Bringing the Future Close: Randima Fernando on AI, Agency, and the Inner Life

An Awakin Call with Randima Fernando, co-founder of the Center for Humane Technology, hosted by Birju Pandya and Srinija Srinivasan.

Randima Fernando spent seven years at NVIDIA writing the book, literally, on the programmable graphics chips now driving the AI boom, then walked away from it in his late twenties to spend a decade building Mindful Schools and later co-founding the Center for Humane Technology. In this Awakin Call, he traces that unlikely arc, from a childhood where the precepts and the pixel were taught side by side, to a seat across the table from tech leaders who, when asked directly what happens to the jobs their own automation is built to erase, "suddenly become very obtuse." He is unusually honest about the compassion he still struggles to extend to the people making those choices, and unusually practical about what the rest of us can actually do: bring the future close, find where you have real leverage, and keep a well deep enough that the world's noise doesn't disturb it. It closes, fittingly, not with optimism or despair, but with something sturdier — intention, effort, and a bet that people rise when they finally hit the wall.

How to handle feelings of overwhelm and powerlessness related to AI: What's most helpful is knowing where you can be helpful. See things in a balanced way. At CHT, our goal is to paint that in a fair way, above the camps — a systemic view, an honest view.

But there are a lot of dumb things happening in the AI space, because the systems tend to tilt towards the harm. The rules don't keep up with the tech. This can be a little disturbing. Some practices anchor you on equanimity — for me, I listen to a lot of Buddhist talks and just get your mind in that kind of space. Something grounding. That does a lot.

Also, agency: trying to figure out what is the thing you can do if you care about any problem deeply. Find your place. That gives me a lot of peace. Even if the trend is not good, there's something I can do to be helpful. That spirit of service is a great protection.

Background growing up in a devout Buddhist family and early exposure to technology, intertwining threads: On the Dharmic side my parents were devout Buddhists and very knowledgeable. I took the precepts at age five. My mother would sit with us and say, here are the Pali words, and then she would translate and tell us what it means. They made sense to me. It was super helpful, because I had a very clear operating system, and I was never seeking — I was seeking to deepen my understanding of that operating system. It was very clear to me, and it has really worked for me. On the tech side, my father is an electrical engineer and loves playing around with devices, so as soon as there was a chance to get a computer, we got one, a green monochrome Apple. Over time he showed me programming books. I learned how to light up the different pixels, to draw things on the screen, and I thought that was so fascinating. Over time, as the computers became more powerful, this idea of controlling this machine that was very strong and could do stuff — that was very fascinating to me.

About the leap from NVIDIA to the inner life, to Mindful Schools: Early at NVIDIA, like a lot of Silicon Valley companies, you get to do stuff very early — age is not an issue. I was very lucky because I got affirmation quite early. By 25, I was looking at different non-profit-y things, starting to volunteer, and then NVIDIA launched their foundation. In 2007 I was on that board, and that got me really interested in where can we allocate money, how can we get employees to give back to the nonprofit sector, in expertise, not just money. I came across a project teaching mindfulness in schools, and that sounded really good. So I started volunteering for about two years, commuting to Oakland from Santa Clara. It dawned on me that if I quit, I could put all my time on the nonprofit work, and I could do a lot. So I sat with it for over two years, and eventually decided to give it a shot. It didn't feel as bold as it sounds. I had a big argument with my parents, they were really worried. "You studied computer science, you're at this good Silicon Valley job..." Apparently I was quite vehement. It didn't feel as risky.

Jensen himself said, "Come back if it doesn't work out." It was a lot of changes, but a good thing to try.

On systems, conditions, and why technology isn't neutral: Early on I saw the problem of attention. I didn't fully understand the downstream effects of social media on polarization, on the impact on democracy. Over time it's become obvious. From a Buddhist perspective, one of the fundamental things is that everything is conditioned. Things arise because certain conditions come together, and stick around until conditions change. So you start to see these systems are very connected, and govern the stuff that's in the systems, like the technologies we build. We have these Principles of Humane Technology, like think in systems, don't just optimize narrowly. Or protecting the systems we all depend on. Or understanding that technology is never neutral. Now, as the products become much more active, they're extra not neutral. Why are you building technology? To answer that well, you have to understand the human condition. What causes suffering? What are you trying to solve? Often we don't talk about those things.

