Opsani Corner is the place where we pull back the Opsani curtain and share stories about our team and our internal goings-on. This month, we sat down for a chat with Eric Kalosa-Kenyon, one of our talented machine learning engineers.
Hey Eric! Let’s start simple. How would you explain machine learning to a 10-year old?
You know how Netflix knows what you like to watch? That’s machine learning. You know how each Minecraft world is unique and interesting? That’s machine learning. It helps us find planets in deep space, design the rocket ships to get there, and do just about anything people already do – just faster and better.
How did you come across Opsani, and what made you decide to join the team?
Peter found me. Joining a company is always such a fluid, multifaceted thing – it was a good fit.
What challenges do you think you’ll be facing at Opsani, moving forward?
Hiring is hard. But we need all the talented people we can find as we’re growing so fast. Please check out our careers page and send along your resume!
What do you find most exciting about Opsani as a product and a company?
In the US, data centers account for as much carbon emissions as air travel (~2% each). Opsani saves compute cycles, which saves both on cloud and carbon.
In what direction do you want to take Opsani’s evolution as a product? Which capabilities do you want to add or develop in its current system?
We can probably optimize more than just computer infrastructure. But I’m not going to push Ross or Peter on that! I admire their business sense and appreciate that we need to focus on our bread and butter first.
An apocalyptic fear of AI is quite rampant in pop culture. What are your thoughts on this?
We can be confident that AI, in concert with existing military technology, will enable new capabilities exceeding those enabled by existing tech alone. What’s more, the barrier to entry for weaponizing AI is relatively low in comparison to other advanced weapons like nuclear ICBMs. For example, both the US government and the Taliban control weaponized (semi-)autonomous drones, albeit with different designs. If you haven’t seen the Black Mirror clip about small drones with facial recognition, I’d recommend it – it’s entirely feasible.
We should be extremely concerned about asymmetric state-controlled AI capabilities. With AI, not unlike the one Google’s DeepMind has trained to win at the strategic war game StarCraft, we have the capability to incinerate hundreds of thousands of people, quietly target specific ethnicities for re-education, cure cancer, provide universal access to clean water, or – most likely – some combination of the above.
Learn more about Opsani in our blog, What Makes Opsani Unique.