Understanding Atoms & Bits: Harnessing the Superpower

I was at a dinner a few weeks ago where each guest was asked to introduce themselves by saying what their superpower was. Fortunately, I had a few people up before me so I had time to think. My superpower? Changing diapers at 3am? Nope, most parents have to do that. Conducting team meetings while walking the streets and playgrounds of San Francisco? A nice skill, perhaps, but hardly a superpower.

I finally settled on something that I think is pretty cool, and certainly applies to what we are doing at TC Labs: I understand both atoms and bits.?

Atoms means I understand the complexity of the physical world. I got a mechanical engineering degree in college and spent the first part of my career in the energy industry. I get first hand the complexity of factories taking raw materials in one physical form (like crude oil), running them through a conversion process (like refining), and delivering a different physical form that’s more useful and valuable (like fuel and plastics).

Bits means I understand the stuff we do in technology and Silicon Valley. We move electrons around to process information. Our factories are data centers. The raw material flowing in is data, the process in the middle is software algorithms and large language models, and the valuable deliverable is information and knowledge.??

I was reminded of the similarities and differences of atoms and bits on a recent visit to Houston. In Silicon Valley, Google, Apple, Meta, Nvidia, hundreds of startups, and dozens of major venture firms are all within a few minutes of each other. Oracle, SalesForce, and many other technology companies are just up the road. It’s a tech playground. In Houston, when you walk around downtown every building is a big energy company: Shell, Conoco, Philips, Kinder Morgan, and so on. I ran into a friend of mine from college who’s an executive in one of those companies walking from one meeting to another, just like I might bump into a former Google colleague at a coffee shop in San Francisco.??

That’s a big similarity between atoms and bits: the impressive concentration of all the major players in a global industry in a relatively small space. The differences, though, are more profound, and, in my experience, much harder for someone from the bits world to comprehend. It’s so easy for bits people to get caught up in our techie ecosystem and change-the-world ethos and underestimate the utter complexity of the real physical world. This is why I think so many tech companies fall short when it comes to creating high impact, quick-to-value solutions for the physical world. They get the bits, but don’t appreciate the challenges of atoms.??

What are those challenges? In our experiences working with refineries, they are scale and safety. It’s critical to understand these things first hand if you are trying to use bits to solve atom problems.

The bits world has plenty of experience building big production facilities: data centers are big and getting bigger. The biggest data center in the world right now is the China Telecom-Inner Mongolia Information Park, located in Hohhot, Inner Mongolia, China. Its physical size is a little shy of a million square meters. Impressive, right?

The biggest refinery in the world is the Jamnagar Refinery Complex in India. It’s about 30 million square meters. 30 million to 1 million. Now this isn’t exactly an apples-to-apples comparison - a refinery is a very different animal than a data center - but it’s clear that the physical scale of refining atoms is much much bigger than that of processing bits.??

Greater scale translates into greater complexity. In my prior role, I worked with the complexity of Google every day. For example, we’d make code changes to improve Android’s performance so your phone wouldn’t crash and the battery lasted longer. In the data center, code optimizations led to huge cost and energy savings. Optimization opportunities always existed (and still do), you just needed to know where to look.?

However, moving a physical substance such as crude oil from where it is extracted from the ground to where it is refined into fuel to where that fuel is pumped into cars, trucks, and planes takes complexity to the next level. It is an incredibly intricate, multi-faceted process, one that’s much more complex than converting data to information.??

Here’s one way to think about that complexity: how many ways can the whole thing be brought to a screeching halt? In a data center, there are few blocking processes that bring an entire data center down. There are a range of problems that can occur, but there are also robust failover systems and processes so you can get workloads back up and running quickly. Bits are weightless and can be quickly re-routed, which makes it easier to avoid system-wide outages.

With atoms, though, it’s much harder to alter the flow of oil and fuel. There are complex thermodynamics at play that you don’t get with bits, so a failure somewhere in the process is more likely to cause a major outage, a safety issue, or a pollution event. There’s redundancy built in many places, but even small or medium scale failovers can have ripple effects all the way upstream to the oil well and downstream to the pump.

For example, refineries often experience incidents that result in materials being accidentally released from their primary containment vessel. Such “loss of primary containment” could be due to equipment or operational failure. Either way, the ripple effects can be severe, reaching well beyond just the immediate issue. Then, after such a failure, getting to the bottom of what happened and generating a root cause report can take a team of people several weeks.??

The scale and complexity leads to a more fundamental concern: safety. It’s hard for a technology-minded Silicon Valley person to grasp how deeply safety is built into the culture at energy companies. Early in my career I was working in the offices at an energy company and we would regularly have safety stand downs, where we would stop what we were doing and talk about potential safety issues.? People took them very seriously. At first, that struck me as a bit unnecessary - we were office workers! But then one of my colleagues was in a plant testing some valves with an engineer. Something went wrong and the engineer was burnt. I realized that we did safety stand downs in the office because our colleagues working in the plants and rigs had to do them to avoid people getting hurt.

This doesn’t happen with bits. When machines fail in data centers it's very rare for people to get hurt or die. When there are failures with atoms in refineries, the consequences can be more dire.

We imbue everything we do at TC Labs with an understanding of the very powerful real-world differences between bits and atoms. We strive to leverage AI technology to help operators spot problems earlier in their very complex plants so they can address them before they lead to failures. When problems happen, we strive to help teams get to solutions much faster by tapping into a broader knowledge base and virtually testing a broader set of possible solutions. We strive for a future where the best operator and the best engineer is at every plant, every day.? All of this is in service to safer operations, fewer failures, faster problem solving, and faster post-issue analysis and reporting.

When it comes to atoms and bits, we understand both. That’s the only way to deliver high value, fast-to-value solutions.?