Howso is a fast-growing technology company that’s building an understandable AI engine for use in diverse industries. Howso’s instance-based machine learning platform is fundamentally different from the neural networks powering most black-box AI solutions, including ChatGPT and Google Bard. Their AI engine allows organizations to build explainable AI applications; users can understand why and how the AI makes decisions, intervening when necessary. Calibrate first invested in Howso in May 2021, leading their $2.5 million seed round, and, impressed by their visionary founders, powerful technology, and strong market traction, we supported their $25 million Series A round last September. Calibrate recently sat down with Howso Chairman and Co-Founder Dr. Mike Capps to discuss the burning need for ethical AI, why having mentors matters, and how to productize your early technology for future growth.
What drove you to start a company?
I had such a fun career in video games, culminating in my role as President of Epic Games. But I had kids on the way, so I retired in 2012 to focus on being a dad. Once they hit school age, I was getting itchy, and I knew I wanted to work on something with a positive social impact. I made a list of three areas that I thought were important enough to take me away from my family. One of those three was black-box AI, and how our society was heading toward leaning more and more on AI that was fundamentally inappropriate for use in critical decision making.
My kids won’t drive a car for themselves. AI will decide if they get into college, get a job, receive medical treatment, and probably even who they should marry. And all those systems right now are black boxes, which is terrifying. My great friend Chris Hazard had been working on explainable AI for years and had a breakthrough, and we decided we’d work together to make AI trustworthy, ethical, and safe. That was in 2017. Few others in the industry were worried about the dangers of AI making mistakes back in 2017, but, suddenly, almost everyone is buzzing about Chat-GPT “hallucinations.” So the time’s right, and our tech is ready.
Looking back to your company’s earliest days, what is one thing you would have done differently?
Likely, take more risks. Chris and I are both scientists and cautious by nature. We were focused on building powerful, flexible technology. But we likely missed some opportunities to productize for specific business problems. For example, we would speak to one prospect about using it for submarine fleet readiness testing, and then to the next one about how to use it to improve financial services, and then we’d work on a system to predict foreign exchange markets. Mind you, every one of those projects was a success, and our tech was improving—but that approach was too scattered and didn’t allow us to build a scalable, repeatable go-to-market strategy. We also found ourselves with a few early big-ticket licensing deals, which meant we got very protective with our software IP. We know the best way to impact everyone’s relationship with AI is to make explainable AI more available, and we should have started working on that earlier.
On the flipside, what is one thing you got absolutely right?
From day one, my co-founder Chris and I agreed on what matters, and we haven’t wavered from that: building AI with social impact. We’re here to replace black-box AI with something that’s equally powerful, but also understandable, transparent, and controllable. We believe we can also build a successful long-term business that brings AI to companies in multiple sectors, since our technology is powerful and scalable. But we’re about impact first, profit second. We make every business decision through that lens, and all our employees and investors are “warned” up front and excited about our mission.
What is one milestone you’re particularly proud of?
For toolsmiths, the only milestone that matters is the tool being used. It’s gratifying to see global financial institutions, world-class hospitals, and governments deploying Howso tech at scale. For example, one state government agency is using our software to improve behavioral health treatment for inmates. A massive insurance company in Europe which is legally required to delete claims data after three years uses our synthetic data to make “privacy-free backups” of their data. They can retain the synthetic data indefinitely, which lets them do the analysis they need to make their business more efficient. And we’ve been active in defense projects since before the company’s official founding date.
If you could give one piece of advice to a new founder, what would it be?
Oh man, I’m still writing my own founder story. I guess I’ll say that this is a roller coaster, by definition. It’s going to have great days and hard days, and that’s all natural, and just enjoy the ride. Having been on a wild growth path with Epic Games, it’s the same darned thing again leading an A-round startup. Also, working with people you trust is absolutely huge. I’ve worked with Jason Schoettler at Calibrate Ventures since long before there was a Calibrate. We agree about 70% of the time on business direction, but we agree 100% of the time on what’s ethical and the right way to treat people—the real things that matter.
Who do you admire and why?
There are so many people who have helped me on my path, but it’s a few special ones whom I admire and appreciate the most. Back in 1980 in my public school, I had a third-grade math teacher who quickly saw my potential and let me write computer code during recess. And in middle school, my math teacher started pulling problems out of college textbooks for me. Then in college, a professor encouraged me to pursue computer science when he saw I loved the idea of problems that can be proven unsolvable. I was a smart kid who worked hard, sure, but without teachers who went the extra mile, it wouldn’t have been the same path.
What does success mean for you?
Success for me is measured by impact. Trying is nice, but it doesn’t really matter unless you’re trying to get better at trying. So, I’m trying to use my D&D stats, such as they are, and my weird experience to help humankind. Our tech helps find the wisdom hidden in data–basically, the “truth” in the facts. There’s so much we can do with that wisdom to accelerate discovery in many fields to create a lasting positive impact.