These posts were removed from the frontpage owing to their staleness. They remain here for posterity.

Government and the internet

Over the last two decades, tension between government and the changes caused by the internet has been a recurring theme. Today, they’re almost seen as opposing forces. This is somewhat strange when you think about it. Most technologies don’t cause so much ongoing upheaval.

Why is the internet so challenging? I decided to make a list of reasons and came up with 11. Though none of them are very novel, I found the catalog interesting. (For one thing, I expected fewer.)

1. Better communication tools mean that events simply happen faster. The Arab Spring would probably have been less problematic for incumbent governments if things had happened more slowly. They’d have had more time to react. But governments still operate on timescales of days and weeks while it’s becoming easier for events to play out hour-by-hour.

2. Communication and publication can no longer be censored. Wikileaks, the Arab Spring, and Snowden’s revelations all depend on governments being unable to prevent mass dissemination of information.

3. Individuals can now cast a spotlight on government action. (That is, publication isn’t just uncensorable; it’s also decentralized.) With the internet, we can rapidly rewire our communication networks when a node becomes an important source of information. (This is most noticeable during live events, when people rapidly figure out which Twitter users are worth paying attention to.) If a government does something wrong, and even if the mainstream publishers ignore it, the information can still be rapidly broadcast to millions.

4. In a global community, local aberrations stick out more. Because online news and social structures don’t pay much attention to national boundaries, laws and government actions are increasingly held to international standards. People are more likely to know when they’re being screwed over by their governments.

5. It’s harder for governments to broadcast directly to their citizens. Fifty years ago, it was easy: just run a prime-time TV segment. You can still do that today, but many fewer people will pay attention.

6. Governments are more fragile and hence weaker. It’s almost as easy to leak a database as it is a file. It’s much harder for governments to maintain secret structures, and they must contend with the omnipresent risk of a calamitous leak.

7. Governments are more powerful and hence more likely to overreach. Because it’s now far easier to eavesdrop on communications, maintain intrusive databases, etc., it’s much more tempting to do it. Thirty years ago, you needed to adopt extreme GDR-style tactics to eavesdrop on everyone. It was logistically prohibitive, and most governments would probably reconsider when they realized what doing so would actually entail. Today, technology improvements mean that it takes much less effort—and evidently that it feels much less wrong.

8. Because a lot of internet activity is oblivious to national borders, jurisdictional questions are now much murkier. Internet companies and services aren’t so much proactively global as casually global: the default is to be available everywhere. As a result, internet services don’t consider the legal implications of expansion into each jurisdiction. They don’t really expand into jurisdictions at all. They’re just there. We don’t have a good framework for regulating these entities. Should a US website be subject to Australian libel law? (To the union of all libel laws?) The most obvious solution is to have an internet company be subject only to the laws of the country in which it resides, but even this doesn’t work: companies can easily operate from multiple countries. And, of course, countries affected by a non-resident service aren’t likely to be satisfied by this solution.

9. A lot of laws make less sense today. Remix culture, Airbnb, YouTube, Lyft, and Bitcoin challenge existing regulation. The harm of bad laws is often invisible, which makes weighing the trade-offs difficult. (How many good things aren’t happening and how good would they be?) It’s hard to argue against regulation before the potential of a new technology has been realized—the possibilities are, by definition, merely hypothetical.

10. Since a lot of industries are being reshaped by technology, more incumbents are seeking regulatory protection. Goverments sometimes acquiesce and enact bad laws (DMCA), which in turn angers citizens.

11. Legislators lack the conceptual framework to reason effectively about internet and software issues. I think this might be the biggest problem of all. As industry insiders, we have an advantage: we know it’s inane to talk about "getting data back"; we know that metadata and data are often distinctions without differences; we know that large datasets are very hard to anonymize; we know that a large dataset will rarely be used only for its originally intended purpose. We know this simply because we’ve watched these issues play out many times before. Politicians haven’t, and when policy questions hinge on understanding technology, they don’t tend to fare well.

Looking back over these, the main thing that strikes me is that it’s very hard to imagine any of them going away or being resolved any time soon. The internet was an epochal jolt and our governing structures have yet to catch up. We’re now twenty years in; given the pace of change, I expect these issues to persist for most of my life.

I find it interesting that the internet is still politicians’ example of choice when they need to show how the future is exciting. The US is today, in Obama’s words, “the nation of Google and Facebook.” The problems don’t simply stem from obliviousness. Even leaders who recognize the internet’s importance and value are stuck in an edifice almost guaranteed to yield a dissonant relationship.

Could Stripe have been started in Ireland?

I’m sometimes asked whether Stripe could have been started in Ireland. It’s impossible to really know the counterfactual, but I suspect not.

Stripe wasn’t easy to get off the ground. It required convincing banks to work with us, and to take a bet on an unproven startup. I actually spoke to some banks in Ireland when we were starting out, and it was pretty clear that getting them to work with us would have been a very tough battle. I doubt we’d have been able to pull it off.

But I’m not sure that asking whether Stripe could have been started in Ireland is the right question. Most technology start-ups don’t have to convince banks to work with them. The interesting question is probably “how hard is to start a successful start-up in Ireland compared to doing so in Silicon Valley?” (I’m not claiming that Stripe itself is yet successful—it's still very early days.)

The answer is almost definitely “substantially harder”. Among the reasons:

Despite these issues, I’m struck by how things seem to be improving every time I come back to Ireland. Dogpatch Labs recently opened in Dublin, and Eamon Leonard has been organizing a successful tech events series. Conferences like the Web Summit, f.ounders, FunConf, and Úll have substantially increased Dublin’s profile. Thanks to firms like Atomico, Accel, Index, DST, and others, the European investment scene generally looks much better than it did five years ago. Early stage tech companies like Hubspot and Zendesk have opened offices in Dublin. In general, things appear to be moving in the right direction. Given its size, Dublin compares surprisingly favorably to Berlin or London as a tech hub.

It also goes without saying that it’d be a mistake to look at any of the problems identified in this post as immutable, and it’s worth keeping in mind that they apply to most areas of the world that aren’t Silicon Valley. The first two issues mentioned above have gotten better since we started a company in Ireland in 2007. But given how much effort is (rightly!) being put into making Ireland a better place to start a start-up, it’s probably worth being clear about what should be improved.