you apply it to cooking, but Joe Ray at Wired published that article last week. He did the work of talking about the problems of asking generative AI for recipes, so I do not have to. Definitely go read it! It makes many of the points I wanted to make, and more.

In any case, I also spent time – especially on long flights and Amtrak rides – thinking about how people use generative AI in the kitchen. Apparently, people use it to “save time” with knowing when food is going off or to use food, plan their meals and grocery shopping, look for recipes, or figure out what to cook. 

I think there are tons of problems here. Ray goes in about them in the food context, and other people – especially Ed Zitron and Baldur Bjarnason – have written about these issues in other contexts. But for cooking, I see two big problems. One is that cooking is embodied. Many of the things we do when we cook, we do through physical sense and actions that we take without describing well. How we beat an egg, see that something is browning, or how my blind friend listens to hear if a sponge cake is fully cooked. Generative AI does not have a body, much less knowledge one gains through a body. So the “advice” it spits out is already suspect.

The other is that cooking is inherently unpredictable. What happens when, as occurred to me last week, you cut into a cucumber to find a worm, and you need to rejig the salad you made? Or when you accidentally spill too much salt into your soup? AI usually looks for the “most likely” thing – but sometimes, we need to take unlikely steps in cooking. I worry about what happens to someone’s ability to cook and feed themselves when they become reliant on a tool that cannot handle the unpredicted. (And as an autistic person, I know surprises are hard! But they are part of life.)

Friends have also made a point that much of the dependence on AI is a direct consequence of the parlous state – and degradation – of search platforms, something that I have seen as I have written this blog over the past decade. Google and Bing, for example, both return far less reliable results than even two years ago. People are looking for something that seems useful to them – even if, as in this case, it causes more problems than it solves.

Modernist cooking is lots of things, but this use of AI is not one of them. At best, it is gimmicky. At worst, it becomes another way to pressure people to spend more time, energy, and money cooking than they can afford. What I see most vis-a-vis AI and food is that it is papering over an unhealthy or unsustainable relationship someone has with food. The problem is not that you need to better track your produce or plan your meals – the problem is you are trying to cook or eat in such a way that is not working with the way you live your life. And there is no shame – and it is probably better for your health, the planet, your wallet, and your well-being – if you choose to eat some more processed or prepared foods, or eat that sandwich, or do something simple, because that is what you have time for. And there are many established, low-tech, and more reliable ways to do this tracking. (I use a notebook.)

And if you need recipes, why not go to the original source? Recipe writers are humans whose work deserves to be supported. The best way to get free recipes, besides food blogs, is to support your local library and check out a few cookbooks – which you can now even do online. These books’ recipes are tested with the embodied knowledge that AI can never have. Here in Greater Washington, I have immensely enjoyed the cookbook collections in both Montgomery County and Washington DC’s library systems.

The blender and food processor are modern miracles. 

My business startup period has coincided with a renewed love of Southeast Asian vegetable salads. These are magnificent, hearty creations that feature shredded vegetables and fruit, often with tangy dressings and tofu or even fish for heft. (Vietnamese mango salads are a particular favorite.) While traveling, I was also lucky enough to have many delicious things that prominently feature grated carrots – fritters, pickles, noodle dishes, and even desserts. Grating or julienning by hand is a slow, dangerous process – and I am slower with a knife than most people. And besides that, I do not always have the time to do such an intense chop – especially with all of the tasks of getting a business launched and starting business development. My workaday, mundane food processor and blender have been a lifesaver. I can satisfy my cravings, safely, and do it in a reasonable amount of time. This machine is a win for society, not a cop out.

I have also, after nearly 34 years on this planet, finally come to truly understand why people love smoothies. Not as a meal replacement, but it is nice to have something somewhat heartier than my typical coffee (normal or decaf), tea, or sparkling water to sip on. It is especially comforting while I am trying to learn QuickBooks Online. Now, hearty drinks have a long history – in Viking Age Scandinavia and pre-colonial Mesoamerica, hearty grain-based drinks were very common. But the smoothie as we understand it now, with pureed fruit, yogurt, and anything else, is completely enabled by modern cooking equipment such as a blender. The miracle of cooking in 2025 is not a predictive model that can tell you to combine tarragon and fennel to flavor your pasta (pro tip: do not do this), but the fact that I can plug a machine in that spins a knife and liquifies a mango for me. What a time to be alive. Baruch Hashem.

Read more after reading my scribbles

And now, some resources for each of my points:

1. You can learn more about convenience stores in Japan from this article in the Tokyo Weekender and this book chapter, if you have access through an academic publisher. You can learn more about Korean convenience stores from this article from CNN. If you like Rachel Laudan’s article, check out her magnificent book, Cuisine and Empire.
2. AI skepticism is hard to find amidst the absolutely monstrous amount of propaganda for Generative AI we see today. I recommend looking at work by Baldur Bjarnason, Ed Zitron, Nik Suresh, Allison Morrow, Bryan McMahon, Edward Ongweso Jr., Emily Bender, and Alex Hanna. I have heard good things about Karen Hao’s new book, Empires of AI, but I have not had a chance to read it.
3. I found a cool history of blenders (PDF) from Purdue University’s Extension Service. Also, given I mentioned the Viking Age and Classic Mesoamerica, I have two archeology books to recommend to you. Children of Ash and Elm, by Neil Price (for the Viking Age) and Collision of Worlds, by David Carballo (Mesoamerica) are some of the best books I have ever read, and changed the way I think about certain parts of food history. 

Thank you to my husband, David Ouziel, for marrying me, traveling with me, putting up with my increasingly unhinged rants about AI hysteria, and eating my green mango salad with gusto. Thanks to Emma Greenstein, Mikaela Brown, Michael Faccini, Jonathon Epstein, Dexter O’Connell, Maryam Sabbaghi, AJ Faust, Matthew Marcus, Benjamin Gammage, Joe Conrad, Rachel Ouziel, and Jad Atoui for talking through some points in number 2 with me. Thanks to longtime readers Alex Strauss, Aaron Rubin, and Adelin Travers for taking us on wonderful food adventures in Japan.