“Scientific progress on a broad front results from the free play of free intellects, working on subjects of their own choice, in the manner dictated by their curiosity for exploration of the unknown.” – Vannevar Bush

We need to talk about how misaligned incentives have eroded the quality of the Labs. I will discuss the important role these labs have for the nation, its short- and long-term security, and how to fix the problems. At this moment of extreme change and danger, the lack of high-performing federal Labs is dangerous. They exemplify the decline in American science.

Before I get to the bigger, deeper issues, I’d like to recount an episode that shows how far the moral and ethical behavior that’s accepted at the labs has fallen. This was one of the most abhorrent episodes while I was at Sandia. This gets to some deep issues with the mindset of people in parts of the lab.

I had a student who was quite good. He has his PhD now and is a staff member. He was working on his degree at Texas A&M University. His advisor at the time was an Iranian citizen who had lived in the United States for quite some time. He is a citizen now.

“Ignorance, allied with power, is the most ferocious enemy justice can have.” – James Baldwin

Ar required, the counterintelligence part of the lab demands that we elaborate on any external collaborations we have with foreign nationals. I complied. When I looked within the database for the information to make sure it was correct, I noticed that the line item associated with this professor was labeled as “terrorist.” In other words, the professor was labeled as a terrorist. I found this alarming and investigated. It turned out to simply be a general label for Iranians, not a statement about the person.

This only made the label worse.

I found this to be a genuinely appalling lack of judgment on the part of the counterintelligence people. I called it out, and they declined to address the issue, saying that it was too difficult to fix within their database. This should have been a three-alarm fire for the counterintelligence office. Nothing the government funds should ever use this kind of language or slur to describe anyone. Can you imagine what would happen if they took someone who came from an African country and labeled them with the N-word?

I will just note that the label “terrorist,” when given to anyone of Arabic descent, is a racial slur these days. The lack of any sort of sense of this on the part of people at the national laboratory was one of the grossest and most disgusting things I had seen in my entire career. As far as I know, this practice persists, and they still label people of Iranian descent as terrorists. I don’t know whether this applies to other nationalities, but it is an appalling practice. Full stop.

This repulsive example is just the tip of the iceberg of fucked up things ruining the Labs.

“The prospect of domination of the nation’s scholars by Federal employment, project allocations, and the power of money is ever present and is gravely to be regarded.” – Dwight D. Eisenhower

For all the talk of alignment with AI models, the thing we rarely talk about is the misalignment of incentives for corporations and institutions. Increasingly, corporations, for all their wealth, are viewed as the model for institutions. I saw this at the labs, where corporate governance became the way to do things. In the process, the incentives totally aligned towards money and profit, and power became aligned similarly at the labs. This has led to destruction of capability and incompetence.

“The best minds of my generation are thinking about how to make people click ads. That sucks.” – Jeff Hammerbacher

One of the big problems is the advent of the gold standard science from the Trump administration. They are bootstrapping off of genuine problems in the scientific community to make things worse. They are pushing forward an ideological project that is undermining science. I would not mince words on how much the national labs declined during my tenure. There are real problems, and those problems are huge. Nothing the current administration has done addresses those problems directly or makes things better.

They are only making things far worse.

The Lab’s missions are entirely different from corporations. I would argue the alignment of incentives at corporations is similarly misaligned. Societal good and the well-being of the people in society are not even considered. They often suffer under the decisions that are most beneficial for profit. We’ve seen this unfold decade after decade, most acutely with the nature of social media. Their software is engineered to get humans’ attention because that attention is a source of money. This engineering is done without regard for the damage that it does to society as a whole or the individuals in it. Children are particularly vulnerable to this.

Now we have AI that is being written by these corporations.

Somehow we think that the misalignment of the corporations won’t seep into the alignment of these models. They have the power to do even more damage than the social media software has already done. This is absolute madness!

For a long time, the labs have suffered under the governance of the NNSA and the Department of Energy. The labs have become completely allergic to anything that looks like failure. Without failure, you also sacrifice success. This is particularly true in safety, but it’s poisoned research as well. There is this element of fear and hostility to any risk-taking that has led to a decline in everything the lab does. This is the pre-existing condition the current administration inherited. They have done nothing to address it.

