Big Ideas Lab - Genesis Mission

Episode Date: May 19, 2026

The Genesis Mission is a national effort to accelerate discovery by uniting AI, supercomputing, experiments, and the infrastructure of the national laboratories. This episode explores the problem Genesis is trying to solve, the growing gap between the pace of scientific discovery and the scale of the challenges facing the nation. At Lawrence Livermore National Laboratory, this vision is turning into a reality through infrastructure, workforce training, and governance.  Guests featured (in order of appearance): Dario Gil - Department of Energy Undersecretary for Science and the Director of the Genesis Mission Rob Neely - Associate Director for Weapon Simulation and Computing, LLNL Lori Diachin -  Principal Deputy for Computing, LLNL -- Big Ideas Lab is a Mission.org original series. Executive Produced by Levi Hanusch. Sound Design, Music Edit and Mix by Matthew Powell. Script by Caroline Kidd Story Editing by Levi Hanusch. Audio Engineering and Editing by Matthew Powell. Narrated by Matthew Powell. Video Production by Levi Hanusch. Brought to you in partnership with Lawrence Livermore National Laboratory. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

In the late 1670s, voices rose in a room full of arguments in Paris. At the edge of it all sat Gottfried Wilhelm Leibniz. To him, the arguments felt predictable. Not wrong, just imprecise. Because days earlier, he'd been at a workshop. Gears turning. And a machine that didn't argue. Just rules producing answers.

The contrast stayed with him. The problem wasn't the ideas. It was the language. What if reasoning worked like the machine? Not argued, but broken into symbols. He wrote his thoughts into words, quote, When controversies arise, it will suffice to say, let us calculate. He already envisioned converting all knowledge into a binary,

presentation and imagine calculators, machines that would take that binary representation and calculate with it. And he envisioned that as a possibility of then problem solving. A vision from the 17th century. So that was already hundreds or years ago. Problem is, they didn't have the technology to make it happen. Even now, science can model parts of reality. Computers can process data, but no system has been.

able to do both, at the scale that scientific discovery now demands. I think that lightening's dream of making all knowledge computable is one that is going to be realized for the first time now. High performance computing, artificial intelligence, and emerging technologies like quantum computing are beginning to work together, expanding what's even possible to model. The timing is right. The moment is now. That convergence is no longer theoretical.

It's being built. Across all 17 Department of Energy National Laboratories, alongside universities and industry, a new engine for discovery is taking shape. The Genesis Mission. Welcome to the Big Ideas Lab, your exploration inside Lawrence Livermore National Laboratory. Hear untold stories, meet boundary-pushing pioneers, and get unparalleled. paralleled access inside the gates. From national security challenges to computing revolutions, discover the innovations that are shaping tomorrow today. Scientific advancement still moves

the way it always has. A question, a hypothesis, an experiment, a result. Then the cycle begins again. That process has carried humanity for centuries. It has revealed the laws of motion. Unlock the atom, map the genome, and push the frontier of knowledge further and further forward. What has changed is the scale of the world science is being asked to understand. The problems facing scientists today are larger systems, denser data, tighter timelines, and higher stakes. The challenge is no longer just discovering something new. whether that discovery can happen fast enough. President Trump wants to streamline the regulation process for artificial intelligence and

the Trump administration is pushing to expand AI and federal scientific research. An historic mission. This is reminiscent of the Manhattan Project. The Genesis mission is a coordinated national effort to accelerate discovery by bringing advanced artificial intelligence, high-performance computing, experimental science, and the secure infrastructure of the national laboratories into a single, engine for science. This is an all-in national effort to take the power of AI and pair it with the 40,000 outstanding

scientists and engineers at our national labs to advance innovation and science. It is also a way to help the nation move faster from scientific insight to real-world capability. Dario Gill is the Department of Energy Undersecretary for Science and the director of the Genesis mission. We seek to double the productivity and impact of America's R&D engine. And the goal is to do that by doing two things, by building a platform that helps us accelerate discovery and by constructing and executing a portfolio of national challenges of great importance across a variety of domains, from energy to discovery science to national security. At its core, Genesis is about helping the United States.