Those leading the tech are aware of the effects and are building it anyway. How do you hold compassion for them? When we say it's the system, not the people, it doesn't mean there aren't bad people or people making bad choices. Facebook is a good example, where there's a lot of evidence, mainly from whistleblowers — things were raised, and they went up to the top, and the choice was to not fix it, repeatedly. Whoever is at the top would make that choice again and again. That is very common. The incentive structures really lock people into making bad decisions, but some of these were just bad choices, they wouldn't even cost that much revenue.

So I don't do that well on the compassion piece. I know the reality is, from a conditions-based lens, even the worst choices are perpetuated by the conditions that lead to the mind moment when you make the choice. I get that, but it's still hard. I just expect more. A lot of these tech leaders are extremely smart and capable, but when you have interview topics that are inconvenient, they suddenly become very obtuse. Just admit it: yeah, we're building automation technology, and in this new era we're going to need to redesign everything, because we're violating all the assumptions of the systems we built — so let's talk about that. That would be a much more interesting conversation.

What responsibility do the rest of us have in using this technology, particularly AI? It seems like we can't influence systems. To me, the strategy is to bring the consequences of the future close to us as fast as we can, to show what's going to happen if we keep going the way we're going. Jobs is the easiest one. AI slop. Knowing what's true or not. There are a few things you can do: affect legislation, write to your representative, but even there, think about whether you have a path to those making decisions — do I have a path, do I know people who know them? Affect litigation, harms that are not accounted for: research what's going on and make it public. Or if you build technology, build the right products and show they're actually the thing people want. Be aggressive in creative marketing against dumb choices. Being a whistleblower is a difficult choice, but there are nonprofits that can support people who are considering that choice. There are various forms of agency. Each person has a different role.

Preeta Bansal talks about this idea of the governance gap: every year, technology is exponential and governance is linear, so it's only going to get worse. What does that mean for us as responsible citizens? Yes, I completely agree with this idea of the governance gap, and there's also the issue of regulatory capture. Not only can governance not keep up, new versions of tech are rolled out all the time. So one of the things you have to do is upgrade the governance mechanism, to be able to iterate faster. I'd claim strongly that AI is the opportunity, the reason, to rethink the systems. You're forced to do it. So one of the strategies is to apply strong public pressure. Why are people so latched onto data centers? Because they represent a physical instantiation of a problem you're fighting — you can put your finger on it. AI is much more abstract.

Different views held by those close to AI, e.g., slow down, speed up, expand who the players are, make sure China doesn't win... How do you try to frame all these views? The overall principle is matching power with responsibility. Concentration: you just want a small number of companies to control everything. Diffusion: make this tech available to everyone, democratize everything. Obvious trade-offs with both. Closed, no transparency, accountability is difficult, regulatory capture — versus malicious actors get a huge advantage. There's an inherent mismatch between offense and defense. Even if you have software that can find the bug on both sides, the offense just has to use the bug to get in, but the defense has to fix the bug, and that takes much longer.

Another principle: the answer is not to just do everything as fast as you can. The people who want to speed things up say, just let things break and fix them — but the problem again is this offense-defense asymmetry. You create problems that are very hard to fix after the fact.

And there's the principle of protecting systems we depend on. If you break epistemology, how we know what's true and not true, how we make sense of the world — if you break that, you need epistemology to fix things.

Can AI serve inner transformation? Absolutely. One thing it does very well is external pedagogy. It's very well informed, there's lots of training material, that training is reinforced from many different pieces of material that say the same thing. So that's easy and quite helpful. Because how much access do most people have to a qualified teacher? After a talk, or in a retreat, you get very little time to ask questions. So it's a model where you're really starved for expertise. And a lot of the models give very good answers on many of the topics.