Part of the reason for this is the inability of the agency managing the labs to accept failure in a way that is positive. They make little or no attempt to ever cast the failure as an opportunity to learn. It is an opportunity to punish. Anyone who screws up infects everyone with punishment. Every failure just generates more bureaucracy and oversight, and it creates a vicious, self-perpetuating cycle of decline.

“A ship in harbor is safe, but that is not what ships are built for.” – John A. Shedd

I watched this happen for virtually my entire career. I saw a wonderful, generous environment at Los Alamos savaged and destroyed. There, everything good was destroyed along with the problems being wrapped in red tape. They took a hatchet to the Lab where precision surgery was needed. Sandia already had its issues when I arrived with its corporate takeover in 1994. Those issues have gotten worse and worse over those 20 years. This is twinned with the priority given to money. By the time I left, the only thing that really got managed was money. The science, the engineering, the technology were not managed, except for reduction of risk.

AI is a great example of how this manifests itself. For starters, the labs have lost their supremacy as a source of superior scientific and technical knowledge for the nation. AI is this in spades. Private companies have all the expertise, and the labs are playing catch-up. Furthermore, the labs can’t take risks and do the research necessary to close the gap. In fact, the gap is just widening. The nation needs their expertise, and needs it for national security. They need to weigh in on the big subjects of our day. Today, the Labs are impotent.

I’ll get more of this in terms of how cowardly and weak the lab’s leadership is. They have capitulated to the Orwellian demands of the administration. This includes a new version of political correctness in terms of terms that can be used. First, the terms related to climate change were outlawed. No longer could researchers write their results freely. They were censored. Now they have outlawed “artificial intelligence”. It has been replaced by the President’s edict of “super intelligence”. This edict is pure propaganda and completely un-American.

All this does is undermine the integrity of the lab’s leadership and show how weak and craven they are. All they care about is funding. They have no moral certitude or virtue to lean on. In their capitulation, they’ve shown how unfit they are for leadership. They should have lost everyone’s respect.

It is the same thing that is poisoning the law firms that simply collapsed under the pressure from the administration. The ones who gave in to everything based on their threats, threats that were illegal. It violates freedom of speech, and it basically implements a form of chosen stupidity as policy. This simply accelerates the decline of the Labs.

Another example where safetyism has completely destroyed important national capabilities is launch to orbit and the satellite business. The obsession with safety and fear of failure means that NASA is no longer an organization that can lead in space. All our launch capability has been given over to business interests such as SpaceX or Bezos’s Space Launch Enterprise. This is completely unnecessary. Rather than empower and fix NASA, the government just gave up. The key is that NASA should be allowed to take risks and fail, just in the same manner that SpaceX does. Just as NASA used to do. The only reason it doesn’t is the attitudes that we have in government, which hamstring these agencies.

What I can say is that the NNSA is going the same way NASA went. It is in decline, and the way we have governed these institutions is destroying essential national capabilities. It may already be too late. The labs may already be in such a horrendous shape that they are just as incapable of executing their mission as NASA is capable of executing its mission.

“Failure is only the opportunity more intelligently to begin again.” – Henry Ford

I can start with one of my last experiences at the lab around AI. This was the nature of the chat LLMs the labs could access compared to what was accessible outside the labs. Invariably, when I tested the various chat capabilities at Sandia, the internal chat (which we should have had a free, unlimited license to use) was dramatically inferior to any of the LLMs on the outside in free mode. I was working at a premier National organization that had lobotomized these tools on purpose.

It’s my understanding that this imbalance has persisted. The whole reason this is occurring is fear: the lab’s fear of information getting leaked, their information control protocols being violated, and a host of other concerns. The concerns are very legitimate, but the fear of all this means that the labs are using dramatically inferior products to do seemingly essential research. In the end, the fear is completely undermining what ought to be essential work for the nation. AI is such an enormous transition and opportunity that the fear will have long-lasting, profound consequences.