move faster on the problems that matter most by bringing AI, supercomputing, and experimentation into the same scientific workflow. A closed-loop system where each result helps refine the next question. Challenges like making fusion energy reliable at scale. Discovering new materials that could reshape entire industries. Strengthening national security and American competitiveness through faster, more trusted scientific and building the next generation of computing itself. The kinds of problems that don't just advance science,

they reshape how the world works. Because in areas tied to energy security, advanced manufacturing, and national defense, speed becomes a strategic advantage. President Trump signed an executive order launching the Genesis mission. But this is very all-inclusive. We're signing tremendous... When the executive order was launched, order was launched, the president entrusted the Department of Energy to lead the Genesis

mission on behalf of the country, and the president and the secretary appointed me as a day-to-day leader of Genesis. But we recognize from the very beginning that we cannot do this alone. We're executing a strategy that will involve extensive partnerships with industry, partnerships with universities and philanthropy, and partnerships with our sister federal agencies. The ambitious goal of Genesis requires equally immense scale. This is a project that all 17 natural laboratories are executing together, all of them, all 17, which is quite unprecedented that this is happening. And that is because there's not a single domain in their missions and in the practice of science and engineering that is not going to be impacted by what is unfolding with AI and the broader

computing revolution story. Artificial intelligence is, one piece of the system. But on its own, it isn't enough. There's a disconnect between where AI is advancing and where science actually happens. Private sector has an amazing amount of AI compute capacity in the United States, but it's not true for the labs in the universities. The AI has done a really good job on things like language and coding, but it is still in the infancy of having a transformative impact in physics and materials and chemistry and mathematics and engineering. One limitation that we are addressing with the Genesis mission is we want to push AI towards the frontier of science and engineering so that we can use it as an instrument, as a tool,

to help us do better science and engineering. What makes Genesis different isn't just the systems it's built with. It's what it's built for. We are science first and not AI first. Genesis isn't another AI platform. It's an entire system designed for national acceleration and scientific advancement, including environments where trust, security, and human oversight are non-negotiable. The mission around that is to evolve or ability to ask better questions,

apply the scientific method, everything from hypothesis generation to the design of experimental campaigns, the collection of that data, how we share and distribute, how we integrate knowledge across disciplines. So because we're science first, we are focused first on the domains, first on the communities that have been the ones that have been accumulating the knowledge and the experimental capability and the scientific instruments and the wherewithal to push those frontiers forward.

The proof is already here. Not in theory, but in scientific groundwork, is trying to scale. In biology, that foundation took shape in the protein data bank, a growing archive of experimentally determined 3D molecular structures that began at Brookhaven National Laboratory and was built out over generations. It expanded to a broad set of collaboration

where generations of scientists and microscopes and experimentalists systematically collected samples, used crystallographic techniques to be able to, create three-dimensional representations of the structures. Fifty years of effort of scientists and engineers and user facilities created 200,000 protein structures, and then also the community created its own benchmarks of how you validate it. And even that wasn't enough.

For decades, scientists had to measure each protein directly, one experiment at a time. Then something changed. Instead of measuring them, We started predicting them with AI. In that case, the methodological changes is I give you a sequence of amino acids and you're able to issue the protein prediction

now using a neural network page technique. The amazing result of that is we went from 200,000 proteins to 200 million in the period of a couple of years. This kind of acceleration isn't isolated. It's beginning to appear in pockets across the world of research. For the National Nuclear Security Administration lab, within the Department of Energy, that possibility carries a specific weight. It means applying AI and advanced supercomputing to some of the nation's most demanding security

and deterrence missions, where speed matters, but trust matters just as much. It's a new way of doing science. Rob Neely is the Associate Director for Weapon Simulation and Computing at Lawrence Livermore National Laboratory. It's not going to necessarily replace the way that we've done science. But it's a new way of thinking about the processes and how to come up with new innovations. And all that's now culminated in the vision that is Genesis mission. At Lawrence Livermore, that shift isn't just about faster science.

