• Skip to primary navigation
  • Skip to main content
  • Skip to primary sidebar
  • Skip to footer
  • Skip to navigation
Close Ad

The Spoon

Daily news and analysis about the food tech revolution

  • Home
  • Podcasts
  • Events
  • Newsletter
  • Connect
    • Custom Events
    • Slack
    • RSS
    • Send us a Tip
  • Advertise
  • Consulting
  • About
The Spoon
  • Home
  • Podcasts
  • Newsletter
  • Events
  • Advertise
  • About

AI

May 22, 2024

We Used CloudChef’s Cooking Guidance System to Cook Like a Chef

Earlier this month, we visited Google in Chicago, where we got a chance to put the Cloudchef cooking guidance system through its paces.

For those not familiar with Cloudchef, the company uses computer vision to monitor a chef working through a recipe. Sensors and cameras monitor everything from the temperature of a protein to the moisture lost while reducing a sauce to the brownness of an onion and put it all into a machine-readable playback file that can be executed in a kitchen powered with Cloudchef’s software.

We put this playback system, which acts essentially as a cooking guidance system, into action. In the video below, you can watch The Spoon’s Tiffany McClurg being introduced to the system by Cloudchef CEO Nikhil Abraham, and then watch as she cooks a meal of fried rice using a recipe that was created and “captured” in a Google Mountain View kitchen by one of their in-house chefs.

According to Tiffany, she’d never cooked fried rice. Using Cloudchef’s system, it took her about ten minutes to make a meal that tasted pretty darn good!

You can watch The Spoon’s new newly trained chef cook the entire meal using Cloudchef in the video below.

The Spoon Cooks a Meal With CloudChef

May 15, 2024

The Story of Samsung Food with Nick Holzherr

Nick Holzherr, founder of Whisk and head of Samsung Food, is this week’s guest on the Spoon Podcast.

Those in the smart kitchen industry know Nick, in part because his company helped pioneer the early tech behind shoppable recipes, but also because his acquisition by Samsung is the culmination of one of the true success stories in this market.

Today, the technology that Nick and Whisk built is what powers Samsung Food, the AI-powered food and recipe platform that the consumer electronics giant debuted at CES 2024.

Some of the things we talk about this latest episode of The Spoon Podcast include:

  • The story of how Whisk was the first recipe startup to explore how to use AI and apply it to recipes.
  • Nick’s experience going on the British version of The Apprentice and appearing before Lord Sugar (the British version of Donald Trump) to pitch the company. 
  • The growth of the company as Whisk started working with grocers in Europe and eventually appliance brands
  • Nick fielding calls from three companies who presented offer sheets to buy the company.
  • We talk about what Nick is excited about and how he sees technologies like AI being applied in the kitchen in the future.

You can read the full transcript of the conversation below and listen to the full interview by clicking play on the podcast player below or heading to Apple Podcasts, Spotify, or wherever you get your podcasts.

If you’d like to connect with Nick in person, he will be at the Smart Kitchen Summit on June 4th. You can get your tickets here.

Conversation Transcript:

Michael Wolf: Nick just told me whatever I throw at him, he’s ready for, especially with that new microphone. You look great in that, in your little room there, the microphone.

Nick Holzherr: It’s brand new. It was three months old. I realized I needed to start upping my game when everyone’s podcasting now, right?

Michael Wolf: Everyone’s podcasting. So when’s the Nick Holzherr podcast start, by the way?

Nick Holzherr: I haven’t got my own podcast, but if anybody wants to invite me onto theirs, I’ve now got a good mic.

Michael Wolf: Yeah, well, Nick, I feel I’ve known you probably since 2016, 2017. I think Whisk was the first company I remember that was really talking up AI and recipes. That was one of your original stories, right?

Nick Holzherr: Yeah, I mean, we started with the problem that users want to use recipes anywhere, not just from one manufacturer or one publisher or one grocery company or one tech startup with a couple hundred recipes. And the only way to really make that happen is with AI. If you want a smart experience, that is. Use AI to parse all these recipes and make them smart. So we did that, but that was before deep learning, before it was accessible to startups. We were literally flying out on planes to America, buying the latest processors, latest chips, graphics cards, and using them with basically PhDs, ex-professors that we had hired to build these models. It was really early in AI, before TensorFlow, before these libraries. Of course, now it’s totally changed. Super early.

Michael Wolf: I was looking back in the history a little bit and I didn’t realize Whisk had started back in 2012. Back then, you were actually on a British TV show, British Apprentice on the BBC. Was it pitching Whisk at the time?

Nick Holzherr: Yep. It actually was. It’s called the Apprentice, and it’s the same thing as basically Donald Trump has in the US, right? And we have a different person running it called Lord Sugar. The guy in the UK is called Lord Sugar and he was an advisor or member of the Labour Party, which is the left side of the government. So it’s kind of a little bit the opposite of Donald Trump.

Michael Wolf: Lord Sugar. Way better name.

Nick Holzherr: Yeah, he was originally Sir, I think, and then they made him a Lord. This is the British monarchy system, which is also hilarious. It was amazing fun. You do 12 weeks of tasks competing against other people. And then if you get to the final, you get to pitch your business plan. If you win, he gives you money and becomes a partner in your business. I got to the final and I didn’t win. But because it was such a popular TV show, and I got to pitch it to him and his advisors and therefore the UK, everyone wanted to invest in it, but not everyone, but enough people wanted to invest in it. Most people would take the meetings with me. So I got inbound from a lot of the grocery stores and publishers and stuff saying, hey, we’d love to work with you.

Michael Wolf: The Tescos the world came in with you say, hey, let’s meet. Was it immediate like the next day?

Nick Holzherr: Yeah, literally that night, my inbox, my LinkedIn and my inbox just filled up with loads of people who wanted to work, who said, ‘I love it. I know he (Lord Sugar) didn’t invest, but can I invest or can I work with you? And then the ones who hadn’t, when I messaged them months later and said, Hey, this is the idea, I pitched it to on the Apprentice, they would take the meeting, not necessarily because they’re interested, but just wanted to meet the guy who’s on TV.

Michael Wolf: There was a curiosity in meeting you, right?

Nick Holzherr: And for a salesperson, that’s what you need, right, a foot in the door.

Michael Wolf: I feel The Apprentice was the predecessor to Shark Tank, which is also Dragon’s Den in the UK. The value for you as an entrepreneur is it’s just a giant commercial. The value is in the millions of dollars in terms of free publicity.

Nick Holzherr: Exactly. It’s huge. And it’s also actually fun. It is really, really fun. It is also challenging because it’s not a show about business success. It’s about conflict, essentially, right? Of people conflicting and trying to make something work. It’s like a business assault course. They put you through psychometric tests and psychology interviews to basically choose people that they think will conflict with each other and create good TV. So that part is stressful, but you also get to do some really cool things on the show, like build products super quickly and pitch them to big people. So it’s really fun. But ultimately it was the platform I needed to get Whisk started.

