China’s AI Advance—Hype or Bonanza?

Plus, a Mossad official explains the Hamas-Israel negotiations, and more

DeepSeek claims it can offer artificial intelligence cheaper than its U.S.-based competitors.

The market woke up today to the significance of DeepSeek, an artificial intelligence company based in China that recently released a new model.

A technology investor who is Jared Kushner’s brother, Joshua Kushner, posted skeptically about DeepSeek, portraying it as a security risk and as a setback for American dominance: “‘pro america’ technologists openly supporting a chinese model that was trained off of leading US frontier models, with chips that likely violate export controls, and — according to their own terms of service — take US customer data back to china.”

A fellow at Stanford’s Hoover Institution, “Grumpy Economist” John H. Cochrane, says cheaper AI may be good news longer-term: “The winners will not be the producers of AI, which looks to become a marginal cost commodity with remarkable speed, but the users of AI. …The profit, and ultimate benefit, of railroads was not so much in the railroad itself, but in the wheat fields of Kansas.”

Cochrane also writes, “imagine the world some of the ascendant right wants to take us to, in which anything “China” is banned from the US. We would not see DeepSeek. Competition is the main source of efficiency, and competition needs to be global in the world of 100 million dollar upfront costs.”

I’m all for competition. One of the ways American companies can compete against China is by making the case to Americans that, because the American companies aren’t subject to influence of or domination by the Chinese Communist Party, the American customers of those American companies therefore are free from having to worry about a sudden interruption of service in the event of a sudden escalation in U.S.-China tensions.

An example of this is hospital masks, gowns, and surgical gloves. If you are an American hospital that gets all its supplies from China, that works well enough, until March 2020. Then, suddenly China is using all those medical supply materials itself and not shipping any more of them to America. There’s a tradeoff between the cheaper prices of the Chinese products and the reliability of knowing that a supply chain that isn’t totally China-dependent is likely to be more reliable and resilient in a crisis that involves China. (That said, the price differences are vast. For example, here is a pair of made-in-China men’s shoes, available at Target, for $35. Here is a pair of made-in-America men’s shoes, for $706.)

There are also laws regulating competition, notably in the U.S. Constitution’s patent clause, which gives Congress the power “To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries.” What qualifies as an inventors’ discovery is a question that provides plenty of lucrative work for patent lawyers; stealing an inventors’ discovery isn’t competition, it is theft.

Anyway, how you view news of the DeepSeek release depends on what storyline you primarily place it in.

Kushner sees it in the context of U.S.-China geopolitical competition.

Cochrane sees it in the context of productivity growth and technology-fueled dynamic prosperity. As Cochrane put it: “the stocks of companies that can potentially use cheap AI did not jump up. Maybe they will, once more people figure out how to use cheap AI to make profits for a while. Maybe those companies haven’t been founded yet. That is now the Wild West.” If you bought the Wild West when it was the Wild West, today you’d be sitting on a lot of nice real estate in the San Francisco Bay Area, Scottsdale, Beverly Hills, Santa Fe, Aspen, Telluride, and similar locations.

My own take is that it’s not an “either-or” but a “both-and.” We’re both in a period of conflict and potential conflict with Communist China and in a period of growth and potential growth of artificial intelligence. The China competition carries great risk of economic disruption or even chaos and war if mismanaged, but also the chance of growth and benefit if China moves in the direction of economic and political freedom and rule of law. Likewise, the AI revolution will bring disruption and also tremendous upside potential for efficiency and productivity in applications including pharmaceutical discovery, dynamic pricing, investment management, marketing and customer acquisition, and risk analysis.

The DeepSeek news is a plot point in both storylines simultaneously. If you just see the techno-optimism storyline, the Kushner China concern seems misplaced. And if you just see the China threat storyline, the Cochrane techno-efficiency storyline seems misplaced.

Transformative education: The dean of the Harvard Divinity School, Marla Frederick, sent an email to the community today announcing Diane Moore’s retirement as director of Religion and Public Life, David Holland’s appointment as interim director, and Terrence Johnson’s takeover as director on July 1, 2025. The email concluded, “as we move forward in service of our students, we are committed to continuing to deliver a transformative educational experience.”