It's about enabling mission outcomes in environments where speed, accuracy, and trust are critical. When the COVID-19 pandemic hit, the need for faster scientific workflows for national security concerns like pandemic preparedness became. immediate. There was a lot of effort going on in the science community to understand how to come up with rapid responses to help stop this pandemic. And one approach was developing antibodies. And that was all work that can be done largely computationally. At the end of the day, the result was they went and actually did some experiments in the laboratory using synthesized molecules to understand. Did what AI pointed us to and what simulation told us was possible. work. And that was an eye-opener to me because it's like, wow, that is something that would have

taken years and years of manual effort to do. And now we're able to do this with the help of AI assistance in a matter of weeks and months. And thinking about that in the context of other science problems, the possibilities are almost limitless. What emerged in that moment was more than a rapid response. It was a glimpse of a fundamentally different way of doing science, where AI, simulation, experimentation, and validation continuously inform one another. So then the question became, what happens when you apply that same approach

to the biggest scientific challenges we face? We've put a list out of national challenges from fusion, materials, to a whole host of areas that are really important to us. Those challenges stretch across energy, materials, advanced manufacturing, simulation, and national security. Fields where faster discovery can change what the country is able to build, protect, and deploy.

Are we able to realize a world where fusion is part of our ecosystem? That would be an amazing thing. And if people see that, hey, look, you guys did a good job on that. And we trust because you deliver the outcomes that you wanted. That is the promise of Genesis. Not one breakthrough. Not one field. Not one model.

a unified mission. It's starting to take all these individual small pieces and get them more unified in looking at things in a common direction. For centuries, the idea was ahead of the tools. Now, for the first time, the tools are beginning to catch up. Artificial intelligence, high-performance computing, experimental science, the infrastructure of the national labs.

A lot of the pieces are there. It's just taking this big vision of Genesis mission to finally get us to bring them all together. Across labs, universities, industry, government agencies, and all the data connecting them, that's where the real challenge begins. There's tremendous opportunity and there is some tremendous risk. Imagine a scientist has a question. The data that can help answer it lives in one system. The model that could search it lives somewhere else.

The computing power to test it lives somewhere else again. Genesis is the attempt to bring those pieces into one working system. 2004, Lori, didn't really even have AI on her radar. Meet Lori Dyschen, the principal deputy for computing at Lawrence Livermore National Laboratory. She works at the cutting edge of integrating AI into the national laboratory system. Computing was very different than compared to what it is now. She's not alone. In the span of just a few years, AI went from something most people barely used to something shaping how we write, code, search, plan, and solve.

But inside Genesis and Lawrence Livermore, the shift carries a different weight. Accuracy, security, and trust are no longer aspirations. They're essential. You can't necessarily trust what's coming out of an AI model because it doesn't always have that context that you have in your room. mind, right? It has a probabilistic understanding of what's likely next. And it's often right, but not always. They're probabilistic models that predict what is the most likely thing that's next in a sequence, right? And that sequence could be code, it could be language. And when you understand that, you're able to better assess what the risks are associated with it, where there may be biases,

where there may be errors. And I think ultimately the buck has to stop with the human. At the same time, new models of artificial intelligence are being released. That means that scientists have to balance the speed of innovation with the importance of trust, reliability, and security. It's very, very frequent that new frontier leading models are coming out from the various different companies. And those different models have different strengths and weaknesses. We do need to make sure that all of our new uses of AI

align with our policies, with our ethics, with the fact that we need to be very, very strong. sensitive to the risks that it introduces. How can we streamline it in various different scenarios while still maintaining confidence in what we're approving? Genesis lives inside that tension. Building it within the national lab system grounds this initiative in the scientific rigor needed to manage risk while still accelerating discovery. Big science is one of the things that the labs do really well, and we mean by that is that they are uniquely positioned to create some of the most sophisticated scientific instruments that the nation creates and the world creates.

So think about all the light sources and particle accelerators or telescopes that we do in partnership with our sister agencies like NSF and others. The national labs matter not just because of their equipment, but because they combine mission, infrastructure, people, and rigor. One of the unique aspects of the labs is you can collect teams and put teams together that do, big scientific projects. And then, not only we build them, but we operate them. So we operate a large number of user facilities that support 40,000 scientists and engineers across the United States. So that's one thing that is unique.

