The Good Guys Get to Fail

A few years ago, I read Peter Thiel’s Zero to One. I was interested in how an idea becomes a business that does more than copy what already exists. His thinking about markets, competition and monopolies gave me plenty to consider. The hostility he attracts here in Germany has often surprised me since. At times, even being curious about his ideas seems to count as taking a political side.

On September 24, Thiel received the Axel Springer Award in Berlin. Outside the publisher’s headquarters, protesters opposed the award. The criticism concerns, among other things, his support for Trump and Palantir’s work for security agencies. His conversation with Mathias Döpfner, released alongside the event, covers AI, Europe’s decline and America’s political future. It prompted me to look more closely at our own record.

Debates like these remind me of a date at an art museum. It was our second meeting, and the conversation turned to American companies. I mentioned that I invested in some of them, including Palantir. “You have blood on your hands,” she said. Afterwards, we went back to talking about the art and her studies.

My reasons for investing went unspoken. They have to do with a conviction that matters to me particularly as someone who values individual liberty: a free society must be able to defend itself. That requires the ability to identify threats, analyse information and distinguish military targets from civilians. Better information can help forces target attackers more precisely and spare bystanders. That is one potential benefit I see in Palantir’s work. How the company performs in actual operations, and what harm occurs in the process, deserves close examination. That would have required a longer conversation at the museum.

I oppose indiscriminate surveillance of the population. Sharing a government’s foreign policy aims does not entitle it to look into my private life. Its goodwill is too little protection for me. Individual liberty requires limits on state power that hold even when an intrusion is popular and the reasons offered for it sound impeccable.

In the interview, Thiel distinguishes two ways of being bad: evil, in the moral sense, and bad, as in incompetent or unfit for the job. A society can become so preoccupied with evil that it loses sight of the incompetence of those responsible for running it.

I keep noticing this indulgence in German political life. A politician clearly meant well, governed under difficult circumstances and made mistakes that others have made too. All of that may be true. But such generous understanding can obscure how much time and influence that person had to change a failing course.

Before Russia’s full-scale invasion of Ukraine, Germany obtained more than half its natural gas from Russia. That dependence had grown over many years, including under Angela Merkel. Her much-praised experience went hand in hand with a commitment to a policy that left Germany vulnerable over a vital energy supply. Households and businesses were eventually left to deal with the consequences.

I see a similar indulgence in the debate over migration policy since 2015. Taking people in requires housing, schools, language teaching, jobs and reliable rules. Whether such a policy succeeds becomes apparent in the daily lives of public servants, businesses and neighbourhoods, long after public attention has moved on. Questions about limits and practical outcomes are reasonable. Too often, the discussion then turns to the political character of the person asking them. Perhaps he simply wants things to work.

The difficulty of these decisions cannot remain the answer to every criticism of them. Governments had room to act, however constrained it was. The public, meanwhile, is asked to be patient: to understand the constraints and adjust to the consequences.

Thiel wonders in the interview at how secure Friedrich Merz’s position remains, despite what he sees as a poor record. He also suspects that potential successors have no better answers. Without the prospect of a different course, even a poor record eventually becomes something people accept; the alternatives would be no better anyway.

Even a government with better answers would face the same institutions and constraints. Coalition partners disagree, a government department objects, European rules stand in the way. Treaties and constitutions limit room for manoeuvre beyond individual electoral terms; at the European level, different national interests have to be reconciled. Under these conditions, an election can change the people in charge more noticeably than the direction of policy.

A serious change of course would therefore have to address these obstacles too: clarify responsibilities, reform procedures that block progress and build the political support needed to change European rules. Protecting individual rights is distinguishable from preserving rules that make a project more expensive or slower without a corresponding benefit.

Entrepreneurs receive rather less indulgence. A failed venture is readily taken as proof that its founder overestimated himself; a successful one becomes a reason to resent his influence and wealth. Someone stakes capital, years of their life and their reputation on an idea. An environment that mocks the attempt and envies its eventual success does little to encourage others to try.

Thiel’s suggestion of a fear of success concerns founders themselves: why they sell or stop rather than build much larger companies. It also raises a question about the conditions in which they make that choice. What makes continuing to build here attractive once a business is large enough to choose between locations and sources of capital?