Michael Wolf: And so you got Whisk started and I think I connected with you in around 2016, 2017. From those early days when you had the business plan, what you pitched on Apprentice BBC versus when I met you around 2016, what had changed? Obviously you’d started the company, went from business plan to actually starting this thing, but had the vision changed at all?

Nick Holzherr: The core mission was to help people cook, make it more joyful as we talk about it today, and to help them in that process. That hadn’t changed. But what had changed was we started off with a consumer app and we couldn’t get it to what we thought was enough scale. So we pivoted from trying to build a consumer app to looking at all that tech we’d built. And you started off talking about AI. We’d built a lot of AI way back, 10, 12 years ago. And so we were really early in that. And businesses wanted to use that. We knew that because we had, when we built our consumer, we had built widgets that publishers could integrate on their sites. And through that, we had loads of conversations with big publishers, big retailers and big CPG brands. And a lot of them said, ‘we love this widget. Yes, we’ll integrate it, but can we also use some of the other tech?’

So, in2014, 2015, and into 2016, we couldn’t make the consumer app work. So we pivoted into a B2B platform, offered all the tech that we had built for anybody to use, built an API platform for people to use. We were powering half a billion monthly consumer interactions at one point, at the height of it, which was pretty impressive and we were powering it for a lot of the big players in the world like a lot of big CBG brands. If you made a list of the top players, maybe half of those were using our technology.

Michael Wolf: Were you on big publisher websites?

Nick Holzherr: Yeah, the biggest, right? Food Network, Allrecipes, BBC Good Food, About.com, the big, big sites with many tens of millions of hits a month, users, those were integrating us and we were sending them across to grocery retailers like Walmart, Amazon, Instacart, the UK ones like Tesco. And the big CPG brands wanted to be part of it because they wanted to have their products featured. So when it’s butter, they want it to be the Unilever butter, or when it’s stock, the Unilever stock, right? So they cared about that. And they also wanted to connect their own websites to commerce. Because we had those integrations, they then said, can you power more of our platform? We said ‘yes’. Then all the big IoT companies came along, the appliance manufacturers came along and said, ‘hey, can we use your platform as well?’ So we were powering a whole bunch of different IoT players. And that’s, I think, where you and I first started talking.

Michael Wolf: So just real quickly, you were with the publishers and you were essentially doing shoppable recipes. So, like Bisquick, Nestle wants their yogurt or whatever. You were one of the pioneers in the whole shoppable recipe concept. And that was around the 2013, 2014 time frame.

Nick Holzherr: That’s right. I mean, there was Constant Commerce (now Constant.co), which are no longer there, but Constant Commerce was there at the same kind of time as us, super, super early. But it was basically only Constant Commerce and us. And we took a different approach to them. They were more enterprise-y in their proposition. So some of the big enterprise players wanted them. And, but we were more like open and like, let’s integrate the whole market kind of sort of game. So someone wanted us and we had a bit more of a B2C to B2B2C kind of play where we integrated onto their sites and then they added to a shopping list. It would still be our shopping list and it users could save the recipe into our recipe box. Our consumer experience is always there, but it was integrated on other people’s sites and shops and brands.

Michael Wolf: And the appliance guys showed up around in the 2015, 2016 timeframe. And what were they asking you? ‘Hey, we want apps. We want people to have recipes.’ What were they showing up with, and what questions did they have?

Nick Holzherr: It was ‘we want to connect our appliances, so can you give us recipes in a structured way?’ Or can you make it work with our products?

Michael Wolf: It was the guided cooking recipe era.

Nick Holzherr: Exactly. But also shoppable recipes. I think everybody always thinks and starts in the early stages of developing an experience that thinks, ‘I want to make money. How do I make money? I make it by connecting to commerce. So, let me connect to commerce.’

So that was definitely part of the game as well. And then what happened was really at that point what we were offering was in its infancy. It was not a mainstream thing to do. And it was really around the 2016 and 2017 mark when suddenly the e-commerce market had grown. Penetration of e-commerce had doubled or tripled from what it was back in 2012. And so you suddenly have everyone wanting it. And it was kind of a crazy time where we had struggled as a business until 2016, 2017, and suddenly they were inundated with everyone saying, ‘hey, can we use your platform?’ And we suddenly became profitable. We went from five people to 30 people in a year, not from any investment, just revenue. We had no investment at that point. Everything was bootstrapped. We were making money, and the business was profitable. It was like, wow, this is really fun. What is it like to make money? A startup making money? What is this?

Michael Wolf: It’s nice to make money. And so how were you making money? You were doing these recipes, and were you kind of taking every recipe impression that went through, you got a little bit of money? Was there a big appliance company that wanted to pay you to do a custom integration?

Nick Holzherr: All of those. So, we had three revenue models. We had the license fee model, a monthly fee for the API. We had a grocery per click or per new customer fee, which was a small fee per transaction. But when you add it all up, it works out relatively well. And then we also had an ad model where you, as a CPG company, if you’re Nestle and you want to get your yogurt sold, you pay to have your yogurt featured. It doesn’t say Nestle yogurt in the recipe, but yes, Nestle yogurt the customer gets. And so they don’t pay on a per acquisition there, they pay on a per view because you are, you’re inspiring the customer. You’re not just selling the customer.

Michael Wolf: Okay, and so this thing’s going crazy. You had 30 employees and you’re making money. Life is good. And then, at some point, or probably I imagine around this time, you’re the kind of the apex of this ride, you get a call from Samsung?

Nick Holzherr: Yeah, I mean, there’s three companies, another one of the big appliance companies, and one of the world’s slash one of the world’s biggest technology companies all rang up at roughly the same time. Like in the six month period of time, we had three serious acquirers wanting to buy us and a bunch of other ones on the side saying, ‘maybe could we, would you be open to it?’ But three of them were kind of term sheet level interest, like real interest. Flying people into our offices saying, ‘hey, can we buy you?’ And it was a really good time; we were profitable, and we didn’t need to sell. And then, ultimately, we chose to go with Samsung.

Michael Wolf: And you’re looking and you’re just thinking in your head. “Lord sugar. You should have made a better offer’. So you had these three offers; a technology platform company, two big appliance companies, it sounds like. Were you inviting them in, or was their head of acquisitions calling you up?

