There’s an old saying in business: being early is the same as being wrong.
That applies to pretty much any market, but especially kitchen tech. I should know: over the past decade, I’ve watched dozens and dozens of startups try to build the next big thing in the kitchen, both through my event, the Smart Kitchen Summit, and in my coverage for The Spoon.
Of course, some were just plain bad ideas that wouldn’t take hold no matter when they showed up. But many were just too early, and the technology simply couldn’t deliver on the vision.
Which is why, when I got together last week at SKS 2026 with two founders in the space – Chris Young of Combustion (and formerly of ChefSteps and Modernist Cuisine) and Tom Klaff of Revolution Cooking – I wanted to better understand how many of those ideas were fundamentally flawed and how many simply arrived before the technology could deliver.
For Young, the question was especially relevant. Much of what he’d tried over the past 10 years, first at ChefSteps and later at Combustion, was far ahead of what others were attempting. The technology simply let him down.
One of his early experiments involved using a Facebook chatbot as a cooking interface, well before tools like ChatGPT emerged and when companies were just starting to use voice assistants like Alexa to control appliances. The thinking was straightforward. Someone making dinner knows how they want their food to turn out, but translating that preference into a temperature, cooking time, and sequence of steps requires knowledge they may not have. Giving them a way to simply explain what they wanted seemed like a promising solution.
The interfaces, however, weren’t ready.
“It’s a good example of it failing to get the adoption we would have liked, not because people hated the idea; it was because the technology just sucked,” said Young.
The founder of Combustion believes companies should now revisit those early experiments and distinguish between ideas consumers never wanted and those held back by the available technology. In his view, recent advances in AI are making some of those ideas more practical.
One longstanding challenge has been the math behind cooking time and carryover heat, the rise in temperature that continues after food leaves the oven. Young said he’d spent more than a decade working on that problem, often with the help of accomplished mathematicians, investing more than seven figures along the way. Now, he says, AI is helping his team make progress on problems that had resisted those efforts.
“How long will it take something to cook?” asked Young. “When I pull it out of the oven, how much is it going to overshoot? What’s blowing my mind is just in the last couple of weeks, it’s become solvable. I have taken the millions of cooks we have, I have been using training AI models, and we can answer that question now.”
The benefit he described is easy to understand (especially for me): helping someone decide when to take action so dinner turns out the way they want.
For appliance companies, the maturation of AI raises a question they will increasingly need to wrestle with: Who will provide the intelligence that helps people cook? Young thinks that, in the future, AI will coordinate activity across multiple appliances. An assistant that understands the whole process could become the place consumers turn for guidance.
“The AI layer is going to exist on top of a lot of appliances and help orchestrate those experiences.”
What does that mean for appliance brands? Will they see their value go down if someone else controls how people interact with their appliances?
As AI interfaces mature, appliance brands could end up supplying the hardware while an AI company controls much of the interaction with the cook. Young thought the risk to appliance makers was they could surrender the opportunity to companies such as Apple and Meta, which can bring new software experiences to consumers without waiting for them to replace an oven or refrigerator.
During the panel, we also discussed why desirable features like steam cooking haven’t gained more traction in home kitchens. According to Young, combiovens can produce excellent results in the hands of experienced cooks, but they also require decisions many home cooks don’t know how to make.
“And part of it is nobody knows what to do with the steam or the humidity control. But robots and AI are like really good at making those decisions. And if you know how to use it, you get phenomenal results.”
If AI can handle those choices, an appliance that previously required considerable learning could become useful to a much wider audience. The hardware wouldn’t necessarily need a new capability. The cook would need a better way to make use of what it already does.
This vision fits with Klaff’s belief that the person using the product should remain at the center of its development. “The human being is still the most important thing,” said Klaff. “And we’re designing experiences for the human being.”
For years, kitchen companies have offered consumers more precision and new ways to cook. According to Young and Klaff, as technology matures to deliver on the promise of audacious new approaches to cooking, the next opportunity could be helping people use those capabilities without first becoming experts.
You can watch the full session below.