It is at a scale that a university could not build and operate or even the private sector. And so we fill that gap. Another thing that the national labs do is they have a unique and distinct mission on pushing the frontiers of energy, of the physical sciences, and of national security. And as a consequence of that, there are incredibly talented group of people that have the infrastructure,

the domain expertise, and the mission to carry out things that are very special. At the labs, those tools have to operate alongside some of the most sophisticated systems in the country. And that is where operational complexity begins. Some of that work happens in sensitive national security workflows where AI systems must operate with rigorous safeguards, traceability, and human supervision. A lot of the Genesis mission infrastructure will be built on tools that were developed

at Lawrence Livermore in our Livermore computing program. A vision this large is tested in smaller places, and the daily habits of people ask to carry it, person by person. We have an incredibly capable workforce at Lawrence Livermore. A great set of folks who have proven over and over again that they can adapt to new technologies that are coming out, new ways of doing business, but they need to be trained. There are 9,500 people at this laboratory, and there's a lot of work to do to make sure that they're all literate about AI and what's needed for their jobs and how to make them more efficient, more effective, et cetera. So that's one area where we really need to make some investments to make it durable. It's a big job. So there are a number of initiatives that we have undertaken

at the institution to help upskill the entire workforce. Not a few specialists at the edge of the lab, but thousands of people adapting to a fast-moving new technology and making it part of real work. For Genesis to work, data cannot simply exist. It has to be findable. It has to be accessible to the right people. And it has to stay protected from the wrong ones. Those requirements became critical as Lawrence Livermore works with production plants and sites to build a digital thread, connecting data, models, and decisions across the mission.

We have recently at Lawrence Livermore stood up a data governance board that's thinking about how do we make our data findable and accessible. retrievable, and how do we do that in a way that doesn't disrupt the fact that we've got hundreds of people who have stored data in various different places. So we're creating a metadata catalog, and we're talking about the interfaces that you need to allow people to be able to find it and access it. And this is going to be critical across all the plants and sites within NNSA and across DOE, and Genesis Mission is really accelerating that development as well. In Genesis, data governance is a key part of building a trusted national security AI infrastructure.

That tension between collaboration and security is not new. And for Rob, it echoes an earlier effort where openness at a national scale had to be built without compromising the environments where their work ultimately had to live. The X-scale computing initiative ran from about 2016 through 2022 or so and was an all of government effort led by the Department of Energy to understand, understand how can we advance our modeling and simulation. That effort required the same balancing act now facing Genesis, broad collaboration, national scale, and work that ultimately had to function inside environments where access could not

be taken for granted. We had to have an unclassified analogy to that so that we could partner with people who didn't have the clearances. But at the end of the day, this had to work in the classified environment. We solve a lot of the infrastructure problems in the unclassified network. We come up with similar problems to the kinds of things we're solving that are classified that we can talk about openly. After all of the models, the infrastructure, the workforce, the data, and the secure environments,

Genesis still comes back to one question. Can it be trusted? Do we trust the people, the groups and the institutions that are carrying them out? We're talking about scientists and engineers. putting AI to the uses that they think are most important, and also that they are people who have been trained and they have a rigor in how they approach it. If Genesis succeeds, it will leave behind something lasting,

an AI platform that can scale, a system that can endure, and a stronger engine for scientific discovery across the country. That includes the national security missions that depend on getting the science right, and getting it fast. It could reshape how the U.S. designs, manufactures, and deploys critical technologies

in an increasingly competitive world. Impact that will not be abstract. Did we make Americans' lives better when we finish all of this? Because we were able to deliver more affordable and reliable and secure energy sources to them. I would say Genesis is going to be a great accelerant of the United States ability to do big science faster.

It will be felt in science, in energy, in national security, and in outcomes people can actually see. If Genesis succeeds, it will not just accelerate discovery. It will help define how the nation solves what comes next. Thank you for tuning in to Big Ideas Lab. If you loved what you heard, please let us know by leaving a rating and review. And if you haven't already, don't forget to hit the follow or subscribe button in your podcast app to keep up with our latest episode. Thanks for listening.