In stressful times, one of my favorite acts of self-care is to hunker down behind my desk and binge-watch TED Talks, riding the swells of emotional well-being that come along with, say, learning how David Blaine held his breath for 17 minutes underwater, or uncovering the secret power of introverts. It was in this context that I first discovered Neil Grimmer, the CEO of Habit, a company that aims to liberate humans from fad diets forever, using a concept called personalized nutrition.

Grimmer took the stage at TEDx San Francisco clad in thick-rimmed glasses and black jeans, a pair of tatted sleeves descending from his black T-shirt. The dad-punk aesthetic belies his past as an organic baby-food tycoon; Grimmer is the founder of Plum Organics, the leading purveyor of vacuum-packed kiddie purees, which he sold to the Campbell Soup Company in 2013 for a cool $249 million. But the topic at hand wasn’t what babies eat (or even what babies think, another fascinating TED Talk). Rather, it was Grimmer’s own personal health journey, which follows an arc that has become a trademark of the self-help genre.

Grimmer was running one of America’s fastest-growing food companies, and the strain began to take a toll on his health. The year 2013, when he sold Plum Organics, was rock bottom. Grimmer had gained a lot of weight. He was prediabetic. He took stock of his life and did something that most of us non-baby-food-millionaires could only dream of: Grimmer traveled the world, submitting himself to the most sophisticated battery of genetic and metabolic tests available, and then enlisted a team of experts to translate that information into a diet.

Within six months of his new regimen, Grimmer reached his goal weight, regained his energy, and restored his blood work to healthy numbers: the holy trinity of modern dietary victory.

Ever since, Grimmer has been on a mission to give others the same experience. If he could harness the power of his own biological data to design the perfect diet, why couldn’t everyone? That, in a nutshell, is the promise of Habit: to collect the right biological inputs and convert them into a perfectly tailored diet. “Powered by big data and computational biology, for the first time ever we’re able to tap into nutritional insights that live inside our DNA, our blood work, our gut microbiome, and even our metabolism,” Grimmer told the TED crowd.

It’s a tantalizing premise: What if for millennia we’ve been groping blindly around the edges of human nutrition, making crude food decisions based on one-size-fits-all dietary research, when the the precise cocktail of nutrients needed to optimize health and longevity has been with us all along, written into our very DNA?

It seems remarkable, when you stop and think about it, that humans have managed to invent self-driving cars and tiny robots that can do surgery before we’ve nailed down what counts as a healthy lunch. Seven billion eaters on this planet — untold trillions of meals served since the first Homo sapiens carved their first stone blade — and we’re all still grasping at dietary advice that is both ever-changing and flatly contradictory. Low fat, low carb, low sugar, paleo, keto, vegan, gluten-free: They all work for someone, but no single diet works for everyone. Few of us feel we’ve cracked the code when it comes to eating for weight loss, let alone optimal health.

“It’s not inconceivable to imagine walking up to a vending machine, and through Bluetooth and a smartwatch carrying your genetic information, the vending machine is able to dispense a snack tailored your genetics.”

Today, nutrition researchers are coming around to the inconvenient reality that two people can stick to the exact same diet and experience vastly different results (an idea that anyone who has ever had a diet buddy will accept intuitively). Christopher Gardner, a nutrition scientist and professor of medicine at the Stanford University School of Medicine, tells me there is now a solid body of clinical evidence — including his own recent, high-profile study into the efficacy of low-fat versus low-carb diets — to support this phenomenon. “If you get a group of people eating almost exactly the same thing, they don’t respond the same way,” Gardner says. “There’s massive variability. It’s just staggering. One would think that given that variability is there, there’s a way to explain it.”

One potential explanation currently gaining traction: that variations in our genome determine how each of us processes the nutrients in food.

Consider, for instance, the Inuit people living in the Arctic. Scientists have long puzzled over the fact that, despite eating a diet consisting almost exclusively of fatty fish and meat, the Inuit have a relatively low incidence of heart attacks — an observation that flies in the face of conventional nutritional wisdom and has helped get everyone popping omega-3 supplements for heart health. But a 2015 study published in the journal Science identified several gene variants in the Inuit population that appear to help them break down fatty acids differently than the rest of us. That finding suggests that for the majority, downing omega-3s may have no positive impact.

Caffeine is another example. Coffee toggles on and off the naughty list, with some studies celebrating its health benefits while others raise red flags about heart risks. In 2006, researchers at the University of Toronto pinned down a single gene that affects how caffeine is broken down. They found that people with a “slow metabolizer” variant — about half the population — were at greatly increased risk for heart attack when drinking four or more cups of coffee per day, whereas those with the “fast metabolizer” variant could safely consume higher doses of caffeine. (Side note: Whether or not you are a fast metabolizer of caffeine has no relationship to how long you can feel its effects.)

