Origins of a Bipartisan

Growing up with 69 first cousins in a small Arizona town where his dad was the nonpartisan mayor, Flake explained that he arrived in Washington in 2001 in shock at the hyperpartisanship of Congress. “There are people in Washington on both sides of the aisle trying to do good, but the incentives are terrible,” he lamented.

In 2014, Flake and New Mexico Democrat Martin Heinrich attempted to prove that Republicans and Democrats could work together in perhaps one of the more unique episodes in legislative history. The pair, with the help of the Discovery Channel, filmed a Survivor-style reality TV show on a deserted island in the Marshall Islands, learning to survive with nothing but each other and a machete. 

June 2017 brought home that hyperpartisanship is no game: along with other congressional Republicans, Flake was practicing for the annual Congressional Baseball Game when a gunman opened fire, critically injuring two. Just months later, Flake announced he would not seek reelection to the Senate, saying he was out of step with the direction of the Republican party under Donald Trump.

The Diplomat

Türkiye was no small post. Normally occupied by a career diplomat, the Biden administration entrusted the ambassadorship in Ankara to Flake during what would prove to be a contentious time in Turkish-American relations.

For years, Turkish president Recep Tayyip Erdoğan had steered Türkiye further away from democracy, jailing opposition leaders and vilifying unfriendly media. Flake, who prided himself on speaking his conscience during his political career, suddenly had to walk the line between trying to improve the situation and getting expelled from the country. 

“You never want to be condoning human rights abuses or democratic backsliding,” he told the audience. “That doesn’t mean you lecture at every opportunity,” he added.

To balance out his criticisms of the regime’s actions, Flake noted that he tried to find common ground and give credit where credit is due. For example, he explained, Türkiye is home to more refugees than any other country, and Flake found it important to praise that humanitarianism both in public and directly to Turkish leadership. “In truth, it’s a difficult thing for them to do, politically,” he said. “I tried at every avenue I could to compliment them there.”

The War

Six weeks after he arrived at the U.S. embassy in Ankara, Russia invaded Ukraine, and Flake spent much of the next year trying to build a bridge between NATO and its most Russia-friendly member.

The ambassador spent week one persuading the Turkish government to close the key Bosphorus Strait to Russian military ships, and they eventually did. Flake maintains that Türkiye’s ultimate decision was not just due to international pressure. “They did it because [NATO] was important to them,” Flake said of the Turkish government.

And when Erdoğan came under fire among NATO members for blocking Sweden’s accession into the alliance, Flake positioned himself at the front lines of negotiating with Türkiye to lift its veto.

To do this, Flake tried to understand the veto’s roots in Türkiye’s resentment of Sweden’s position on the PKK, a U.S.-designated terrorist organization which promotes Kurdish nationalism in Türkiye. “[The veto] wasn’t just something Erdoğan was trying to use for leverage,” Flake explained.

When he went to Washington to negotiate the sale of F-16s to Ankara as part of the deal to lift the veto, Flake described finding that the issue cut across party lines, reassured that not all matters were constrained by partisanship.

A Hint of Bipartisanship in the 119th Congress

Having ushered Sweden into NATO (earning a knighthood from the Swedish king along the way), Flake left Türkiye at the end of the Biden administration as the Trump team assumed a more alliance-skeptic international outlook.

When Kathryn Stoner, the Mosbacher Director of the Center on Democracy, Development and the Rule of Law at FSI and professor in the Political Science department moderating the discussion, asked Flake about the prospects of bipartisanship after the November midterms, he responded optimistically. He pointed out that the Lindsey O. Graham Sanctioning Russia and Iran Act of 2026 passed both the House and Senate with large majorities, revealing members’ continued belief in promoting American principles abroad. “That’s where they are, that’s where they want to be,” Flake said. 

“We’ve been through bouts like this before,” he said of President Trump’s change in foreign policy, expressing optimism about a future return to transatlanticism. “If his policies were more popular, his influence would be stronger.”

Flake also described a “longing for U.S. leadership” in Europe, but he added that the U.S. is an unreliable partner. “Our allies have found themselves turned upside down,” Flake exclaimed.

To reassure allies of America’s support, Flake called on his former colleagues in elected office to continue speaking their minds on foreign policy issues. “Heads of state listen,” he said.

Towards the end of the talk, a community member asked what role Secretary of State Marco Rubio could realistically play in shaping U.S. foreign policy. 

Flake’s reply: the example of Mike Pence taught him the necessity of having reasonable, level-headed members in Trump’s inner circle, and Flake is convinced Rubio is working effectively behind the scenes. “I don’t fault those who feel they can influence a president,” Flake said. “I’m glad he’s there.”

