Jorge Torres of MindsDB On The Future Of Artificial Intelligence
Using technology to bring goodness to the world is really about making sure that two things happen. One is to make sure that technology such as AI is democratized and fully accessible. The other is making sure that the process of democratization is ethical and takes into account everything that might go wrong.
As part of our series about the future of Artificial Intelligence, I had the pleasure of interviewing Jorge Torres, CEO and Co-Founder at MindsDB and visiting scholar at UC Berkeley.
Prior to founding MindsDB, Jorge worked for a number of data-intensive start-ups, most recently working with Aneesh Chopra (the first CTO in the US government) building data systems that analyze billions of patient’s records that led to savings for millions of patients. He started his work on scaling solutions using Machine Learning in early 2008 while working as the first full-time engineer at Couchsurfing.
Thank you so much for joining us in this interview series! Can you share with us the ‘backstory” of how you decided to pursue this career path in AI?
I believe that there is enormous power in data. The more a company has, the more they’re able to propel their businesses forward. But only if they’re able to get meaningful insights from it.
In the fall of 2017, my best friend Adam Carrigan (COO) and I came to the conclusion that too many businesses faced limitations when it came to extracting meaningful information from their data. They realized that one of the biggest limitations was in how many of these businesses were severely under-utilizing the power of artificial intelligence. We believed that machine learning could make data, and the intelligence it can provide, accessible to everyone so we designed a platform that would allow anyone to use the power of machine learning to ask predictive questions of their data and receive accurate answers from it.
We called this platform MindsDB, it is backed with over $7.6M in seed funding from Walden Catalyst Ventures, YCombinator, OpenOcean, the venture fund launched by the creators of MySQL and MariaDB, SpeedInvest, and the University of California Berkeley SkyDeck fund. MindsDB is also a graduate of YCombinators recent Winter batch, and was recognized as one of America’s most promising AI companies by Forbes Magazine, one of the 8 most innovative AI and ML companies by TechRepublic, and recently nominated as a “Cool Vendors in Data for Artificial Intelligence and Machine Learning” by Gartner.
We’re focused on continuing to make it incredibly easy for businesses to extract information from their data by creating tools and platforms that enable everyone to become a ‘data-scientist’ and democratize machine learning along the way.
What lessons can others learn from your story?
Well, one of the most important lessons we learned came from the debate over whether or not to go open source. When you’re building something new, it often makes sense to go open source because you get a lot of validation that would otherwise take a lot of time and effort. You might have a great idea in mind, but not necessarily know how to get there. By leveraging the open source community, you get a lot of active feedback that you don’t have access to when developing in a vacuum. So that would be the biggest takeaway from our journey — if you’ve got a product that is quite ephemeral in scope and doesn’t have a standard benchmark that allows you to validate it, making it open source is a great idea — and also a great honor. So if you’re on the fence about whether to make your product open source or not, the odds are you probably should and then you’re free to start thinking immediately about turning that open source product into a viable business.
I’d also say, consider business accelerators. Sometimes it’s hard to think about giving away so much equity for something fairly intangible, but those intangibles can be invaluable on your path to growth. Having someone that can offer you immediate validation and financial backing can help you advance in your journey, and they’re usually the kinds of people that can teach you the ropes and make you think differently about things from a business perspective — that’s hugely important if you’re operating somewhere competitive like the Bay Area with lots of venture backed businesses all competing for attention.
Can you tell our readers about the most interesting projects you are working on now?
Oh, absolutely. I think one of the most exciting things we have going on at the moment is extending in-database machine learning to cover all sources. Businesses today have very diverse data stacks; they don’t just marry into one technology. Their data flows around all of these different technologies depending on how data is being used or how it needs to be presented. We’re looking at how machine learning can work across all of these different technology stacks to add value right across an organization. Really, machine learning is an end-to-end process — you need data analysts to be able to forecast, you need engineers to be able to automate those forecasts, and you need decision-makers to get value from those predictions in order to make better business decisions. When we started out, our intention was to just bring machine learning to databases, but we quickly realized that people wanted machine learning predictions from databases, data lakes, data source connectors, and even while streaming real-time data using tools like Apache Kafka. We’re the glue that binds machine learning with data, no matter where that data originates or flows.