The examples go on. Researchers have identified genes that influence whether you’re the sort of person prone to deficiency in a number of vitamins and minerals (including vitamins C and D and folate), whether sodium makes you hypertensive, and how you tolerate lactose. There’s even a gene that appears to determine a person’s chances of losing weight on a high-protein diet. It all adds up to a pretty compelling explanation for why nutritional studies so frequently contradict each other.

“Pick your nutritional factor, pick your health outcome—if there have been enough observational studies, you tend to see results all over the place,” said Ahmed El-Sohemy, a prolific nutrigenomics researcher at University of Toronto, in a recent lecture. “We used to call these people outliers and try to figure out how to get rid of them from the statistical analysis. But more and more, we find these so-called outliers are consistently there.”

The Habit test kit costs from $299 up to $449, if you spring for an additional “four weeks of digital guidance and tools to accelerate your adoption of new habits” (“BEST VALUE,” the website promises), and comes in sleek black packaging designed with unboxing pleasure in mind. A customer first fills out an online survey with information about her physique, activity level, blood pressure, etc. Next, she takes cheek swabs to send off for DNA testing. But what Grimmer and Josh Anthony, Habit’s chief science officer, consider to be the kit’s secret sauce is the “challenge beverage”: a 900-plus-calorie cocktail of carbohydrates, proteins, and lipids meant to simulate a large American breakfast. A fasting blood draw followed by two more at 30- and 120-minute intervals after gulping down the shake measure how efficiently your body processes each macronutrient.

Back at Habit HQ, the more than 70 biomarkers collected in the testing process are crunched through proprietary algorithms. Six to eight weeks later, results become available on the company’s website or its snazzy app. The customer finds herself characterized as one of seven “habit types”: archetypes that determine what percentage of her calories should come from fats, carbohydrates, and proteins.

That, in a nutshell, is the promise of Habit: to collect the right biological inputs and convert them into a perfectly tailored diet.

A “Protein Seeker,” for instance, could be advised to get 35 percent of her calories from protein, whereas the optimal diet for a “Plant Seeker” might include only 20 percent. Habit’s readout also includes personally tailored recommendations for things like calorie count, sodium intake, and recommended daily doses of a range of vitamins and minerals. A list of “hero foods” — items like salmon, Swiss chard, almonds, blueberries — highlights ingredients that Habit believes to be a particularly strong fit with your biology.

Across all habit types — and all the individual test results I saw — the dietary recommendations seem sensible and of a piece with conventional notions of nutritious eating. No one’s hero foods are pork rinds and vodka. Most people following their Habit recommendations would find themselves eating more vegetables, fiber, and lean protein, although a paleo adherent might be counseled that his body in fact has a hard time dealing with proteins, which means he’s better suited to a higher-carb diet.

I spoke to three Habit customers I found via Twitter — all self-professed data geeks who were as eager to dive into the raw results from the company’s testing as the ensuing dietary recommendations. All tell me they’re glad they did the kit, and though the results haven’t drastically changed the way they eat — or resulted in weight loss — the process has nudged them to be more vigilant about things like incorporating fiber or limiting caffeine. One customer notes that the results of her metabolic challenge, which suggested that her body was particularly good at handling sugars, seemed to validate her childhood attraction to sweet foods. The dominant complaint I hear is a shortage of transparency in terms of how the biomarkers inform the final results. “I think there’s an amazing education piece they could add that would connect the dots between the biomarkers and the diet,” Shane, a 41-year-old design director from Washington, DC, tells me. “Being able to share more of that information would be a huge benefit.”

And there’s the rub. Habit views the algorithms it uses to convert the raw data into recommendations as too valuable to share. Fair enough — it’s a business, after all — but it means the company’s methodology cannot be vetted by the broader scientific community.

More problematic than secret algorithms, however, is the absence of study data to indicate whether Habit’s approach is actually effective. The company stresses that its approach is evidence-based, meaning that it has been formulated with the input of scientific literature; however, that isn’t the same as validated, meaning that it has been tested through established scientific methods, like randomized controlled trials. Habit is currently in the beginning stages of a clinical trial, which may eventually add some scientific heft to the company’s promises. But for now, the company can only point to an eight-week pilot study of 20 subjects. And according to Frank Hu, chair of the nutrition department at the Harvard School of Public Health, the science on all this might not be ready for prime time. “There has been a lot of hype about the benefits of precision nutrition,” Hu tells me. “I think there’s a disconnect between the science and the commercial application. The principal is appealing, but the evidence is limited.”

Habit’s core product launched in January 2017, well under two years after the company came into being; a short runway for a highly technical, and novel, health product. But that “build the plane while flying it” approach makes sense when you consider its pedigree. Founder Neil Grimmer is an accomplished packaged-foods entrepreneur, but he doesn’t come from a science background; in fact, his academic training is in conceptual art and sculpture, as well as product design. The company’s chief science officer, Josh Anthony, who holds a PhD in cell and molecular physiology, is a food industry veteran who has spent much of his career working on infant formula. What’s more, all of Habit’s venture funding to date — a hefty $32 million — flows from the Campbell Soup Company, makers of tomato bisque and Goldfish crackers. That’s the rub, too: Habit isn’t a tech company. It’s a food company.