In Ukraine, inexpensive drones, software updated overnight, and weapons built in weeks instead of years have reshaped the front lines. In the Middle East, waves of missile and drone fire involving Iran, its regional neighbors, and the United States have tested air defenses and the economics of intercepting cheap threats. Both conflicts demonstrate that the advantage goes to those who can build, adapt, and field new technology fastest. Keeping pace takes more than new weapons. It takes a generation of technologists, founders, and public servants who understand the problem. On October 7, Stanford DEFCON set out to build that understanding, bringing together military officials, technology funders, and private industry leaders into one room on campus.

The day featured keynotes from former Taiwan President Tsai Ing-wen; Gen. Dan Caine, chairman of the joint chiefs of staff; and Owen West, director of the Defense Innovation Unit. Panels addressed the issues shaping modern conflict, from producing unmanned systems at scale to lessons shared by speakers who had traveled from Ukraine to describe building weapons under fire. Together with an advanced technology career fair the day before, which connected more than 400 students with 40 companies, the two days drew about 1,200 people.

The conference was sponsored by the Gordian Knot Center for National Security Innovation at the Freeman Spogli Institute for International Studies, but students ran it. Over several years, center director Joe Felter has built a network that links veterans, engineers, policy decisionmakers, and aspiring founders across Stanford’s schools. Members of that network raised funding from 22 sponsors, recruited speakers from the Pentagon to Taipei, and hosted more than a thousand guests, all while carrying their course loads.

We spoke with three of the event’s co-presidents, all second-year MBA students at the Stanford Graduate School of Business. They talked about how they pulled it off, their favorite moments, and their advice for students who want to work on the nation’s hardest security problems.

Meet the Event Co-Presidents

Ryan Casey, MBA '27, served for eight years as an infantry officer in the U.S. Marine Corps before coming to Stanford. 

Gordon Kim, MBA '27, is founder of Praetar Systems. Before Stanford, he served for ten years in the U.S. Army as an infantry and Special Forces officer.

Luke Zerrer, MBA '27, spent six years in the U.S. Army, most of it in Special Forces, and then worked as an engineer at the Johns Hopkins Applied Physics Laboratory on autonomous systems for air and missile defense.

What was your role in organizing the conference?

Ryan Casey: I was one of DEFCON’s co-presidents, and with the rest of the planning team I led our work recruiting panelists and participating companies. I also moderated two panels: “Effectiveness Under Fire: Command and Control, Communications, Maneuver, and Power for Contested Logistics in the Pacific,” and “Beyond the Valley: How the Government and Private Capital Are Underwriting National Security.”

Gordon Kim: I was the lead planner and one of DEFCON’s co-presidents. I was responsible for driving the conference’s vision and managing the full DEFCON team. I brought President Tsai Ing-wen to Stanford and moderated our fireside chat with her. I also moderated a discussion on Taiwan’s democracy, emerging technology, and the U.S.–Taiwan partnership with Director-General David C.H. Wu, Admiral Lee Hsi-ming, and Jeremy Chih-Cheng Chang.

Having lived and worked in Taiwan, I wanted students to hear directly from leaders who have shaped its future. Taiwan sits at the intersection of democracy, global trade, and critical technology. Understanding its importance is essential to understanding our own national security.

Luke Zerrer: I was one of this year’s co-presidents. I led sponsorship, working with the 22 companies and firms that funded the conference, and I helped with general logistics. On the day itself, I moderated two panels. “Mass, Autonomy, and Attrition” covered building and buying unmanned systems at scale, and “What Ukraine Taught Us” covered building weapons in weeks instead of years.

What were the highlights of the conference for you?

Kim: Moderating the fireside chat with President Tsai was a defining moment for me. We discussed how she helped shape Taiwan’s place in the world despite the constraints it faces. Her observation that sovereignty is a mindset and a value stayed with me. It captured the determination of people who insist on shaping their own future, and the resilience I came to admire while living in Taiwan.

I was also proud of what we accomplished as a student team. DEFCON drew more than 850 registered guests, and the advanced technology career fair the day before brought together more than 400 students and 40 companies.

Zerrer: The Ukraine panel is the one I will remember. Speakers traveled from a country at war to sit in a room at Stanford and tell students, founders, and investors what they have learned about building and fielding weapons under pressure. It is hard to get closer to the problem than that.

The other highlight was watching the whole day come together. We had three keynotes, from former President Tsai Ing-wen of Taiwan, Gen. Dan Caine, and Owen West, plus a full day of panels. Students organized all of it while carrying a full course load.