Nick Holzherr: The latter. They were ringing in us. And this was kind of what’s crazy about it because initially I was even saying no to them, not to the acquisition teams, but to the initial integrations. I was like, why do you want to integrate to an appliance? Are we going to get any users from this? I was in the user’s head and thinking I’m not going to make any transactions. How many fridges have you got live? How many users are going to click on it? Am I going to make enough money on this to make it worthwhile while visiting your office? Because often they’re these large enterprises, their request is, can you visit our office to pitch this certain person? And I was apprehensive. I was like, is it wasting my time?

Michael Wolf: You’re busy. You have 30 employees.

Nick Holzherr: We’re busy. And the funny thing is, of course, those conversations ended up with the M&A team suddenly getting interested. Because they’re looking at it going, ‘hang on, how many people in the market are using it? Our competitors are using it. They’re thinking, ;who’s going to end up owning this?’ They probably knew that M&A stuff was happening because the people in the industry know each other quite well. So they probably all started worrying about what happens if the other person wins it? And that’s a great place to be as a startup. It was literally the perfect time. I could probably have gotten a better price for it if I’d held on another few years. It’s impossible to price it. It’s impossible to get the right time, and it was certainly a good time and I’m happy with what worked out.

Michael Wolf: I think it was a good time, and Samsung is an interesting company. Did you have to fly out to Seoul to pitch?

Nick Holzherr: No, I had phone out to Seoul multiple times after we had a partnership with them, because we started powering loads of stuff for them, right? It became a really good partnership where we were powering way more than initially was requested. We were building, even building some new features based on their budgets and based on their requests and their needs. It was being used across lots of appliances. And I had a vision to being used by lots of other divisions as well, which is what ultimately ended up happening, of course. But that we couldn’t predict that at the time. But then I did fly a bunch to San Francisco because the acquisition team was called Samsung Next. It was an innovation arm of Samsung. And basically HQ, because it’s quite hard for the Korean team to buy stuff because of how something is structured, they asked the innovation arm, hey, can you guys take a look at this? And they did, and they did the acquisitions. So actually, our bosses, if you like, or our acquirers was the San Francisco entity of Samsung.

Michael Wolf: And so that deal comes through, you accept it. I think I remember being in Chicago at the Housewares show. I think when I interviewed you about that deal, I think is right. So there must have been around the March timeframe, it was at Housewares, it was 2018, 2019, one of those years. And it’s been an interesting ride.

Nick Holzherr: Yep.

Michael Wolf: A lot’s probably happened. So maybe talk about you get acquired by Samsung. What’s that like initially kind of absorbing you into this giant Borg, which is it’s a massive company, even though it’s like the next company like the San Francisco companies, but it’s still Samsung.

Nick Holzherr: Yeah, it is. Actually, what I would say is, what I’ve been really pleasantly surprised by is how much autonomy they’ve given us and how much they’ve let us keep parts of the culture that matter. We were always a distributed team and they’ve let us keep that. And that was a massive advantage during COVID when it was so difficult if you were not a distributed team, but it also allowed us to kind of operate, I think, quite effectively. And keep some of the culture we have, like we go on annual offsites where we fly everyone to one place in the world. Like that’s not a normal thing in an enterprise to explain to your HQ that, hey, we’re going to fly the whole team to Greece or something. And they support that. So that’s been awesome.

Michael Wolf: You go to Greece for your offsights? What the heck?

Nick Holzherr: Well, we have to fly somewhere and actually Greece ends up being quite, we actually haven’t actually, Greece is our next one. We actually haven’t been to Greece yet, but like the previous one, Cyprus, we did actually do one in Korea. That was the most expensive one we did. We did Lisbon, Madrid, Budapest. And these places are actually, if you look at them on spreadsheet, the cheapest places to fly everyone, but it is also nice. It’s also fun to actually spend a week together and meet people in person, especially if you’re working in a distributed environment. So we spend a lot of time building a distributed culture.

Michael Wolf: Nice!

Nick Holzherr: and making sure that we, people felt valued and motivated and knew what they’re doing. And Samsung let us keep that. So I’m really grateful and respect Samsung for that. What was crazy was some of the enterprise ways of managing a business, right? That definitely added a layer of complexity and that’s inevitable with any enterprise. And then I think the other thing that was crazy was scaling. So they wanted us to scale our team from 30 people to 120, and they asked me to do it.

in ASAP, I did it in nine months. So going from 30 to 120 in nine months is an experience that I’m not, maybe I’ll do it again, but it was intense. The amount of interviews you have to do, the amount of like how to integrate that at the same time as integrating your company to a large enterprise like Samsung. So all of that in one go was probably the most intensive six months of work that I’ve ever done. Massive learning curve again, hard work, but I learned a bunch and I’m grateful for the experience.

Michael Wolf: Were you in on almost every one of those interviews and decisions? You were pretty hands-on, it sounds like. That is crazy. That’s crazy. Okay, so you grow to 120, that’s impressive. And I think the company over time, I saw more and more getting integrated. And I think the culmination of that was this past CES, when I think it essentially transformed from Whisk to Samsung food. So Samsung food was a big announcement for Samsung at CES 2024 in January.

Nick Holzherr: Yeah.

Michael Wolf: But the heart of that, the beating heart of that was really the Whisk acquisition, right? It’s almost a rebranding of Whisk.

Nick Holzherr: Yeah, that’s right. And there are two sides to that, whether it was a good idea or a bad idea. So I built the Whisk brand, right? I own a bunch of Whisk. I’m drinking out of a Whisk flask right now. So like, hey, kill Whisk as a brand. And some users felt that too. But ultimately, there is a thing with any enterprise probably. If you’re trying to build an app that comes across as not from here, you’re going to struggle getting integration into the different hardware units and you’re going to struggle integrating into the Samsung ecosystem. And ultimately, if you’re building a consumer experience, you have to leverage the distribution of a platform like Samsung. Samsung has one of the most devices in the world of any company. I think it probably has the most devices of any company if you add all the different units, because it’s so broad in what it does in TVs and mobiles and kitchen appliances, watches and now rings, everything. So we had to use that distribution. And the most effective way of doing that was to call ourselves Samsung Foods. And actually that has worked, in terms of the integrations we have planned and the ones we’ve achieved, but ones we’ve now have on the table that are going to happen over the next six, 12 months. I think it’s part of what I’m actually most excited about in where we’re going. It’s especially with the health, because food and health are so closely coupled with obesity and diabetes and other things. Yet, no one has really solved that. Actually people have worked on it, but no one’s really solved it. And I think Samsung has got an important part to play because they have health and food appliances. And I think if we do that, that’s actually a really important thing in the world that we’re doing. Something I feel proud of and passionate about. And that I find awesome. So that is, Samsung food as a brand, I know people are split on the opinion. I think it’s a good name and there’s lots of advantages of the name. If you’re looking at 10 food apps or 20 food apps in your…

Michael Wolf: Yeah, you’re going to see Samsung food. Yeah.

