Keynote Session: Under Secretary for Research & Engineering HON. Emil Michael

Well, the vibe of it all is sort of dystopian, I guess, in so far as you've got a category of employees, and I worked in the Silicon Valley for a long time, and I know sort of how the place operates in a large way where you motivate a lot of young people to work really hard on a mission-driven path and there's some goodness in motivating them with a sort of a north star like we're going to build this, in this case they're trying to build super intelligence. And I think there's ways to get caught up in what that could mean and you can get into this doomer sort of lane. That doesn't mean that these AI companies don't have things to consider. I mean when I was growing my business at Uber, you know we're always pushing the envelope and we had to think about safety and regulations and how we're vetting drivers and we always got through it because we had a product that people loved and that was adding value to a city and I think these companies all have good products that can add value to society. But it's easy to get caught up in this sort of doom loop of, you know, we're going to unemploy 40% of Americans, we're going to destroy humanity, we have a cyber nuclear weapon, a bio nuclear weapon, and not expect that there's going to be some adjustment in society that pushes back on some of these externalities, let's call them, or force them to internalize their externalities, which has the effect of them putting more resources in the places that can moderate the potential harms in these things. So, I do think the free market tends to work on these things. Not that there shouldn't be some sort of industry-wide collaboration on this or some form of work with the government. That makes sense. But, you know, 125 million people saw that tweet, something like that. It was quite interesting. It just goes to fear of data centers, fear of change, you know, all the fears that have been coalesced in one well-written tweet.

I mean, when you say that I think the quote was 10% chance that all of humanity is wiped out this decade. So, we're late 26. So that gives us three years. In three years 10% of 8 billion people. So your expected value is 800 million people perishing. You know it seems that something would have to go so catastrophically wrong and uncontrollable by governance systems by sort of the competitive dynamics by the desire of the people who work at these places to be creating something that has that potential. That sounds pretty far out there for me.

You know, I had this when we had our dispute with Anthropic at the Department of War, our dispute was fairly well defined. We wanted to use AI for general purposes. They're building artificial general intelligence. We want to use it lawfully. We committed to use it lawfully. You have to use it lawfully. And then they have their additions to the law that they wanted to add. Right? When I buy a jet from Boeing, they don't tell the military how to fly it or where to fly it or what country it could go into. So, they sort of had this ideological view on it. And I think the balance is that they have to figure out what's going to make, you know, it's a capitalist economy. They get to build what they want. They're liable for the damage they cause. Their employees have their own decisions to make about whether they're going to want to work at a place that's going to be doing something like that. State regulators, local, they're going to be looking at the impact of these companies. But that system's tended to work for 250 years or so. So I'm betting that it will continue to work.

Well, there's a more important court ruling that has yet to drop. So Anthropic was designated as a supply chain risk under two different laws. One of them was in the Northern District of California, a judge where they forum shopped that place. The more important one is the second case only is heard by the DC Circuit Court which is a higher court, goes right between there and the Supreme Court and I think there is sort of more of a national security oriented forum so we should hear from them soon in the next month or two. So there's still a supply chain risk because that court didn't designate a preliminary injunction so the supply chain risk designation is still in place so maybe that wasn't widely known if you're not a legal eagle. And I'm glad that they fixed their issues with the rest of the government, but Department of War has been very consistent about our need to be working with AI providers in a way that benefits the war fighter and defense of the nation. And that means we need to use software if we're going to use it for all lawful purposes in an unconstrained way.

Yeah. Do you tweet or does your staff tweet?

That's you behind the keyboard.

Yes. So, our research and engineering, my component was split off from acquisition and sustainment, you know, in 2018. And the idea was that we had these cold war weapons systems and we had to acquire more of them because we were still growing our stockpile and we had to sustain them because they were complicated machines, lots of parts, but we didn't want that to divert from the new things we had to build from a tech standpoint. And so Congress split them into two. So I have the R, essentially the R&D portfolio or research and engineering. So we can be investing in new technologies, drones, autonomous AI, design for manufactured weapons and weapon systems. And that split has been sort of I think really beneficial to the department because now we're, that was before the Ukraine Russia war started because before sort of what we've seen in Iran with the drone warfare and now we have sort of clear paths where we can develop things at DARPA with the research labs, universities with the Defense Innovation Unit and then as they mature they get brought into the main line of the services. And we all know that we're in a technological age like we've never seen before with the advent of AI and autonomy software. So I'm excited about that.