Casey: For me, the highlight was seeing students from across the Stanford community come out, eager to learn and to be part of serving our nation. Most of these students have no prior military service. They still want to serve their country outside the uniform, in industry or government, and that is truly inspiring.

What has being part of the Gordian Knot community meant to you?

Zerrer: The Gordian Knot community is where my military and engineering background meets what I am learning at Stanford. It puts people from the military, engineering, policy, and investing in the same room, and it trusts students with real responsibility. Joe and the team let a group of students run a conference of this size, and that has been one of the best parts of my time here.

Casey: Being part of the DEFCON planning team, with the support of the Gordian Knot Center, has been the highlight of my time at Stanford so far. Under the leadership of Dr. Joe Felter and Jack Faherty, the center was instrumental in helping us plan and execute an operation as complicated as DEFCON. Their insight on national security and technology helped us build a day of discussion with leaders from across the national security community.

Kim: Gordian Knot has given me a place to carry my commitment to public service into a new chapter. After ten years in the Army, I came to Stanford interested in how entrepreneurship and technology could help address the national security challenges I had seen firsthand. This community has helped me turn that interest into action.

Through DEFCON and my work building Praetar Systems, I’ve connected with people who bring different skills but share a sense of responsibility for the future. What means the most to me is the community’s willingness to take students seriously, to give us meaningful responsibility and the support to do something with it. That trust changes what you believe you can accomplish while you’re still in school.

What advice do you have for students interested in the advanced technology sector and the Gordian Knot community?

Casey: I would ask them: Do you want to make a difference for our country, and do you want to work on the hardest problems facing it? Serving your country does not only mean signing up for military service. Solving the toughest problems our nation faces is admirable, whether you’re working for a company in the defense industrial base or as a federal civilian in national security. Both are honorable. We need young, motivated, mission-driven Americans serving at all levels.

Kim: Start with a problem you care about, and get close to the people who experience it. Understand what they need, where existing approaches fall short, and how your skills could help. There is room in this sector for engineers, entrepreneurs, policy thinkers, and people who can connect those disciplines.

Then show up and take responsibility. Attend a Gordian Knot event, join a project, volunteer to help organize something, and follow through. Anyone can contribute. DEFCON showed what students can accomplish when they take an ambitious idea seriously and commit to doing the work.

Zerrer: Show up and ask for real work. You do not need a military or technical background to contribute. DEFCON is run by students from across the university, and we learned most of what we did on the job: raising sponsorship, recruiting speakers, and hosting hundreds of people for a day.

I would also spend time with the people who use and build the technology. The fastest way I know to understand this sector is to listen to operators and founders describe the problems they are stuck on.

This article was supported by research using Anthropic Claude. All outputs were reviewed by humans and the final expression was produced by the authors.

Chinese President Xi Jinping's three-day state visit to Washington, D.C., delivered lavish pageantry and a two-month extension of the U.S.-China trade truce but few tangible breakthroughs, according to Mark Lambert, APARC's China policy fellow. He cautioned, however, against scoring the summit as a win or a loss for either side. 

"Having people talk to each other, particularly the leaders of the two strongest economies on planet Earth, is a good thing," said Lambert, a retired U.S. diplomat who served as one of the State Department’s leading China experts, most recently in the role of Deputy Assistant Secretary of State for China and Taiwan and as the Department’s China Coordinator.

Lambert joined APARC Director Kiyoteru Tsutsui on the latest episode of the APARC Briefing video series to assess the outcomes of the Trump-Xi summit and analyze the issues shaping the U.S.-China relationship.

Sign up for APARC newsletters to receive expert analysis and insights from our scholars and guests >

A Truce Extended, Big Questions Deferred

Lambert described Xi’s state visit as “the third movement in a five-part concert,” following the two leaders’ October 2025 meeting in Busan, South Korea, on the sidelines of the APEC summit, and Trump’s May 2026 visit to Beijing. These engagements spurred talks about a framework to stabilize the tense U.S.-China relations. Looking ahead, Xi has invited Trump to the APEC summit in Shenzhen in November 2026, and Trump hopes to host Xi at the G20 in Florida in December.

Each leader has domestic reasons to keep bilateral tensions low, Lambert pointed out. In the United States, the midterm elections and soaring gas and grocery prices are front and center. In China, high youth unemployment and demographic pressures loom large, and Xi, facing a Party plenum next month, needs to show he has the American relationship in a good place from a Chinese perspective.

"Both leaders have bigger fish to fry right now than antagonizing one another," Lambert explained, noting that the White House did an effective job of keeping expectations from Xi’s state visit modest. Washington and Beijing agreed to extend their trade truce through January 2027, but announced little else beyond that and a commitment to hold further talks, and provided no specificity on the economic and trade aspects of the U.S.-China bilateral relationship.