Nick Holzherr: you’ll notice it and you’ll probably try it. Your propensity to try out the app is higher. But of course, Samsung hasn’t got a huge plethora of successful apps. So that’s part of also what we’re helping Samsung with.

Michael Wolf: I always have an 8:30 call, so I’m going to tell them to be a couple minutes late.

Michael Wolf: and late. I’m like, I wanted to go for an hour and then all of a sudden.

Nick Holzherr: We can do part two later, a different day if you want.

Michael Wolf: Yeah, I’ll edit that part out. So, I think that when I think of Samsung food or Samsung with your appliances, you have the wearables, you have what you built in Whisk. There’s a lot of different parts of the puzzle there. And I think you’re almost on a collision course with some of the other folks who are just coming at it from the precision nutrition side. Like I look at January AI, which is a pretty cool app. They do like take a photo of food and you can predict blood sugar. It seems like makes sense that ultimately that type of function would be built into what Samsung has because you guys have so many other parts of the puzzle. So I see you guys have a great foundation. A lot of it’s built on the appliance and the wearables and then that and the app side.

Nick Holzherr: Yeah, I think that’s what’s exciting about the sensors, right? If you’ve got access to sensors and distribution through the hardware and the brand, that gives you some good starting Lego bricks. It’s not the whole solution. It’s such a big problem in the world. A big proportion of the world is going to suffer from the health challenges of not eating healthily. Of course there’s going to be thousands of companies. I hope a lot of them are successful alongside us.

Michael Wolf: I mean, I think a lot of people are just thinking, hey, now we can just take a pill and I will no longer be at risk for type two diabetes. I’ll be always 50 pounds. But I think that’s not realistic. You can’t have half the world’s population at risk for type two taking medication or a shot. I think what you guys are building could be really interesting. It’ll be a few years though, a lot of these different things need to come together. That sounds like what that’s driving you. What else is driving you as you look forward towards five years down the line, 10 years down the line in this space? What is really exciting for you?

Nick Holzherr: I think the advances in AI are exciting for everyone in the space, including us. We tried to do stuff like vision AI, use your camera to detect items, food, two, three years ago and failed. It wasn’t good enough. That was not despite having good engineers on it, it was hard to do. Yet now you come along and use an open source library by open AI and boom, it works. Sure, it needs tuning, it needs some work, it doesn’t maybe it takes a little bit of work to get it to production level, but it’s so easy. Stuff that we spent 10 years building, you can now use the OpenAI API for it and get a good way there. Which is so scary for people who’ve invested a whole bunch of time and effort into building really, really advanced technology that so much of it’s now easier than ever to do, which means it’s open to anybody. The scary part is, how do you win if everyone can do it? The fun thing from a consumer side is it’s going to be possible to build some of the stuff that we all hoped would be there five years ago. Now it will finally get to a point where it’ll be better and usable and actually add starting value. The smart kitchen, there’s always the question is how far off a truly smart kitchen that actually adds value to the user are we. When will it stop being a gimmick and when will it start being smart and useful? We’re getting closer and closer to that being true. Of course it’s true in some ways already, some devices are fantastic in adding value. But on a macro level where I can talk to my mom and say, do you want a smart kitchen? And she goes, yes, I do because it adds loads of value.

Michael Wolf: It’s not just turning things on and off with an app and making things more difficult by adding more processes in the way and kind of app friction, it’s actually making it more useful. So going from those early days where you’re buying GPUs and trying to figure out to build AI and going on TV and then ultimately to where you are today, it’s been quite a ride. It’s been fun talking to you, Nick, about hearing and hearing about all this. Thanks for spending some time with us.

Nick Holzherr: Thank you, I look forward to seeing you in Seattle.

Michael Wolf: You were seeing Seattle at Smart Kitchen Summit. Everyone who wants to see Nick in person, famous TV star, Apprentice star. I’m just embarrassing him. You can see him in Seattle. All right, Nick. Thank you, man. Don’t hang up yet. Don’t hang up yet.

Nick Holzherr: Thank you, thanks Mike, bye bye.

April 22, 2024

Micromart Wants to Create Just-Walk-Out Convenience Anywhere With Its Just-Plug-In Cabinets

Earlier this month, we learned that Amazon is phasing out its Just Walk Out technology at its Amazon Fresh grocery stores. The company didn’t say much about the reasoning behind it, but one likely reason is customers never valued skipping the checkout line in a traditional grocery store shopping experience as much as Amazon anticipated.

But that doesn’t mean shoppers don’t value speed to completion and low-friction shopping experiences. Getting in and out quickly is highly desirable when watching a ballgame or picking up something quickly for lunch during the workday. That’s why Amazon will continue to keep its Just Walk Out technology in sports stadiums and in its Amazon Go fast-format convenience stores, which are typically located in busy downtown office corridors.

Still, do we need whole stores outfitted with cameras and sensors? What if we could condense all this down to a couple of cabinets that can sit in any condo or office lobby?

That’s the idea behind Micromart, an eponymously named micro-market platform from the same Toronto-based team behind Kitchenmate. Micromart’s solution uses AI-powered image recognition technology, putting it into standalone refrigerated cabinets that fit anywhere with a little floor space and a power outlet to plug the cabinets in.

To open the locked refrigerated or freezer cabinet, the customer taps with their phone. They open the cabinet, grab the item(s) off the shelf, and once they close the cabinet, a receipt is generated. If the item is a meal that needs to be heated, the customer can then heat the meal in a “smart cooker” that is attached to the cabinet.

The addition of a food heating system is one of the major differentiators for the Micromart solution, something that company CEO Yang Yu says they developed for Kitchenmate. Kitchenmate, which The Spoon covered way back in 2019, started as a combination food-to-go service for condos and offices. According to Yu, it was while looking for available technology to enable easy unattended purchases of their Kitchenmate meals that the company realized they would need to make their own smart fridges and commerce system.

“We started with the heater,” said Yu. “That was the only thing we had, but then we realized we needed to put the food somewhere, so we built a fridge. When we built the fridge, we were looking at AI companies that did just-walk-out technology, but all of them had issues, and they were all very expensive. And none of them were very accurate. So we had to build our own.”

After building just-walk-out technology for their fridge and deploying it in different locations, they realized the refrigerated cabinets and the heating system were the business. Not long after, Micromart was born.

One reason that Yu and his team saw this as a potential big business is the realization that many office buildings are shutting down cafeterias, often replacing them with just a couple of vending machines. While some solutions, like Farmer’s Fridge, provide fresh options, there aren’t many choices for fresh and hot food.

“Nobody wants to eat vending food,” said Yu. “There’s definitely success stories around healthy vending, but you’re not going to get the variety and the hot food that people expect out of a cafeteria.”

