In December, the actress Sarah Snook, best known for playing the icy Shiv Roy on Succession, commented just one word on an Instagram post by NYT Cooking: “Meatball.” And who could argue with that? Ali Slagle’s Thai-inspired chicken meatball soup looked good, and getting the recipe required only that one leave the word “meatball” in a comment.
Do so, and a message from NYT Cooking pops into your inbox in seconds, offering a direct link to the recipe. This new format for engaging readers circumvents the clunky “link in bio” maneuver, a workaround necessitated by the photo app’s incompatibility with clickable links in captions and now considered the norm for publications and creators who use the platform to promote work that lives on other websites. Recently, a slew of new add-ons — including Manychat, which NYT Cooking uses — has allowed creators to automate messages and replies in this way. Food52 uses them too, as do recipe developers with unwieldy follower counts, like Yumna Jawad of Feel Good Foodie (4.7 million) and Deb Perelman of Smitten Kitchen (1.8 million).
Influencers and creators have taken advantage of automation like this for a while, whether it’s to send followers recipes or to share shoppable affiliate links. The effect is twofold, saving individuals from the tedious act of manually responding thousands of times and guaranteeing higher engagement since it prompts people to leave comments. These tools have become common enough to have instilled a habit: Some people now attempt to trigger chatbots even when a creator doesn’t use them or instructs other steps for getting recipes.
“It doesn’t actually matter as a content creator/pusher whether you use the bot thing — it’s so standard now that people assume you do,” Perelman of Smitten Kitchen told me in a DM. For viewers, these tools are easier and less confusing than asking people to click the link in her bio. “The actual conversation I had with myself was, ‘Am I going to ignore hundreds of comments a day like this, or am I going to cough up $100/month(!) to give people what they want? With social media, the latter is my default — just make it easy; meet people where they are.”
It’s true: Recipe developers and creators use these tools because Instagram isn’t the best place to share their recipes. Dropping instructions and measurements into a caption is easiest for viewers, but for creators, that means losing the potential revenue and the boosts to their engagement statistics that come from someone clicking through to their blog or signing up for their newsletter. However, since it isn’t in Instagram’s best interest to direct people to leave the app — or empower them to do so easily — the workarounds for highlighting off-platform content are annoying. Today, many people still don’t understand their way around a “link in bio,” even though the strategy has been in use since around 2018. Thus, recipe-sharing chatbots have emerged and taken hold. Do creators like them? Not necessarily. Do users? Begrudgingly.
For the people who use them, these automation tools are a new necessary evil, just like being beholden to the whims of an algorithm. At best, these tools ensure that creator and commenter both get what they want. For one, that’s a click; for the other, a recipe. At worst, they undermine the social nature of social media and depersonalize the experience of sharing food online.
Jesse Sparks
I went to Instagram — where I post pictures of food, pointedly without recipes — to ask food creators for their thoughts on these tools. The responses were overwhelmingly negative. “Yes I hate it if that’s strong enough of a sentiment,” said one. “HATEEEEE,” said another. “I HATE IT,” said yet another. Non-creators felt strongly enough that they had to write in too. Words that came up often were “scammy” and “desperate,” and some people resented them for being too obvious a play for engagement. Indeed, in one ad, which claims “No Follow. No Freebies,” Manychat promotes that it allows creators to “request a follow” before they “give away content.”
A common throughline was the idea of transactionality. “On a deeper level, as a content creator who puts a lot of thought into how I create my recipes and corresponding content, I don’t want people to simply think of me as a robotic recipe mill, constantly churning out recipes for consumption,” Lisa Lin, who runs the blog Healthy Nibbles, told me. “An automated tool seems antithetical to that sentiment,” she added.
This has long been the situation with food on social media. Get enough eyes on a picture of food online and you’ll certainly become familiar with the “recipe?” commenter. Not all pictures of food warrant a recipe, and not all people who post food are recipe developers; sometimes, the point is just to be proud of a nice lunch. Yet the “recipe?” commenter sees no distinction between the professionalism of a published recipe meticulously shot and developed, and the individual’s personhood, preserved and savored. At best, it’s a well-meaning follower’s detour into modest annoyance; at worst, it’s the prelude to a total internet stranger becoming put out when a poster doesn’t provide on-demand service, tailored to every need. In 2022, The New York Times’s Tejal Rao wrote of this phenomenon, coining it the “endless torment of the ‘recipe?’ guy.”