“It seems like what came out of last week's summit was: ‘Let's carve out a few areas where we can trade, recognizing that the vast majority of sensitive areas are either too delicate or too complex, not ripe right now to deal with,’” Lambert observed, predicting that the likely trend moving forward is more de-risking by both sides. 

Skepticism on AI

The White House announced the U.S. and China have agreed to start a dialogue on AI, with a communications channel to help defuse the technology’s risks. Lambert, however, is doubtful this effort will yield much, noting that a similar commitment emerged from the Biden-Xi 2023 meeting in San Francisco but was never fleshed out.

"I am not convinced that China is all that serious about working on artificial intelligence,” he said. “Every time we've tried to have meetings on AI, they will send diplomats; they will send political-type figures. They don't send real experts on artificial intelligence. And that makes it very difficult to talk about safeguards and protecting ourselves from accidents or other things.”

The American and Chinese summit readouts also describe AI quite differently. While the White House said China had agreed to the term "superintelligence," Lambert emphasized that "not once did the Chinese use the term superintelligence in their readout."

"It's good that they're talking about it," he said, but "I will be very surprised if it leads to real, meaningful discussion about these core issues on artificial intelligence.”

Taiwan, Japan, and U.S. Allies

On Taiwan, according to Lambert, Trump did not give Beijing the language it has long sought: that Washington "opposes," rather than "does not support," Taiwan’s independence. Having put forward the largest arms sales package to Taiwan in U.S. history, the Trump administration has postponed it to avoid jeopardizing the summit with President Xi. With two additional in-person meetings between the two leaders expected this year, Lambert predicts the arms sales deal will be postponed until the new year. 

Asked about Xi's referencing shared U.S.-China history during World War II, when the two nations fought together “against Japanese militarist aggressors,” Lambert replied, "History is history. We don't deny history, but I don't think it's wise to let ourselves become a prisoner of history." He called Japan "the keystone relationship with the United States in the Pacific," arguing the moment was "a missed opportunity for President Trump" to acknowledge the wartime relationship with China while also affirming the alliance with Japan and urging Xi to improve ties with Tokyo. He reiterated long-standing advice to new presidents: talk to allies before going to Beijing.

Iran and China’s leverage

Lambert said that China, "hands down Iran's biggest trading partner," has been reluctant to take on the responsibilities of a global power in international affairs. He pointed to Chinese banks and shipping companies that help Iran circumvent U.S. sanctions. Despite subdued rhetoric from Beijing at the United Nations, "I haven't seen anything to suggest that China is going to lessen its support of Iran." Beijing could help Washington end the unpopular war with Iran, but is unlikely to use that leverage.

Symbolism and what comes next

Lambert called President Trump’s personal welcome of Xi at Joint Base Andrews "really unprecedented" and "overdone," and said Beijing is portraying it as American recognition of China as an equal. Still, he cautioned against writing off the United States as a declining power. “Yes, we make mistakes all the time, and I personally think we're in a pretty difficult time right now. But don't count us out.”

In countries far along in the demographic transition, rapidly aging populations are colliding with a severe shortage of health care and elder care workers. Japan may be the most pertinent case study of this and other impacts of demographic disruption. The country’s fertility rate is decreasing faster than projected, and as of 2025, 30% of its population was aged 65 or older. These trends have contributed to a persistent gap between surging demand for elder care and a shrinking care workforce.

To cope with this crisis, the Japanese government has implemented several programs to promote the use of robots in nursing home care. These care robots (or “kaigo robots”) assist care workers and support residents with daily needs. But how does their adoption affect workflows, staff retention, and overall staffing of long-term care facilities? Few studies have empirically examined these questions.

Shedding light on the debate surrounding new automation technologies and their impact on workers, a recent study, published on August 6 in the inaugural issue of the journal Health Affairs Review, finds that care robots help address challenges posed by population aging and care-labor shortages without necessarily displacing workers or lowering care quality.

The study, titled Robots and Labor in the Service Sector: Evidence from Nursing Homes, is part of a collaborative effort driven by the Asia Health Policy Program (AHPP) at the Walter H. Shorenstein Asia-Pacific Research Center. The study’s co-authors are Stanford health economist Karen Eggleston, the director of AHPP; Toshiaki Iizuka of the University of Tokyo’s Graduate School of Economics, a former AHPP visiting scholar; Yong Suk Lee of the University of Notre Dame’s Keough School of Global Affairs, formerly the SK Center Fellow at Stanford’s Freeman Spogli Institute for International Studies; and Ellie Xi, a rising junior at Stanford studying biomedical computation on the pre-med track and a former research assistant to Dr. Eggleston.