Habit’s approach may be, let’s say, overeager, but that doesn’t necessarily discount the potential for precision nutrition to transform the future of food.

Ahmed El-Sohemy, the University of Toronto researcher behind that caffeine study, is one of the scientific community’s true believers in the potential for genetic information to upend the way we think about eating. El-Sohemy’s lab has produced dozens of studies investigating the link between genetics nutrient response, and in 2011, he founded his precision-nutrition company, called Nutrigenomix.

Nutrigenomix differs from Habit in several ways. It’s available only through health care professionals — no flashy Facebook ads, no unboxing experience, no app — and is currently used by the Cleveland Clinic in its nutrition counseling practice. Where Habit collects vitals, metabolic results, and genetic information, then crunches it all into a ready-made nutrition plan, Nutrigenomix offers more-specific recommendations based on the results of a 45-gene panel.

El-Sohemy says his goal with the test isn’t to recommend a particular eating plan but to help people sift through the overwhelming dietary recommendations they’re faced with to determine their biggest nutritional vulnerabilities. Maybe you need to watch out for sodium-induced hypertension; maybe I need to eat plenty of vitamin C. “Skeptics argue that we know what a generally healthy eating pattern looks like — rich in plants and lean meats — but a tiny fraction of the population consumes that,” El-Sohemy explains. “There are different components to that prudent dietary pattern, and the genetic test can help you figure out what aspects of that you should really focus on.” His research also suggests a promising psychological component: that people are more likely to change their eating behavior as a result of personalized nutrition guidelines than general ones.

“I think there’s a disconnect between the science and the commercial application. The principal is appealing, but the evidence is limited.”

El-Sohemy says that as the science progresses — eventually, beyond individual recommendations to a more holistic understanding of how each of us should eat — he can imagine vending machines reminiscent of those Coca-Cola Freestyle contraptions, only designed to spit out genetically customized snacks. “It’s not inconceivable to imagine walking up to a vending machine, and through Bluetooth and a smartwatch carrying your genetic information, the vending machine is able to dispense a snack tailored your genetics.”

But why stop at vending machines? Soylent, the meal-replacement drink of choice for Silicon Valley brogrammers, seems a prime candidate for trying out genetically customized formulas. Prepared-meal companies like Munchery could end up crafting personalized options. In fact, when Habit first launched, it began by doing exactly that: shipping out customized meals to Bay Area users. That offering is now on hold while the company gets its arms around scaling up the service nationally.

Individualism — the elevation of our own subjective values, emotions, perspectives, and needs to a position of supreme meaning — is a defining ethic of modern life. Dress codes have fallen away; radio stations compete with Spotify playlists; prime-time television lineups compete with on-demand streaming; the names we give our children now span an unprecedented diversity.

Those same values have made their mark on how we eat. Cuisine historically delineated cultures, but more and more, the choices of what and how we eat have evolved to become personal, not collective. Take the standard American grocery store, which now carries six times as many items as it did in the 1990s. Old-school fast-food combo menus have given way to the hundred different permutations of a Sweetgreen salad. Our conventional three-meal schedule has begun to erode in favor of on-demand snacking, which may help to explain why more and more people find themselves eating alone. And I can personally attest to the growing complexity of cooking for a crowd, which these days might well encompass eaters who are gluten-free, paleo, vegan, and lectin-free (whatever that is). Getting a meal on the table can feel like launching the D-Day invasion. One wonders, once each of us knows our saturated fat tolerance down the decimal point, whether we’ll be willing to share the same meal at all.

When I asked Christopher Gardner, at Stanford, whether he worries about these things — the potential for precision nutrition to usher in a dystopia where we all eat off on our own, slurping down customized, life-extending gruel — his reaction was mostly amusement at the idea of nutritional science moving so quickly and decisively as to be threatening.

Precision nutrition is often compared with precision medicine, a rapidly growing branch of health care that matches disease treatment to a patient’s biology. But Gardner points out that up against medical funding, the money put into nutritional research is minuscule — and federal contributions are actually shrinking. Nutrigenomics won’t pick up serious steam without a lot more investment. “It costs something like $800 million to get each new drug approved, but a big study in nutrition today is typically funded at the level of only a few million dollars,” Gardner says. “These are questions that are going to take a lot more than a few million dollars to answer.”

In the meantime?

“Here’s what I’m hoping happens,” Gardner tells me, with the air of a man who is forever handing out unheeded dietary advice. “Start listening to what everyone agrees on. Everyone says we should eat more vegetables. Everyone says we should eat less sugar and more whole grains. That’s the foundation. For a lot of people, that alone is going to be a big deal. After you’ve done that? Then start the biohacking.”