The core intentions of the “‘recipe?’ guy” are rarely bad: Isn’t a desire to imitate a compliment? Yet their assumptions speak to a sense of entitlement around recipes and theto cooks for providing them. With one word, that request turns a shared appreciation of food into a transaction, regardless of whether its creator intended for it to be or if they even benefit at all financially. “It’s a way of treating the people who share their cooking online entirely as products. But I think it’s also a way of becoming a bit less human,” Rao wrote.
Indeed, this use of chatbots and automation tools only accelerates the normalization of treating people who share food online like robots themselves. Automation tools reward this behavior. They make it normal to drop a one-word comment to a stranger, like a caveman grunting a demand, without any effort toward etiquette or building a rapport. They reinforce the notion that creators must always provide, as well as the problematic sentiment that whatever we see on our screens should also be available for us to have.
“I’ve worked so hard to build a community,” said recipe developer and creator Erin Clarkson, known as @cloudykitchen. She chooses not to use automation tools, in part because she feels they detract from the conversational vibe she works to foster on her platforms. “A chatbot destroys comment sections,” Clarkson said.
That sentiment was echoed in the responses I got on Instagram, especially from non-creators. It used to be funny or helpful to read the comments, where people made jokes, shared their candid reactions and experiences, or asked clarifying questions. Now, as people seek to trigger auto-response tools, it’s useless. We might see this as yet another example of enshittification: a once-social space optimized in favor of efficiency, but ultimately resulting in a worse experience for the people using the product.
To Clarkson, these tools have also made readers “even more lazy.” Clarkson says she regularly sees readers’ assumptions that she uses them, even though she doesn’t. She sees those presumptive comments another way: If these people can’t bother to read the captions to figure that out, then they likely won’t fare well with the level of detail on her blog. Everyone wants things instantly and easily, and recipes are no exception.
Still, these tools remain a “stopgap,” Lin said. Despite her ideological hesitation to tools that encourage robotic behavior from both creators and their audiences, the reality for her and most other recipe developers and food creators is that she “primarily earns a living on a website outside of Instagram. At the end of the day, I need eyeballs on my website,” she said. Having now subscribed to one of these tools for several months, Lin has found that they’re useful in getting people to visit her website. (Even when it comes to the established link-in-bio system, “many, many people can’t be bothered.”)
“If Instagram would simply allow us to embed clickable links in our captions, we would not need this ridiculous workaround to deliver links to our audience,” Lin said. “This automated recipe-sharing ecosystem wouldn’t even need to exist. But I don’t see Instagram developers changing their ways any time soon, so we’re all stuck in this situation.”
After hearing the malaise of social media users on all sides of the issue, I returned to the prompt that started it all. Committed to testing it out, I, like Snook, commented on that NYT Cooking post. Immediately, it felt silly — not just to comment “meatball” publicly, but also to add to the mindless cacophony of requests and to masquerade as yet another someone who didn’t bother to Google or search NYT Cooking. Afterward, I felt weirdly embarrassed: What friction was I really removing from my life by commenting?
Sure, the recipe ended up in my inbox immediately, but then again, my mess of DMs is where useful information goes to die. The instant access didn’t make me any more likely to make the recipe, and in fact, it would take me an awkwardly long time just to find the link in my inbox if I were in need of it while planning out dinner. I thought about all the recipes that have piled up in my saves on Instagram and in my screenshots folder. So many of them came to me so easily, offered up by way of too-knowing algorithms, and yet, I’ve never made most of them either. We now have access to so much information that we take its abundance — and the work that went into creating it — for granted. We see recipes as commodities that we are owed by virtue of us simply having seen them, even when we don’t have any intention of following through.
I thought about the technique that always works better for me anyway: just googling ingredients I have and then seeing how other people have already put them together. It makes me think a little more, of course, but especially in the age of AI, the most humanizing thing is to do a little of the work yourself — to have to think through a problem. I end up with something that’s all mine; not something anyone willing to just comment “meatball” can reproduce.