I find the whole “transformative” language, which Harvard College also uses, to be problematic. Sure, part of education is challenging pre-existing assumptions, but in some cases the transformation that happens is that an institution takes incoming students who love Israel and America or who are at least neutral and transforms those students into haters. What if the entering student doesn’t want or need to be transformed but simply wants to be educated, to acquire more skills, background, and training, or to help to advance the frontiers of knowledge? What if the faculty would prefer aiming for educating the students instead of transforming them? These institutions have lots of “stakeholders”—a word I generally avoid unless I am mocking it—but among the students, their parents, the faculty, the government, the governing boards, the alumni and institutional donors, the administrators, potential future employers, you wonder who the “transformative” language is there to appeal to.

Israel-Hamas negotiations: Anonymous sources are like the word “stakeholders”—usually avoided here unless mocking. Yet Israel’s Defense and Security Forum recently aired one of its “Zoom briefings” that it did not post and record, because it featured a former Israeli spy agency official who had been involved in the indirect negotiations with Hamas. His analysis: “Hamas is basically playing for time” and will try to draw out the negotiations beyond the 42-day period that is the first phase of the present ceasefire. It will try to “make it look like Israel’s fault” when the negotiations break down.

“There is no bridge between what Israel wants and what Hamas wants,” the official said. “Israel demands complete surrender by Hamas. I just don’t see Israel allowing Hamas to survive. I just do not see Israel completely retreating from the Gaza Strip.”

As for Hamas, “I just do not see them giving up what has been the best leverage that they have.” Keeping twenty or fifty hostages allows Hamas to fuel political dissent within Israel.

“For Israel, defeat in this war is anything less than the complete destruction of Hamas in the Gaza Strip,” the official said. “Palestinians will consider it a victory.”

As a result, he said, “there really isn’t any future of these negotiations.”

Vandenberg recommendations: The Vandenberg Coalition, a bipartisan group chaired by Elliott Abrams, is out with a set of concrete policy recommendations about the Middle East for the new administration. It identifies key American interests, among them, “prevent Iran, Russia, and China from deepening their influence in the region,” and “deny jihadi terrorists a safe haven.” It recommends moving U.S. Central Command headquarters out of Al Udeid Air Force Base in Qatar. It also recommends removing Qatar’s status as a major non-NATO ally and conferring that status on the United Arab Emirates. Several of the report’s recommendations have already been implemented by the Trump administration.

One new idea to me at least: “The United States should propose a Security Council resolution that states the use of human shields is a crime under international law and that those who use human shields are responsible for the civilian deaths in which they result.” China or Russia might veto, but even so, it’s an intriguing idea.

The approach to Lebanon is also unconventional: “U.S. policy should treat Lebanon as a state captured by Iran unless and until Hezbollah’s grip weakens.”

A down-the-road idea: “The administration should extend the 2016 U.S.-Israel MOU, which expires in 2028. The new MOU should reaffirm America’s commitment to Israel’s Qualitative Military Edge and ensure the timely resupply of weapons. The MOU will strengthen Israel’s ability to simultaneously defeat a multi-front Iranian proxy network.” I hope that by 2028 the revolutionary Islamic regime in Iran and its proxy network will be history, so Israel’s need for arms to defeat it will be diminished, but it is a reminder that one of the deals President Obama cut on his way out of office in 2016 was a ten year, $38 billion deal with Israel and its then Prime Minister, Benjamin Netanyahu, for military aid. The Vandenberg Project doesn’t recommend a specific level for this aid. Some pro-Israel voices oppose it, saying it offers the U.S. undue leverage over Israel’s ability to operate in its own self-defense. Israel is moving to make more arms itself domestically, but some weapons systems, particularly aircraft and helicopters, are so large and complex that the involvement of U.S. manufacturers is almost unavoidable. Republican congressman Jack Kemp championed an Israeli fighter known as the Lavi, but that project was cancelled. Anyway, for those of us who hadn’t realized that this issue is on the horizon for the U.S.-Israel relationship, thank the Vandenberg Coalition for the reminder.