How about the inner? A lot of the newer models give quite good comments if you share your practice. If you use it well, it can really help you with concrete things to practice. It can pull from what's relevant at the time, some text that might not come to mind immediately. But you have to be really careful, it's easy to go off the rails — if you start talking about some personal vendetta, it will go down the rabbit hole and engage very willingly. If you're narcissistic, depressed, or angry, it won't regulate you in the way a human would. If you can choose, you'd always choose to have a very experienced human. A lot of teaching is through role modeling. If you can see someone who's really conducting themselves well, and they tell you something, you'll absorb it better. But we have to be realistic about access being a challenge.

Where can the chatbots be helpful? Stick to a domain that's narrower — give it a good prompt on sticking to practice and my experience, help me stay on track, and let's not go off track. And then check in with a good teacher from time to time to make sure. I'd also recommend not doing very long chats, start new chats relatively often, because when the chats go really long, things tend to go off the rails. When a narrow question comes up — I want to think about how to navigate this incident well, give me some options — and then I'll think about it and talk about it with my wife, with my parents. It's good fodder for speaking about it with another human, but also for internal reflection.

On authenticity and vulnerability: I really like the definition of authenticity as the gap between the inner and the outer.

You may say the wrong thing once in a while, but it just makes it much easier to operate. In management there are challenges to authenticity because of asymmetry of information — you can't say exactly the thing, for legal or HR or finance reasons, it can be sensitive. That's a tension that's a little bit unavoidable, and partly why I've wanted to get out of the ED roles, and I've been doing individual-contributor kinds of stuff.

On vulnerability, it's related — you have to feel safe to be vulnerable. At home I'm extremely safe because I trust my wife fully. You can't go out on the street and say the same things, it wouldn't make sense. But it does make sense to try and make the environments you are in support vulnerability at a level that's healthy. And that's different in the home, in the family unit, in the community unit, in the workplace, in public speaking — each place is different, but finding out what's right for you is a good exercise.

In Buddhism we have this thing called bhava tanha, which is wanting to become. People want to become certain things, and often we use the word "thirst" there, because you want it badly. It's really good to figure out what you're wanting, what you're attached to, and as you let go of those things you can become more vulnerable, because you're protecting less. Then there's one level below that — are you letting go because you're holding on to the things you've already done? You say, I can let go now. This is something I think about a lot. I try a lot at work to be someone who's not seeking "professional advancement" or the big pay raise — that should make me easier to work with, or more helpful. I'm checking all the time: am I achieving that, where am I getting defensive, why did that comment feel so pokey. I do spend a lot of time looking at the mind and the reaction to things.

On living into a future we haven't imagined before: Often we're like, "the future is hard to imagine," but you can imagine common sense things, like food, clothing, shelter, medicine — like we've got those things, we have reasonable education, when you get old you don't have to worry about being exploited, when you're in school you don't have to worry about a sexualized deepfake. There are some really reasonable things we can lay down as, that's what we should aim for. The stuff most people would find reasonable is actually a very good future.

An intention or prayer you hold? One thing I've learned over the years is the systems are very complex, so even if you think you have a plan for something, it's going to play out differently. So it comes down to intention and effort, which come from the Eightfold Path. Think about how you can be of service, set the intention, and then work hard to do that thing. But it's just one input in a very large system, so there's an automatic humility that comes in. A lot of times people ask, are you optimistic or pessimistic? I don't think that way. It doesn't match with the intention-effort thing. What do you think it's going to take for things to go well, and what are we going to contribute to that — I think that's a better way of asking the question, because it encourages you to go do something that would be helpful. I'm not optimistic that this is going to go well, because the default is that it's not going to go well. You can't look at these systems and honestly say, I think it's gonna be fine.

I think that's the most optimistic version that I would see, and I want to help accelerate that process. But even if you don't believe that, helping to reduce suffering at every level is a really noble and useful thing to do. All the time. It's just a good thing.

In closing, Randima offered a direct invitation to anyone listening with real leverage inside a company, a legislature, or a boardroom to reach out to him at randy@humanetech.com, and pointed listeners toward humanetech.com/course for more on the principles of humane technology his team has developed.