In addition to the refrigerated cabinets and the food heating system, the Micromart solution comes with software as a service that lets retailers track and forecast inventory, electric price tags, and built-in digital ad displays that the operator can customize. The company’s offering also includes a Shop consumer app that can be customized with the operator’s branding. Pricing for a three-cabinet system is $19 thousand for the cabinets, plus transaction and monthly SaaS fees.

Micro-markets aren’t new. Researchers estimate that the micromarket business in the US was almost $4 billion in 2022 and expect it to grow by 13% through 2030. However, many of the solutions are not much more than refrigerators with RFID scanning or weight sensors built in. Other solutions, like those deployed at airports, require the customer to pick up the items and go through a self-checkout scan, often with a store employee eyeing them from close by. Micromart wanted to marry the lighter footprint of older cabinet systems with the more advanced Amazon Go-like vision systems.

“The whole premise behind this was that you could literally put it anywhere in North America,” said Yu. All you need is a standard electrical outlet, and you plug it in, and it works.”

According to Yu, the Micromart solution will debut at the NAMA show in May.

April 19, 2024

The Food Tech News Show: A Look at Our Food Lives in The Year 2055

Welcome to the Food Tech News Show! You can watch the show live here at 1 Pacific/4 Eastern or on Streamyard, YouTube, or LinkedIn.

Fast Food Facial Recognition? - FTNS

This week Mike and Carlos be joined by Future Market’s Mike Lee to talk over some of the most interesting stories of the week. Mike Lee will also talk about his new book Mise,  which paints four different scenarios depicting the potential direction of our future food system.

Here are all the stories we’ll be talking about.

Bored & Hungry Closes – Food & Web3: A check in on where things are

In March 2022, NFT and crypto investor Andy Nguyen purchased Bored Ape #6184 along with three Mutant Apes and soon decided to establish a Bored Ape-themed restaurant named Bored & Hungry. The restaurant opened its doors on April 9, and by the end of its first day, it had served 1,500 burgers and had lines stretching around the block.

Two years later, Bored & Hungry has closed.

Last week, Nguyen announced on Instagram that the restaurant’s original location in Long Beach, California, was closing. He shared that they had sold the concept to a franchising company from Asia known as HUNGRY Dao.

Is AI-Powered Customer Interaction at Fast Food & Retail Giving Up too Much Privacy?

A Fast Company article titled “How fast food is becoming a new surveillance ground” looks at how new customer interaction layers using things like bio-authentication, cameras, profile information, and more are a new risk for gathering information about the public. 

And earlier this week, we saw Steak n Shake launch facial recognition nationwide for check-in.

Are we going through an airport or going to buy a surf and turf?

Vow Thinks Imitating Meat We Eat is Bad Approach. Enter the Quail Parfait. 

Green Queen: The Syndey-based startup is today launching its cultivated Japanese quail in Singapore’s Mandala Club, after the country’s regulator gave it the go-ahead to sell the product. But unlike other rollouts of cultivated meat, where chefs are supplied with the meat itself (which they then incorporate into dishes), Vow is taking a novel approach. What restaurant kitchens get is a parfait containing its cultivated quail.

Mike’s Book about The Future of Food

The four future visions in Mise (pronounced “meez”), which range from the year 2033 to 2067, are created to help people understand the potential long-term impact of things that are happening today in the world on our food system. The book identifies 5 major happenings in society, technology, the economy, the environment, and politics (abbreviated and referred to as the STEEP factors) that will have a profound impact on the way the world produces and consumes food.

April 1, 2024

When It Comes to Making Generative AI Food Smart, Small Language Models Are Doing the Heavy Lifting

Since ChatGPT debuted in the fall of 2022, much of the interest in generative AI has centered around large language models. Large language models, or LLMs, are the giant compute-intensive computer models that are powering the chatbots and image generators that seemingly everyone is using and talking about nowadays.

While there’s no doubt that LLMs produce impressive and human-like responses to most prompts, the reality is most general-purpose LLMs suffer when it comes to deep domain knowledge around things like, say, health, nutrition, or culinary. Not that this has stopped folks from using them, with occasionally bad or even laughable results and all when we ask for a personalized nutrition plan or to make a recipe.

LLMs’ shortcomings in creating credible and trusted results around those specific domains have led to growing interest in what the AI community is calling small language models (SLMs). What are SLMs? Essentially, they are smaller and simpler language models that require less computational power and fewer lines of code, and often, they are specialized in their focus.

From The New Stack:

Small language models are essentially more streamlined versions of LLMs, in regards to the size of their neural networks, and simpler architectures. Compared to LLMs, SLMs have fewer parameters and don’t need as much data and time to be trained — think minutes or a few hours of training time, versus many hours to even days to train a LLM. Because of their smaller size, SLMs are therefore generally more efficient and more straightforward to implement on-site, or on smaller devices.

The shorter development/training time, domain-specific focus, and the ability to put on-device are all benefits that could ultimately be important in all sorts of food, nutrition, and agriculture-specific applications.

Imagine, for example, a startup that wants to create an AI-powered personalized nutrition coach. Some key features of such an application would be an understanding of the nutritional building blocks of food, personal dietary preferences and restrictions, and instant on-demand access to the application at all times of the day. A cloud-based LLM would likely fall short here, partly because it would not only not have all the up-to-date information around various food and nutrition building blocks but also tends to be more susceptible to hallucination (as anyone knows who’s prompted an AI chatbot for recipe suggestions).

There are a number of startups in this space creating focused SLMs around food and nutrition, such as Spoon Guru, that are trained around specific nutrition and food data. Others, like Innit, are building their food and nutrition-specific data sets and associated AI engine to be what they are terming their Innit LLM validator models, which essentially puts food and nutrition intelligence guardrails around the LLM to make sure the LLM output is good information and doesn’t suggest, as Innit CEO Kevin Brown has suggested is possible, a recommendation for “Thai noodles with peanut sauce when asking for food options for someone with a nut allergy.”

The combination of LLMs for generation conversational competency with SLMs for domain-specific knowledge around a subject like food is the best of both worlds; it provides the seemingly realistic interaction capability of an LLM trained on vast swaths of data with savant-y nerdish specificity of a language model focused on the specific domain you care about.

Academic computer scientist researchers have created a model for fusing the LLM and SLMs to deliver this peanut butter and chocolate combination that they call BLADE, which “enhances Black-box LArge language models with small Domain-spEcific models. BLADE consists of a black-box LLM and a small domain-specific LM.” 