The economic obstacles are described at length in the Draghi report. Europe has research capabilities and entrepreneurial talent, yet fragmented markets and regulatory burdens make it harder to finance and grow new businesses. An additional grant programme may make a project possible while leaving untouched the procedure that makes it unnecessarily expensive.

A government needs to win majorities to change course and often has to reverse its own earlier decisions. A new company can try another approach without first convincing everyone invested in the old one that they were wrong. In a society where political change is difficult, this possibility should be especially valuable. Additional barriers to starting and growing businesses close off an avenue for change outside politics as well.

This openness is what draws me to America. Elections can bring a tangible change of direction, while people independently find the means to set something in motion with an idea. To me, that is a substantial part of what makes a society free.

That freedom also includes being left alone by the state. European institutions can, however, go a long way in intruding on privacy in the name of a good cause. As early as 2022, European data protection authorities warned that the Commission’s proposal to combat child sexual abuse could lay the foundations for general and indiscriminate scanning of private communications. Protecting children was being invoked to justify measures that would reach deep into the lives of people suspected of no wrongdoing.

Invoking a cause like this makes it harder to debate the means. Extensive surveillance can be presented as necessary protection, while its critics come under suspicion of being indifferent to the underlying concern. Agreement on protecting children tells us little, however, about which intrusions into privacy would be proportionate to that aim.

Camus wrote: “The welfare of the people in particular has always been the alibi of tyrants.” In his critique of supposedly humanitarian punishment, C. S. Lewis considered a related, still more uncomfortable possibility: that those in power sincerely believe in the good they intend to force upon someone. Resistance can then seem to them like further proof that their paternalism is needed.

What interests me most about Thiel’s 2009 statement that freedom and democracy are incompatible is the question of what claims a majority may make on an individual’s life. An intervention can be popular and still restrict that person’s freedom unjustifiably. My distrust of such power applies to European authorities as much as to American governments and their private contractors.

I now spend less time thinking about migration, energy policy and many of our other political disputes than I used to. AI has changed my sense of the future we are heading towards. Work, prosperity and political power could shift so fundamentally that some of today’s controversies may mean something quite different in a few years. Much of what passes for long-term planning in our political debates assumes a remarkably unchanged world. I find that assumption harder and harder to believe.

And yet, with AI, the old politics catches up with me immediately. Data centres need electricity and grid connections. Companies need capital and room to grow. The people who could build them have other countries to choose from. Existing wealth can cushion the effects of poor conditions for years. When it comes to building a new industry, it becomes apparent how little of what is already taken for granted elsewhere is available here.

With AI, circumstances can change while a project is still working its way through our procedures. A permit can look like progress on paper even though the team that needed it is already working somewhere else. Other countries keep researching, building companies and gaining economic and political influence. Standing still would do little to preserve Europe’s present position. It would leave us increasingly dependent on those who keep moving forward.

In the interview, Thiel warns against the idea that stagnation is a safe alternative to technological progress. He also talks about the expectation that our children will have better lives than we do. We need growth to sustain and increase our prosperity. But the prospect of a better life matters personally too: being able to build something of our own, to make plans we look forward to carrying out. A life devoted to keeping things from getting worse is a bleak prospect.

AI could bring us better medicines, enable scientific discoveries and help with problems we have yet to solve. Precisely because this development holds so much promise, its safety matters to me. We still have no reliable solution for controlling systems that are far more capable than we are and act autonomously.

In my last essay, I argued for more work on AI safety alongside a faster pace in Europe. A dangerous step in development deserves careful scrutiny. Research into it requires people, computing power and energy. Our delays in providing these foundations do not make other countries’ systems safer. Their immediate effect is to reduce our ability to contribute to that safety.

Europe’s freedom will also depend on whether people and companies here can find the means to help shape this future. Much of what we lack today can be traced to decisions made long before the current debate over AI. Their consequences help determine which options remain open to us.

The political class has retained a remarkable claim on our patience. Others have long been living with the consequences of its decisions, while it continues to cite its experience and warn of the risks of change. I would like to see less indulgence towards that record.