So the six and then why we took them down to 14, asked a previous question was, you know, how many companies have 14 corporate values? How many times you go in that lunchroom and you say 14 different things I have to remember to walk out with every day to do my job well so I know what's on mission that I know I'm contributing to the goals of the company. And they had things that are pretty esoteric. One of them my favorite was an integrated network systems of systems.

It sounds very DoD. You say you know how do I wake up every day building the integrated network systems of systems? And we tried to make them more digestible because they're not only guideposts for the employees at the department. They're demand signals to industry. They're statements to Congress saying, 'These are the things we recommend you invest in.' So instead, we change them to be more actionable and more digestible and more retainable in people's brains. So the most important one is applied AI, turns out, and that was done before all this brewhaha this year. The second one is scaled hypersonics. We built hypersonic weapons. Now we have to scale them. We have to do them at less than $50 million a copy. Scale directed energy. So which also ended up being perfect for the drone fight we face. You could take a laser out for cents on the dollar relative to a $4 million Patriot. So how do you scale that technology to make it capable? So those are sort of the three I look at every day. In addition to biomanufacturing, in addition to contested logistics, what do you, how do you fight in a contested location? So we're prepared for that in the Pacific.

You know I think it's pretty clear that China's had the biggest military buildup in world history over the last 10-15 years and they started, they had two advantages. One, they started after the cold war. So their sort of design principles were post the beginning of the early technology era starting in the 2000s. So their building sort of had modern principles to it whether it's software architecture whether it's manufacturing practice and so on. Second they took a lot of our stuff. So you're starting from a base of tens of billions of dollars if not hundreds of billions of dollars of American investment in core technologies and if you're willing to take those, steal those and based building that you could skip a lot of R&D right so it worries me that I used to say when I did business in China for Uber that after they stop copying they start inventing that they may be at a start point where they're starting to invent their own things. So that's the, we call it the pacing threat. How do we, who do we compare ourselves to more than any other country in terms of our military might? That would be them and I think for us, you know we still have the most dynamic entrepreneurial environment, the most wide array of technologies, the most in system that's most incentivized to produce things that work in a competitive setting. So I think all those benefits will still be ours in the 21st century. But I think that we've got to make sure that we stay ahead in autonomous and drones, ship building, in the ability to make sure our defense industrial base is resilient against the 20 years of outsourcing that we've done. So these we hear a lot about critical minerals, batteries, magnets. We decided at one point that we were going to have the end of history and we were going to outsource these things and have the lowest cost bid win without recognizing that there are some things that if they fall completely outside of your control into adversarial hands. You can't, you're overreliant on them to the extent that you can't protect yourself in any way you want to. So we've done a massive onshoring effort with the defense industrial base with these choke points and critical areas and we used the Office of Strategic Capital for that among other areas and I think we're doing an excellent job just a year and a half into the administration doing that.

Yeah.

It is definitely the summer of lasers. What we've done on the border with the drug cartel drones. I think there was a report that we reduced the drone incursions into the United States and Mexico by 70%. Because of the lasers we have on the border.

And so that creates a chilling effect. We've also transitioned the Joint Light Tactical Vehicle weapon system which was in R&D to the Army. We're also using these in Iran. So it is becoming the summer of lasers and we're making a lot of progress on it. The last administration didn't invest much into it. We're investing into the core technology, but also investing into the producibility of these systems so that you could proliferate them because they're all about how many can you get out there? How far? So they could take down any threat.

No, in the Middle East. Yeah.

Yeah. In the fight.

More cheaper, faster, right? So the thing with these things is the stronger the beam is, right? That sounds simple to a lay person. The bigger munition it can take down and the farther it could go, but then the bigger it is. So, can you sort of make the size of the thing smaller so you can move it around, make it more mobile, but still have a stronger and stronger beam that get farther and farther out and bigger and bigger munitions. Now, that's not science, that's engineering in some ways. So, how do we engineer to perfect these systems so they're better for every scenario? And I think by the end of this term, we're going to have a wide array of capabilities that are going to be workable in almost any theater that we have. In Golden Dome for America, where I went with Secretary Hegseth, we shot down a cruise missile with a demo.

Which is why we have the Office of Strategic Capital to invest in companies to not only reshore in the US but increase the amount of stuff they could produce. And so that combined with the engineering effort should get us there.