As we envision a food future of highly specific specialized AIs helping us navigate personal and professional worlds, my guess is that the combination of LLM and SLM will become more common in building helpful services. Having SLM access on-device, such as through a smartwatch or phone, will be critical for speed of action and accessibility of vital information. Most on-device SLM agents will benefit from persistent access to LLMs, but hopefully, they will be designed to interact independently – even with temporarily limited functionality – when their human users disconnect by choice or through limited access to connectivity.

March 12, 2024

Announcing The Food AI Co-Lab, a New Collaboration Between The Spoon & Future Food Institute

If there was one thing we learned when we held the first-ever Food AI Summit last October, it is that pretty much every food company believes their business will fundamentally change due to artificial intelligence.

Whether it’s companies building farm equipment, managing food supply chains, launching new grocery shopping formats, or creating new quick-service restaurant chains, no one along the food value chain will remain untouched by the rapid pace of change brought on by AI. In other words, we are in a once-in-a-generational rethink of business as usual, a tectonic shift that demands company leaders continuously learn, strategize, and collaborate to make sure their companies survive and even thrive into the future.

Because of this, we realized that we wanted to find a way to bring together our community and others within the food system to talk about the different impacts AI is having across various parts of the food system more than once a year. While we loved the fact that the big ideas that were shared at the Food AI Summit have already resulted in new partnerships and collaborations, we wondered if we brought together folks more regularly – on a monthly basis or even more frequently – might have an even bigger impact.

Luckily for us, one of my favorite organizations – the Future Food Institute, led by one of the most consequential leaders in the future food space in Sara Roversi – had a similar idea. So when Sara approached me about joining forces for a collaborative new organization to do just that – I jumped at the chance.

So, alongside the FFI, I am super excited to announce today the launch of the Food AI Co-Lab!

What is the Food AI Co-Lab? It’s a collaboration that aims to be a meeting space and learning center for leaders who are building the future of food through artificial intelligence. We will explore different topics, engage with our community, and provide information such as industry surveys about what people are doing at the intersection of food and AI.

To kick things off, we will host monthly industry-focused meetings with thought leaders creating using AI across various parts of the food system. Soon, we will also announce in-person events in the US and Italy where the community can get together, network, learn together, and build their own collaborations.

If you’d like to join us on this journey, we encourage you to join our LinkedIn group and also register for our first virtual event, AI & The Future of Food, which will take place next Tuesday, March 19th. At that event, we’ll interview two thought leaders: Dr. Patrick Story, a professor of Philosophy at Cal Poly, San Luis Obispo, participating in a National Science Foundation-funded project analyzing the impact of automation and AI on the food system, and Kevin Brown, the CEO of Innit, a company building a platform that plugs into generative AI large language models to make them more food “fluent” and power AI-assisted food knowledge systems and services.

I hope to see you there, and I am excited to work with you to learn, collaborate, and build the future with the Food AI Co-Lab!

February 29, 2024

Fresco Locks Up Deal To Bring Kitchen OS to Middleby Consumer Kitchen Brands

This week, smart kitchen startup Fresco announced it had struck a deal with Middleby to integrate the Fresco Kitchen OS platform across the cooking equipment giant’s residential portfolio, starting with high-end kitchen appliance brand Viking.

The first Middleby product line to incorporate the Fresco Kitchen OS platform will be the Viking RVL Collection. The RVL line, a new collection introduced this week that is more modern and tech-forward than traditional Viking lines, will incorporate the Fresco OS via a firmware agent residing on an integrated system-on-chip. In addition to the built-in Fresco firmware, which will power the connectivity to other Viking and non-Viking brands that use Fresco technology, the smart kitchen startup is also providing the appliance brand with a white-labeled Viking app. According to Fresco, they will also work on a smart recipe app for another Middleby brand in AGA(Middleby acquired AGA in 2015 and acquired Viking in 2012).

The move is a nice win for Fresco, in part because it is a validation of its revamped kitchen OS, which the company announced a year ago. The deal also adds a premium built-in appliance brand in Viking to its growing list of customers. The Irish/US startup, which got its start a decade ago with the launch of its connected kitchen scale, has been steadily chalking up wins over the past few years, including with Japanese microwave maker Panasonic and the once white-hot Instant Brands.

Middleby is interesting because it represents many other potential product lineups for integration. Not only are there other built-in appliance brands like AGA, La Cornue, and Rangemaster for Fresco to expand to, but the company also has a couple of outdoor grills as well as a countertop cooking brand in Brava. For its part, Brava was not mentioned as an initial target for the Fresco technology, but my guess is that has as much to do with the complex and fairly unique commands associated with the Brava light-cooking technology as anything.

I asked Fresco CEO Ben Harris how his company has continued to grow its partner list in the connected kitchen space, and he pointed towards the system-on-chip in his hand, which featured an Espressif ESP 32 DSP. According to Harris, Fresco’s hardware engineering and silicon understanding, born almost a decade ago when the company introduced the Drop scale, has helped them win customers looking to take advantage of their technology and their knowledge in this space.

Harris said that having on-chip, on-board integration of a kitchen tech stack via their kitchen OS SoC module is not only a good way to on-board a customer like Viking, it also results in a faster system. Harris said that in the early days, when they went from building cloud-to-cloud integration between smart kitchen equipment from different manufacturers to integrating their own SoC with built-in firmware, the quickly saw how much more responsive and fast the system performed.

“It was like that, in an instant,” said Harris in an interview with The Spoon at KBIS. “It definitely feels like it’s an extension of the appliance, and it makes a big difference in the engagement of the user.”

In addition to partnering up for a product integration, the two companies also announced that Middleby will become an investor in Fresco. In this sense, the deal is somewhat reminiscent of the Instant Brands partnership, which had the pressure cooker maker investing in Fresco at the time of the deal.

You can watch my interview with Harris at KBIS below:

Interview With Fresco's Ben Harris on Connected Kitchen Technology at KBIS 2024

February 7, 2024

Check Out NXP’s Presence-Sensing Cooktop Demo Powered by On-Chip AI

We’re still sifting through some of the cool product demos from CES last month, and one that caught my eye was the demo by NXP of a presence-sensing cooktop powered by an embedded MCU. According to the company, the system used a Neural Processing Unit that runs the machine learning and facial-recognition algorithms within the system rather than relying on a cloud-based compute. The demo featured a device control interface from Diehl Controls.

As can be seen in the video below (taken by TIRIAS Research Principal Analyst Francis Sideco), NXP spokesperson Thomas Herbert shows that you can turn on the burner with either touch or motion sensing (including motion sensing with cooking mitts on). From there, the system is using facial recognition to detect if a person is within local proximity of the cooktop. According to Herbert, presence detection comes into play in a scenario where there is a critical state, such as the pan getting too hot and there is water boiling over. If the system detects a critical state and no one is standing around the stove, it will shut off the heat and can send an alert to the person via a Matter (an open-source smart home connectivity standard) enabled device to alert the cook that the system has intervened on their behalf.