The opportunities that I have described with AI are largely unavailable to the people in the lab. Thus, the productivity, the ability to explore ideas, and the basic capability are lost to them. Unless something changes, the labs will always be significantly behind the cutting edge. In a year, they may have capabilities that are equivalent to what I have now. By then, what I have now will be a pittance to what is available then.

“Political language… is designed to make lies sound truthful and murder respectable, and to give an appearance of solidity to pure wind.” – George Orwell

How can we fix this utterly, completely fucked-up situation?

The key is to start trusting people again. The labs need to be trusted. The people running the labs need to start looking at failures and problems differently. Change the mindset from thinking that these are not incompetent or malicious, but rather failures: things that can be fixed and learned from. They are not reasons to mete out punishments every time something goes wrong.

The key is to actually fail and learn from those failures and create better, more effective, more efficient systems. Every bad incident, every bad thing that happens at the lab, whether it’s a safety problem such as a fire or a security incident, is viewed as an apocalyptic event. It simply puts more bureaucratic measures onto the lab and increases the fear and the reticence of the lab’s management to do things. The government and public need to react differently. Bureaucracy is never the answer, and less of it is better.

Over the span of my career, this became an unremitting and corrosive influence on the labs. The labs are much less capable and effective places than they were. They are worse engineering and scientific institutions. It is all because of a lack of trust and the punishments and policies imposed by the governance. To be blunt, the current governance of the labs is a failure. The current management of the labs is a failure. This failure is something we have created and continue to amplify. It’s time to change course.

The ideology that they are promoting is unscientific and overturned by the casual examination of facts. Even worse, they have changed the priorities and the studies in research so that various ideological bits are now removed from the federal research agenda. This is another place where they’re making our future less wealthy, less safe, and less secure than it would be. Part of doing science is sometimes taking the evidence and accepting a hypothesis that violates the current mode of thinking. When a mode of thinking is simply imposed, the evidence is disregarded. The administration has taken a bad situation and made it even worse.

Let’s take the example of the COVID pandemic as being instructive. The right-wing administration instituted a bunch of conspiracy theories that now drive their actions. There were some genuine problems with how we responded to COVID. The view of expertise, as exemplified by someone like Dr. Fauci. The advice that Dr. Fauci gave was from one extreme point of view. Where it went wrong was that it was unbalanced in terms of what the suggestions meant for society. Fundamentally, it was toxic safteyism. There were other bodies of expertise that should have weighed in to moderate the suggestions made and come up with a better solution for society as a whole.

Trump proceeded to pour gasoline all over the conflagration. The issue, particularly for democratically run states, was that the medical advice was taken out of balance and fed into the rampant safetyism that defines society today. Rather than allow there to be reasonable risks for populations that were not susceptible to disease, they imposed draconian lockdowns on all of society. This is probably most acutely felt in the impact on children and education. The problem is that there was an inadequate level of risk-taking and a balanced perspective. It wasn’t that the expert advice was wrong or bad. It was that the expert opinion and advice came from one point of view that was not measured against other considerations effectively. This has turned into another attack on science that has moved everything in the wrong direction. It’s one of the reasons we have unprecedented measles outbreaks in the United States and falling vaccination rates.

In the end, both the left and the right wing of politics failed: the left wing with too much safety and unbalanced advice, the right wing with unscientific ideas and conspiracy theories.

Returning to the governance of the labs, the impact of both of these trends is clear. By punishing failure and failing to take risks, we are setting the science back dramatically. The science and technology we should be counting on for a better future are being crippled by both of these trends.

“Facts do not cease to exist because they are ignored.” – Aldous Huxley

This is not a left-wing or right-wing thing. Both sides of the political spectrum have damaged the labs. They’ve just done it in different ways. Although the right-wing assault is more current and visceral, the left-wing assault has gone on for decades. In both cases, they are harming the nation and ultimately leading to less wealth, less health, and less security for the vast majority of the population.

“Moral courage is a rarer commodity than bravery in battle or great intelligence.” – Robert F. Kennedy