“One of the aspects of the paper that I found most interesting is its attempt to move beyond simply documenting where robots are being adopted and instead ask what their adoption may actually do to the long-term care workforce, especially given the sector’s high turnover rates and physically demanding working conditions,” says Xi.

Sign up for APARC's newsletters to receive our scholars’ research updates > 

Lessons from Early Diffusion of Robotics in the Long-Term Care Sector

The researchers grouped the robots prevalent in Japanese nursing homes into three main categories:

• Transfer aids: Wearable and non-wearable devices that assist staff with lifting frail residents from beds to wheelchairs, significantly reducing back strain, a leading cause of caregiver burnout.
• Mobility robots: Assistive systems that help residents move, use the restroom, or bathe more independently.
• Monitoring and communication systems: Sensors that alert staff when a resident has slipped out of bed or fallen, as well as social robots designed to offer companionship and cognitive stimulation.

To examine early diffusion of these robots, the researchers used 2017 survey data on long-term care work across approximately 860 Japanese nursing homes. A second source of data is their compilation of prefecture-level care robot subsidy availability, using regional variation in robot subsidies as an instrumental variable. Think of this as a natural experiment driven by a “geographic nudge.” Starting in 2015, Japan’s central government began providing prefectural governments with funds to subsidize the cost of adopting care robots. Because different prefectures introduced these subsidies at different times and with varying levels of generosity, nursing homes in some regions had a much easier path to acquiring and deploying robots than those in neighboring areas.

By comparing staffing trends between nursing homes that received this external, subsidy-driven "nudge" to adopt robots and those that did not, the researchers could confidently isolate the technology’s causal effects on the caregiving workforce.

The results indicate that nursing homes that adopted robots saw a significant decrease in staffing retention problems. Robot adoption was also associated with a 28% increase in care workers and a 39% increase in nurses, resulting in approximately 26% higher total employment at the facility level. Importantly, this employment growth occurred entirely among non-regular employees – that is, staff members on flexible, often part-time, contracts. The findings also suggest that robot adoption may improve continuity and quality of care while mitigating careworker shortages.

“Using nationally representative data from Japan and an instrumental-variable strategy, the study provides evidence suggesting that robots can complement rather than simply replace care workers, including by alleviating retention difficulties in a sector facing persistent labor shortages and potentially supporting quality of care,” explains Xi. “The paper’s central finding is that robot adoption is associated with reduced retention difficulties and increased employment of care workers and nurses under flexible contracts.”

A Human-Robot Collaborative Future for Elder Care

Caregiving is a physically demanding and high-stress occupation. The study’s findings indicate that, by introducing robots to ease staff burdens and monitor residents, nursing homes were able to promote more flexible work, either by increasing the tasks performed by non-regular employees or potentially encouraging part-time work. This type of work structure is encouraged by the Japan Nursing Association as a means for reducing stress, preventing burnout, and promoting a healthier work-life balance.

“Given recent debates about AI use in healthcare and the future of work, I found it especially interesting that this research highlights the potential for technology to augment caregivers, particularly by supporting physically demanding or indirect tasks, while preserving the human aspects of care that are more difficult to reproduce,” says Xi.

The study's findings are relevant for policymakers seeking evidence on the social value of collaborative robots and their impacts on staffing and care delivery, particularly as aging societies facing demographic pressures similar to Japan consider how to meet growing demand for long-term care.

A Mentorship Story

The study is part of Eggleston’s Innovation for Healthy Aging research track, which investigates how East Asian economies adopt and adapt to emerging technologies, design incentives, and adjust their workforces to support older adults. It builds on an earlier research project focused on the impact of robots on nursing home care in Japan.

For Xi, co-authoring the paper represents an academic milestone. "As someone deeply interested in geriatrics, I first began working with Dr. Karen Eggleston as a research assistant during the summer after my freshman year," Xi recalls. "Last spring, we decided to continue working together through a directed reading."

During that directed reading, Xi was invited to help revise and update a project that had previously circulated as an Asia Health Policy Program working paper. She worked to update the research, framing the original findings within the rapidly shifting landscape of care robotics and artificial intelligence in long-term care.

"I worked on situating the original findings within more recent research and developments in care robotics and considering what Japan’s experience can tell us about the role of these technologies in aging societies today," Xi says. "It was an incredible opportunity to work closely with Dr. Eggleston, Professor Toshiaki Iizuka, and Professor Yong Suk Lee. I was especially grateful to see firsthand how a collaborative project evolves and to contribute to that process as the paper moved from a working paper toward publication.”