CES 2024 - INVITE ONLY- Part 3 of 3 NXP Autonomous Home Showcase Based on Matter

The demo is interesting to me in a couple of contexts. One is that, as we’ve written about here on the Spoon, the number one reason for fires in the home is due to cooking mishaps, and enabling your stovetop or other cooking appliances to recognize both anomalies as well as the presence of someone could be a real gamechanger, akin to the dip in automobile accidents in recent years due to the widespread prevalence of blindspot detection systems in modern cars.

The second context that’s interesting is that it could become a significant technology in assisting aging-in-place scenarios, particularly for seniors who become a little more forgetful as they age. One of the key determinants of whether folks can continue to live independently is their ability to feed themselves, and by providing “blind spot” detection like this, my guess is this could extend many seniors’ ability to live independently for years.

January 29, 2024

Chris Young: Generative AI Will Provide Big Payoffs in Helping Us Cook Better, But Overhyping It Will Burn Some Folks

Chris Young has never been shy about providing his thoughts about the future of cooking.

Whether it was on stage at the Smart Kitchen Summit, on his YouTube channel, or a podcast, he’s got lots of thoughts about how technology should and eventually will help us all cook better.

So when I caught up with him last week for the Spoon Podcast, I asked him how he saw things like generative AI impacting the kitchen and whether it was necessary for big appliance brands to invest in building out their internal AI competencies as part of their product roadmaps for the next decade. You can listen to the entire conversation on The Spoon podcast.

I’ve excerpted some of his responses below (edited slightly for clarity and brevity). If you’d like to listen to the full conversation, you can click play below or find it on Apple Podcasts or wherever you get your podcasts.

On the resistance by some to using advanced technology to help us cook better:

Young: “A lot of people are focused on going backward in the kitchen. They want to go back to cooking over charcoal and cooking over fire. That can be fun, but if you look back at what it was really like in the 19th century, the kitchen was not a fun place to be.”

“The modern kitchen is much healthier and much safer. And it does a better job of cooking our food. But we’ve kind of stalled, in my opinion, for the last couple of decades of really innovating and creating a compelling vision of what the future of the kitchen can be. I think the idea that our appliances are too stupid to know when to turn the temperature up or down to cook my food correctly is bizarre in the modern world where sensitive, high-quality sensors are cheap. And we have unlimited compute and AI now to answer a lot of these questions that humans struggle with, but I don’t see the big appliance companies or the incumbents doing this on their own. So, my small contribution was to create a tool that measures temperature and makes it very easy for people to do things with those measurements.”

On why it’s important to create a vision for the future of a technology-powered kitchen:

Young: “My criticism with a lot of people in this space is they haven’t sold a vision of what the future of that your kitchen could be like that resonates with people, that feels human, that makes it a place I want to go that is forward-looking rather than backward-looking. The kitchen of the 1950s, the kitchen of the 1920s, feels more human, feels more relatable, and I think people want that. It’s not to say you can’t create a forward-looking vision of a kitchen where it’s easier to cook food, it’s easier to bring people together and have everything work out right, but nobody’s really creating that vision.”

Combustion’s thermometer runs its machine-learning calculations on the chip within the thermometer rather than in the cloud where many AI compute happens. Young explains how – and why – they made that possible:

Young: “One of the crazy challenges was this is some pretty hardcore math. I think even we initially thought, ‘Oh, we’re gonna have to run this on the cloud, where we essentially have unlimited compute to run these fairly sophisticated algorithms.’ But we have some very clever software and firmware people on our team who have a lot of experience doing these kinds of hardcore machine-learning algorithms. And we were able to basically figure out some clever trick techniques to get the stuff running on the thermometer. The benefit is that it means the thermometer is always the ground truth; if you lose a connection, if you walk too far away, or if Bluetooth gets interrupted, or if any of that happens, the thermometer doesn’t miss a beat. It’s still measuring temperatures, it’s still running its physics model. So as soon as you reconnect, the results are there, and nothing has been lost.”

Young on the benefit of generative AI:

Young: “In the short term, AI as it’s being marketed is going to be disappointing to a lot of people. It’s going to burn some people in the way that IoT burned some people. But there’s going to be meaningful things that come out of it.”

“…When I was playing with ChatGPT 3.5 and I would ask it cooking questions, the answers were mostly garbage, as judged from my chef perspective. When GPT 4 came out, and I started asking some of the same questions, the answers were actually pretty good. I might quibble with them, but they wouldn’t completely fail you and they weren’t garbage. And if you modified the prompt to rely on information from Serious Eats, ChefSteps, or other reputable sources, all of a sudden, I might have given you a different answer, but it’s not necessarily better. And in many cases, what people want is a good enough answer. Building those kinds of things into the cooking experience where, when you run into a problem, or you’re confused about what this means, something like the Crouton app, or the Combustion app, or a website can quickly give you a real-time good enough answer, that actually solves your problem and keeps you moving forward and getting dinner done. Those I think will be really, really big payoffs, and that stuff’s coming.”

Young on whether big food and appliance brands should invest on building their own AI internal competency:

Young: “It’s hard to give advice when that’s not my business. But I have a few observations from having worked with these companies. It’s very hard to sustain a multi-year effort on something like an AI software feature. For these companies, that culture doesn’t exist, the way of thinking about the long term payoff of software tends to not be a strength of these companies. And so while they have the resources to go do this, the willingness to make those investments and sustain them, for years and years and years, and learn and iterate, that hasn’t proven to be their greatest strength.”

“I think that is kind of why there was an opportunity for Combustion, and for a company like Fisher Paykel (ed note: Fisher Paykel has integrated the Combustion thermometer to work with some of their appliances) to recoup the millions and millions of dollars, we’ve invested in the AI in our algorithms team. (Fisher Paykel) could maybe build the hardware, but doing the software, investing in the hardcore machine learning research, I think it would be very hard for them to sustain that effort for three or four years when they’re only going to maybe sell 12-25,000 units a year. We’re in a much better position because we can spread it across the entire consumer base.”

“And so I think you’re going to see more partnerships emerging between the big appliance companies that can provide the infrastructure, the appliance that’s got ventilation over it, that’s plugged into a 240 volt, 40 amp or 50 amp circuit. They’re going to be very good at that. If they basically open up those appliances as a platform that third-party accessories like the predictive thermometer can take advantage of, I think over the long term, they actually take less risk, but they actually get a market benefit.”

“Because as more small companies like Combustion can get wins by integrating with these appliances inexpensively and easily, making our products more useful, I think you’ll start to get a lot of things like the rice cooker no longer has to be a dedicated appliance that you put in a cabinet. Instead, it can be a special pot that goes on the stove. But now it can communicate with the stove to do what a rice cooker does, which is turn the power on and off at the right time. And now a lot of these small appliances can migrate back to the cooktop, they can migrate back into the oven.”

