Two Standard Cognition Founders on Changing the Face of Retail — And Reclaiming Humanity’s Most Precious Resource
How many hours have you spent standing in line so far this year — and what could you do if you got those hours back? For the team at Standard Cognition, making in-person shopping seamless means not just removing friction from your next trip to the grocery store, but fundamentally changing the way people spend their time. Below, co-founders Jordan Fisher (CEO) and Michael Suswal (COO) explain the challenges their team is tackling, how they think about automation and the ethics of computer vision, what makes them grateful to be competing with Amazon, and why they believe transforming retail is just the beginning.
First, give us your vision for the future. How will the world be different if Standard Cognition succeeds?
Jordan: Ultimately, we’re building towards reclaiming literally billions of hours people spend waiting in lines every year collectively. It’s crazy that we’re wasting that much time — it’s our most precious resource. I think getting it back will be a windfall for humanity.
Of course, it will also change the time and manpower we expend on retail work. We’ve been through similar changes in the past — before the Industrial Revolution, three-fourths of the population worked in agriculture; now it’s about 2%. I think we’re better off today and will ultimately be even better off because of automation. But it’s important to understand that transitions like this are really rough for a lot of people, and we need to get there as responsibly as we can. No one has answers yet, but the first step is to talk about it, openly and honestly. That’s the thinking behind the automation pledge we’re writing — companies like us have a duty to be part of that conversation.
Michael: Something we also believe we can help with in the long term is making things easier for smaller retailers. It’s incredibly challenging to run a shop without a large workforce. Store owners are staying open 24/7 but can’t afford the staff — it’s such a brutal business.
How did you and your co-founders decide on this product?
Jordan: It started back when Michael and I were still at the Securities and Exchange Commission. We were working with all these amazing people, building the SEC’s first machine learning infrastructure. And I knew I wanted to start a company in the computer vision space at some point. So I started a research discussion group and invited all those amazing engineers. We read different papers and talked through them once a week, and we settled on this topic about a year in. Making retail frictionless was a fascinating problem, and the technology was not quite there, but close enough to be practical.
Michael: It was also a huge addressable market. That’s a misconception I had before we started out — you hear so much about the growth of e-commerce. But brick-and-mortar retail is actually growing year over year, too, in terms of both number of locations and revenue. And even if you think of it as one pie, physical retail is still almost 90% — it’s a $25 trillion per year industry.
Tell us about the challenges your team is working on.
Jordan: On the technical side, there’s obviously a lot of machine learning work to do. We’re productionizing computer vision in the real world more than anyone else has, so there isn’t a playbook for everything yet. It’s still kind of the wild west — we’re defining best practices for a new industry.
“We’re productionizing computer vision in the real world more than anyone else has, so there isn’t a playbook for everything yet. It’s still kind of the wild west — we’re defining best practices for a new industry.” — Jordan
But at this point, the machine learning piece is actually some of the more straightforward technical work we do. The questions we’re tackling now are how to integrate it into an engineering system, into an account management process, into the entire mindset of an industry. How do you get this into not 10 or 100 stores, but 1,000? How do you design a backend system that solves the unique challenges of this space? The machine learning piece is like the laser in a barcode scanner. When barcode scanners were introduced around 70 years ago, that was cutting-edge physics. But the most interesting work came after that, trying to figure out how to make it work reliably in the chaotic, real-life environment of the front of a busy store.
How do you think about the ethics of computer vision?
Michael: A lot of people who hear about Standard Cognition assume we’re capturing biometric data and doing facial recognition, but that’s definitely not the case.
Jordan: Yeah, we take a pretty hard stance on that. The way I see it — if we succeed, we’ll be putting cameras all over the world. And with a system like that, either you design it so it can’t be abused, or it eventually will be. So it’s not “we do facial recognition, but we’re really careful about it.” It’s that our system is fundamentally incapable of doing facial recognition at all.
People also assume that training machines for this work means you’re teaching them to be biased, as if we’re trying to predict what a shoplifter looks like. But it’s the exact opposite — the only thing our system cares about is whether it scans a product or not. It’s actually much more neutral than a human.
“The only thing our system cares about is whether it scans a product or not. It’s actually much more neutral than a human.” — Jordan
Amazon recently licensed its Go technology. How will Standard Cognition compete?
Jordan: The first big difference is Amazon’s technology is very expensive. Standard Cognition is essentially a retrofit, so it has to be cheap and flexible enough to easily deploy in an existing store. With Amazon, everything is custom, down to the shelves. That costs millions, and the end result is turning your store into an Amazon Go store in everything but name.
Even if you get past the price tag, though, I think the biggest difference is that for a brick-and-mortar retailer, Amazon is their biggest competitor, and using their technology means handing over incredibly valuable data. Toys R Us is a great example — they partnered with Amazon, Amazon started to carry everything they had in their physical stores for a lower price, and the stores went bankrupt. So I do think there’s a lot of room in the market for solutions that don’t require you to work with your competitor. We’re actually grateful for Amazon’s presence in this space, because we can show how we’re a foil for their corporate tactics.
What are you excited about in the years ahead?
Michael: On the product side, it’s exciting from multiple angles. We’re fundamentally redefining how people shop, and every part of the company will have a say.
But I think it’s bigger than that, too. Right now, every other industry is watching what we do, because they know computer vision and AI are ultimately going to have an impact that goes far beyond retail. It’s autonomous driving. It’s manufacturing. It’s anywhere where eyes and brains are useful. We’re contacted every week by camera companies, packaging companies, payment processors, cloud providers — and they’re all trying to figure out how to service this industry. I see this as a revolution in the same way the internet was. Computer vision is going to change the way we interact with the physical world, in retail and beyond, and we get to help lead the way.