We’re also looking at some pretty cool integrations with e-commerce and SaaS data sources like Shopify, as well as integration with visualization tools like Tableau and Looker. Workspace integrations with tools like Slack are also well underway, moving businesses to not only a 360-degree view of their data, but an environment in which decisions pertaining to all aspects of their organization can be made swiftly with deep and detailed intelligence served up automatically.
One specific business example is Domuso, a next-gen financial services platform offering comprehensive payment processing solutions for multifamily properties, that wanted to improve decision making using collected data for forecasting and risk assessment. Using MindsDB the existing team were able to quickly train, deploy, and leverage ML models that reduced chargebacks by $95k over two months saving the company approximately $500,000 a year.
None of us are able to achieve success without some help along the way. Is there a particular person who you are grateful towards who helped get you to where you are? Can you share a story about that?
There are so many individuals and organizations that have helped us along the way. Our work involves so many different technologies and we rely on brilliant collaborations with the makers of those technologies. Beyond that though, I’d say the open source community has helped us more than any single person or company. The open source community is the driving force behind any idea, and if not for their input I don’t think MindsDB would be where it is today.
What are the 5 things that most excite you about the AI industry? Why?
The first thing is knowing that we’re only just scratching the surface of what’s possible. When we started MindsDB, we were pioneers in the machine learning automation space, doing research on it at Berkeley. People were kind of laughing about it and dismissing it at the time, but now the idea that it isn’t possible is just as laughable. That’s how rapidly things have changed.
Second, is the capacity we now have to solve really complex problems. The sheer complexity of the data that these models can now handle is overwhelming. It’s about more than numbers and labels, and now encompasses complex text and images. The fact that you can now feed some very natural language into a text-to-machine algorithm and that machine will extract meaning and make predictions based on it is lightyears ahead from where we were even seven years ago.
I suppose that leads nicely into the third thing, which would be the rise of conversational AI. Compare a chatbot from 20 years ago to a chatbot today and the results are incredible. Imagine where we’ll be in another 10 years’ time?
The fourth thing I’d say about AI is how it’s going to free up humans to devote our attention and cognitive power to more important things. Instead of spending countless hours combing over data, making comparisons and creating arguments, we can automate that form of insight and intelligence gathering and focus on top line objectives and the big decisions we’d rather make ourselves. We’re basically augmenting the decision-making process to make it easier and more efficient. I read somewhere that humans only have capacity to make a certain number of good, productive decisions in any given day — so let’s allow machines to take care of the smaller, less important ones so we can focus on the big ones.
The fifth, and perhaps most important innovation in AI, is the rise of explainable AI or XAI. This basically asks AI to demonstrate it’s working out, so that humans can easily interrogate decisions made by AI and get full transparency over an AI’s decision-making process. This is going to be crucial for combatting things like biases and reassuring those businesses that might be on the fence about using AI for important things like recruitment or compliance.
What are the 5 things that concern you about the AI industry? Why?
Ultimately, AI is created by humans and therefore susceptible to the same mistakes and vulnerabilities. That’s why concepts like explainable AI, which I mentioned in my previous answer, are so crucial. As artificial intelligence gets more and more complex, we need ways to moderate it effectively and make sure mistakes aren’t being made. We also need to empower field experts to be involved in the development of complex AI that will be deployed in their domain, to minimize the risk of errors being amplified as processes are automated.
Another thing is that as AI and machine learning get more complex, there’s a risk that might become a technology that very few organizations feel they can access or adequately control. We need to start democratizing the tech, pulling back the veil and doing our best to make sure it’s transparent and easy to deploy. In many ways, our mission at MindsDB is to counter the increasing complexity of AI and make it more simple to adopt and control.
Another concern would be the understandable apprehension around the use of AI in delicate or high-risk use cases like investment, compliance or healthcare. We need to make sure that we educate people on the benefits of AI, because if we don’t and there’s a backlash due to fear or misunderstanding we could end up over-regulating it and bottlenecking progress.
Counter to that, there is of course a natural concern around over dependence on AI, or even mis-use of AI for malicious or nefarious purposes. There needs to be an ethical framework around the development and use of AI to help us guard against such things, but that framework must be applied in a way that doesn’t stifle or limit innovation. Ultimately, AI is about scaling the human mind and accentuating human ingenuity, but it also has the potential to change the way we work and live as a society.