If you want to hear the full conversation with Chris Young, you can click play below or find the episode on Apple Podcasts or wherever you get your podcasts.

January 25, 2024

Jersey Mike’s Jumps on the AI-Voice Order Bandwagon as it Deploys Soundhound to 50 Locations

The AI-voice bot customer service wave is coming at us fast, and the latest chain to roll out the technology is the sub-sandwich chain Jersey Mike’s.

Soundhound announced this week that Jersey Mike’s will deploy its voice AI ordering system to allow customers to place orders by phone. According to Soundhound, the AI has been trained on the entire Jersey Mike’s menu and can handle order placement and answer queries about menu items, specials, store information, and more, all while ensuring orders are taken accurately and efficiently.

The video below shows Soundhound’s Jersey Mike integration in action.

DEMO: Jersey Mike's Automated Phone Ordering System - Powered by SoundHound AI

Like many voice AIs, it sounds about 90% natural in action, but there’s still something of an Uncanny Valley stiltedness to it. Listening to the order, it seems to handle the natural conversation flow deftly, but I have to wonder just how nimble it is with various dialects and slang, which can be a natural part of incoming phone orders.

Soundhound, a company that’s been around since the mid-aughts and had focused on auto installations and music until its move into customer service interaction layers for restaurants and retail over the past couple of years, announced the acquisition of SYNQ3 Restaurant Solutions last month. At the time of the deal, Soundhound said the merger extended “its market reach by an order of magnitude to over 10,000 signed locations and accelerating the deployment of leading-edge generative AI capabilities to the industry.”

Soundhound isn’t alone in chasing fast food chains to provide voice AI customer service platforms. Par Technologies, ConverseNow, and OpenCity also offer third-party solutions, while some players, like McDonald’s, have brought voice AI in-house through acquisition.

After a decade of pushing towards digital ordering kiosks and new ways to serve customers in-store and through apps, AI-powered customer service layers have moved to the top of the list for many big chains, including Jersey Mike’s.

January 22, 2024

Dodo Pizza Trials ChatGPT-Powered Flavor Generator in App

Dodo Pizza, a thousand-store pizza chain that’s built a name for itself by experimenting with different types of technology, announced last week that it was trialing a new “In-App Flavor Generator” powered by ChatGPT. The new generator, which is only available for now in the Dubai market, allows customers to create personalized pizza flavors from 35 different ingredients.

Here’s how it works: The app’s flavor generator, which uses generative AI technology, presents users with a choice of ingredients, all of which can be combined in different ways, which Dodo says can result in up to 30 million potential flavor combinations. This feature is designed to cater to individual preferences and moods, enabling customers to experiment with unique pizza creations.

For the launch, Dodo Pizza expanded its ingredient list with 17 new items, including some pretty weird flavors like popcorn, duck, guacamole, melon, fruit loops, falafel, and pumpkin seeds. The app’s interface queries the user’s mood and preferences, with options like “movie night,” and from there, users can customize their pizzas by specifying dietary restrictions or ingredient exclusions, such as vegetarian-only options or the omission of onions, pineapple, or spices.

The company, which was founded in Russia and originally grew to become Russia’s biggest pizza chain (but now is headquartered in Dubai), recently brought on a new CEO in Alena Tikhova, while company founder Fyodor Ovchinnikov has moved into the executive chairman role. Under Ovchinnikov, the company gained a reputation for embracing technology early on (Dodo was the first pizza chain to experiment with drone delivery), and this latest move continues that trend.

According to Dodo, early trials have piqued customer curiosity. The company says the AI “pizza card” generates about three times the clicks of other menu positions. Dodo says they plan to roll out the AI pizza generator in other regions this year, including Southeast Asia and Africa.

January 18, 2024

January AI’s New App Uses Generative AI to Predict How Food Will Impact Your Blood Sugar

If you’ve been diagnosed with a metabolic health issue, you might have used a continuous glucose monitor (CGM) at some point to track the impact of your food intake on your blood sugar. However, as of March 2023, only 2.4 million people used a CGM in the U.S., and because of the relatively small adoption rate of this technology, the vast majority of folks with diabetes or who are in danger of metabolic health issues may not have access to real-time insights into what the impact different foods may have on their glucose levels.

January AI aims to change this with its latest innovation: a free app that performs predictive analysis on the impact of various foods on blood sugar. The company, which unveiled its newest tool at CES last week, has developed an AI-powered app that analyzes meal photos and offers users immediate feedback on glucose impacts, macros, and healthier meal alternatives.

January says its app uses generative AI to automatically generate accurate food titles and estimates of ingredients and ingredient quantities within complex meals.

“It uses three kinds of generative AI to tell you your blood sugar response,” said Noosheen Hashemi, CEO of January, speaking at The Spoon’s CES Food Tech Conference last week. “It uses our own generative AI for glucose, and then it uses a vision generative AI to pick what’s in the food, and then it uses that language model to give it a title.”

According to the company, its AI-driven predictions are based on millions of data points, including wearable data, demographic information, and user reports. The company says this approach enables the app to provide personalized glucose level estimates and insights, making metabolic health management more accessible and actionable.

“It’s as simple as scanning a food,” said Hashemi. “You can also scan a barcode. You can also do a search. And we can tell you all the macro, its total calories, how much fiber, protein, fat, and carbs it has. And we can also show your blood sugar.”

According to Hashemi, the company’s platform can be customized and trained for specific users by taking data from a wearable such as a smartwatch, a person’s glucose monitor, or even food logs. With that data, the app can create highly customized predictions around a person’s biomarkers and dietary preferences.

“One out of three people in America has pre-diabetes, and 90% of them don’t know it,” said Hashemi. “And one out of nine people has diabetes, and 20% of those people don’t know it. So blood sugar is something we should all be managing, but we just don’t know that we should.”

Given the increasing popularity of GLP-1 medications, my guess is that more Americans will start to consider how their diet affects their blood sugar in the coming years. And, even if they don’t use a glucose monitor or get a prescription for a medication like Ozempic, increased awareness will push many to use apps like this one to help them better understand how a given food will impact their blood sugar and overall health.

You can hear Hashemi discussing the app and showing a demo in the video below.

January AI CEO Talks About New Generative AI App at CES
Previous
Next

Primary Sidebar

Footer

  • About
  • Sponsor the Spoon
  • The Spoon Events
  • Spoon Plus

© 2016–2025 The Spoon. All rights reserved.

  • Facebook
  • Instagram
  • LinkedIn
  • RSS
  • Twitter
  • YouTube
 

Loading Comments...