Conversations from the Ground: Dr. Thomas Danielson
Dispatch 005: Artificial Intelligence Operational Usage
Date: May-Present
Location: Washington, D.C.
Associated with: The Gigawatt Project
Type: Field Transcript
Status: Published
Savannah River National Lab | Special Competitive Studies Project AI Expo | Washington D.C. May 2026
Recorded: 05/07/2026 10:56 AM
Part Three
In May, I had the honor to sit down with Savannah River National Lab scientist Tom Danielson. Finding myself in Washington, D.C. for the Special Competitive Studies Project AI Expo, I sought to learn how Artificial Intelligence was being implemented on the national level. With the talk of AI in the mainstream regarding consumer facing AI and LLMs being the majority of conversation, I was inclined to discover more about what was happening at the larger scale.
In dispatch 005 and the third dispatch of The Gigawatt Project, I present here a partial transcript from my radio broadcast and recorded interview from the floor and will release the full conversation over a series of dispatches.
Source: Personal recorded interview with Tom in May 2026. Which was edited for special on-air broadcast use for Uncontaminated Sound WGXC 90.7 FM June 2026. Machine transcript (Adobe Premiere Pro) and Otter.ai, lightly cleaned of filler; content unedited. Additionally, a review of the original audio has been conducted. The broadcast version is linked at the bottom of this page. Parts 3 and the following flow beyond the broadcast material taken from the original on the floor recording.
Special Broadcast Listen to the full broadcast Wave Farm ArchiveRob Lundberg: And now what do you envision the program going into next year, even, like with the rapid advance of, technology and AI and, what do you see growing if, funding is readily available, what do you see it going in the next two, three, four years?
And computing, I guess, and and advancing your work.
Tom Danielson: So I think some of the things that we're really excited about are sort of the enterprise data management aspects. Right. Enabling people across the complex to access data much faster,that so, that they can then take that for their scientific or engineering purpose. Right. So in the near term that I think is where a lot of the value is.
At the same time, though, in parallel, we have to continue to explore the technology landscape, for the applications that allow us to do a lot of the science. Right? These deep, chemistry models, any number of different aspects. Right. For it, for any domain, whether it be, you know, meteorological forecasting or subsurface for critical minerals or for the grid.
Power generation. Right. I think that those are kind of the two key aspects of how things need to progress near term, making data accessible so that people can do those developments. And in parallel, using the data sets that we already have to develop out those models, mature them, and then also create these valuation pipelines like we've discussed to get them deployed.
That's the long term view, right? Is having AI systems deployed that make things operate more efficiently.
Rob: Now currently, do you have not just AI systems, say in like a controlled setting, but are there physical bots out there as well?
If you may, if I may ask. Monitoring, the water or, you know, contaminants or whatever.Tom: So, so, so you're talking robotics?
Rob: Yes. So, I was just seeing if the AI systems get out into the field in your realm as well.
Tom: Absolutely. So we collaborate. So another area that DOE has, DOE EM in particular is a process of deactivation and decommissioning. So this is essentially when they want to dispose of old legacy buildings.
So I believe there are about 5000 buildings remaining that have to be dealt with in many cases, those buildings, have some sort of challenge present in them. Right? Some contaminants or lots of rubble or. they don't know where parts are. Right. And so, in many cases some expense comes in because, if there are hazards present in the building, right, you have to send people in and they have to dress out.
They have to do all these types of things. Right. So that's where robotics come in. And really being handy is being able to go into these hazardous environments. And if you are able to put sensors on them, the appropriate types of sensors, then you can get really high fidelity characterization of these buildings You can then, you know, tear them down.
Rob: And not put humans at risk for contamination, per se.
Tom: Exactly.
Another area is in terms of, like the use of drones. Right. So drones, of course, give us the ability to go out and collect things like lidar data. Get a really good characterization of the environment.
And they can do this relatively autonomously, fully autonomous. So if we were to allow them but essentially programing in a flight path, you can characterize hundreds of acres with high resolution centimeter resolution lidar data. And you can start to understand really interesting aspects of how groundwater is moving. So doing things like characterizing where our seat lines, where's the groundwater ultimately discharging out to the surface.
And how does that change over time. Right. Whereas if we were to do these walkover surveys we could take a human, quite a long time. Hundreds of acres and get the and get the same, resolution of data. Yeah. Now in this area. So lots of opportunities.
Rob: Yeah. No, it sounds amazing regarding efficiency and manpower and not just efficiency, but also, practicality of saving, like, not putting, persons in the field at risk.
And then also. Yeah, just better quantity and then data collection.
Sounds really cool. And also just super beneficial for the work you guys are doing right now.
Just curious, like, how do you balance, like, the need for energy consumption regarding AI?
And then this is just a generalized question, I guess. And, and then also preserving, you know, limited natural resources.
So. Yeah. So how do you expand? Well, utilizing resources that may not be there, who knows when or, actually… How do you, I guess, expand technology and energy consuming technology? Well, trying to balance equilibrium, of our resources here?
Tom: Yeah, I think that's a great question. I don't know that I have, fact based or totally.
So my opinion is, are there a few different aspects to it. Right. So first of all, I think that AI can enable us to perform research in a faster manner. So this can allow us to get to, some of the energy production, technologies like fusion in a faster manner.
Right. To, meet the sort of this energy demands in certain places. So there's one aspect there, right? And this is the purpose and the Genesis mission as well.
This is the world's best, scientific discovery platform, right? Yeah. Reaching these, these breakthroughs in some of these technologies that have been worked on for a long time.
That's what Genesis needs to enable.
At the same time, I think we have to recognize that we're sort of not necessarily in the infancy of AI. It's been around for decades at this point, but in its current state, right it's relatively new. What we've started to see. Well, there's a lot of power here.
For generalized use cases. Yeah. Pretty amazing where it's at at this point. It's also very energy intensive. So as we're going along and as we're doing research, we're still researching what is what is that final. And why. Look like we all know the answer to that quite yet. But but I would say it makes a lot of sense to think about potentially less energy intensive, approaches.
Yeah, that could be, that could be taken. Right. How do you get similar or the same capacity, from a smaller model, for these applications. Right. So right sizing that model. But also the research and the breakthroughs in just the AI approaches themselves.