As you know, there is an ongoing debate between prominent scientists, (personified as a debate between Elon Musk and Mark Zuckerberg,) about whether advanced AI has the future potential to pose a danger to humanity. What is your position about this?
I already touched on this, but there’s definitely an ethical component to AI that we need to work hard on as a society. There are key principles that we need to stick to, such as ‘do no harm’, preserving human wellbeing and dignity, as well as preserving the planet and operating in a sustainable manner. AI could potentially pose a risk to all of these things, so it’s going to be vital that there’s a human component in there to moderate AI and steer things in the right direction. Most of us would happily get on a plane from San Francisco to Sydney with AI doing much of the flying, but we still want the pilot there, even if they only need to take control for a few minutes. It’s the same with AI on a grand scale — we still need humans in the loop to counter things like biases and stand up for things like fairness and inclusivity — things that self-teaching AI can easily get wrong through the reinforcement of negative patterns. AI itself doesn’t have an agenda, it’s a tool that we harness and like any tool it needs careful calibration and oversight.
What can be done to prevent such concerns from materializing? And what can be done to assure the public that there is nothing to be concerned about?
We’re almost certainly going to need to course correct a few times throughout our journey with AI. Technologies like explainable AI will give us the awareness we need to do that on a micro scale, but we’ll also need to think about how we adopt AI on a macro level too — government policy, regulation and the like. This will no doubt require investment and innovation, with new compliance frameworks materializing as use-cases increase. We’ll need to come up with these frameworks from a humanitarian perspective, not just a technological or national one.
How have you used your success to bring goodness to the world? Can you share a story?
Using technology to bring goodness to the world is really about making sure that two things happen. One is to make sure that technology such as AI is democratized and fully accessible. The other is making sure that the process of democratization is ethical and takes into account everything that might go wrong. For us and MindsDB, AI isn’t about replacing humans or taking anything away from people, it’s about augmenting intelligence and elevating human capabilities.
As you know, there are not that many women in your industry. Can you advise what is needed to engage more women into the AI industry?
Well for a long time one of our principal researchers at MindsDB was a woman named Natasha, and we were very sad to see her go back to her research after she helped us achieve so much. There’s definitely a disparity now between traditional tech engineering roles, which are still predominantly occupied by men, and machine learning engineering where we’re seeing an overwhelming number of women candidates materializing.
For every woman that applies for a typical tech engineering job there seems to be a hundred men, but with machine learning — particularly when it comes to heavy research roles — I’d say maybe 30–40% of candidates putting themselves forward are now female. And a lot of those women hold PhDs from top universities. So the gap is closing, but there’s still work to be done. I think we need to give women in tech engineering roles more exposure and present them proudly as leaders in the field to inspire other women to get involved. We need to break the notion that engineering is somehow a man’s world, because it absolutely isn’t. I’d like to think that in 5–6 years time, once we have another couple of generations worth of students graduating, there will be a lot more balance when it comes to diversity and gender representation in our industry.
What is your favorite “Life Lesson Quote”? Can you share a story of how that had relevance to your own life? Work your passion, never work a day in your life.
‘Make your obsession your profession, and you’ll never have to work a day in your life’. A lot of us at MindsDB are blessed to be doing something that we’re passionately interested in — we want to make it work and get it right — so it never really feels like work.
You are a person of great influence. If you could start a movement that would bring the most amount of good to the most amount of people, what would that be? You never know what your idea can trigger. :-)
Machine learning has this incredible potential to solve some of the world’s biggest challenges. I’m talking about huge, existential issues like climate change, health and education. Machine learning can make a real, tangible difference here, but only if the technology is well understood and accessible to every person and every organization. So I think the movement I would start would be the movement that MindsDB has, in a sense, already begun. The democratization of machine learning. Taking incredibly complex and powerful technology and making it accessible and easy to use for all.
How can our readers further follow your work online?
As mentioned before, MindsDB helps anyone use the power of machine learning to ask predictive questions of their data and receive accurate answers from it. To see how MindsDB can help you visit www.mindsdb.com or follow us @MindsDB.
This was very inspiring. Thank you so much for joining us!