When Safety Sucks

Safety is good. Except when it’s not.

Most of the time, we equate making something safe with making things better. But what if that isn’t the case? What if we’re actually making things worse?

Greg Ip has a number of examples that force you to re-examine your assumptions in his recent book ‘Foolproof. Like the stats that prove car drivers with snow treads consistently drive faster as the weather gets worse. Not unlike the crew of the Titanic who sailed fearlessly through icy waters, believing it to be invincible…

The theory of Risk Compensation suggests that people adjust their behaviour in response to the perceived level of risk. They take more care where they sense greater risk – and less if they feel more protected. Take American Football as an example: the players wear helmets yet the sport appears to suffer from more frequent serious injuries than other comparable contact sports, such as rugby and Aussie rules football. It’s the same line of thought that led to the removal of headguards in amateur boxing in recent times. In some ways it’s similar to the recent stories about the school playgrounds that are intentionally creating more dangerous surroundings for kids, in order to give them the opportunity to learn.

The problem is that we consistently fail to take into account the feedback loop. In other words, how does our behaviour change when new measures are introduced?

Now, individual risks are one thing. But across an industry, they become far more significant and become systemic risks. Take life insurance as an example. It works because not all policyholders die at the same time. But of course that needs to be wrapped up into a different package to be palatable. That’s seen by the fact that industry markets itself to appeal to emotions and feelings, instead of accurately selling itself on the basis of probabilities. Yet as modern life becomes more complex, we face big problems when risks become correlated. And that situation deteriorates rapidly if that correlation is only visible in extreme situations.

Predictably the financial services industry leads the way in providing us with a powerful example of how things can go wrong. The global financial crisis of 2007/2008 had many causes. But in reinforcing an industry with a belief that others would step in to support organisations because they were ‘too big to fail’, it’s certain that risks were taken that shouldn’t have been. That’s exactly why the decision to let Lehman Brothers fail had such an impact – because it shattered the belief that existed at the time about the invincibility of financial institutions. And by definition, that wasn’t a risk that was seen as likely at the time.

We often take out insurance to guard ourselves against existential risks – whilst forgetting the fact that by definition, no system can ever insure itself against total collapse. So, can we blame the insurance industry? After all, the whole point of insurance is that someone else bears the cost if things go badly.  Does insurance actually just encourage others to act in a riskier acting way?

After all, flood and earthquake insurance enables more people to live in areas where those events are likely. And the constant stream of financial insurance products make it easier for more investors to pile into markets. The result? The bad events that we’re insuring against become both more likely and severe.

So should we be dismantling some of these safety barriers? In many cases, it depends on your own motivations. Unfortunately, the reality in the City is that the greatest rewards are returned to those who pursue the greatest risks. Somehow we’ve created a system where those who pursue higher (leveraged) risk make the most money.

But the question should be posed in many different situations. Take forest fires as an example. Man has tended to view fire as an event that should be suppressed. And yet it has become increasingly clear in recent times that teh opposite is often true: we need regular fires in order to clear away some of the lower levels of vegetation that accumulate. So by making things ‘safer’, when the inevitable fires do come, we’re now seeing them turn far more quickly into unstoppable ‘megafires’ – because they have far more fuel to burn.

So we need to ask: should we actually be chasing a good disaster now and again to reset the system? Perhaps. But most times, that’s not a decision that’s politically acceptable for a whole number of reasons.

Take flying as an example. Flying is now so safe (the highest risks are at takeoff and landing) that there are fewer opportunities for pilots, regulators and others in the industry to ever learn from accidents. It’s a ‘problem’ that’s compounded by the fact that with the more mundane incidents recognisable and under control, any accidents will increasingly be of the truly mysterious, unimaginable variety (e.g. 9/11, the disappearance of Malaysia Airlines 370 in 2014).

And how many people would be happy with ‘a few more crashes’ to help us improve. As an aside, it’s worth looking into how the aviation industry has become so safe. As the book explores, the industry is interesting because it’s full of ‘high reliability organisations’ (which have a preoccupation with failure).

Ultimately, the somewhat counterintuitive result is that any efforts to make the surroundings safe triggers behaviour that frustrates those efforts. So lenders who expect to be bailed out in a crisis will lend more cheaply – but that in itself encourages more investment.

There’s no easy answer here. Perhaps things will become clearer as we gain more visibility about connections between systems work in the modern world. I doubt it though. It feels like the world is becoming even more complex. And the fact that very few saw the financial crisis in 2008 is indicative. As Ip writes in the book, the Dutch who have a history of a thousand years of building dams to prevent their country from flooding have an expression: “There are two types of levees: those that have failed. And those that will“.

More people globally are now living and working in areas that are either risky or unsuitable due to a range of reasons. As a result, environmental disasters (such as floods) are now more destructive than ever before.

If we accept danger, does it make us ultimately more prosperous and safe? If so, perhaps we shouldn’t be working to prevent these things happening. Perhaps this is the price that we pay for living in desirable and productive places. Maybe the focus is better spent on minimising the damage when they inevitably do.

Bitcoin v The Blockchain?

“blockchain technology…is presented as a piece of innovation on a par with the introduction of limited liability for corporations, or private property rights, or the internet itself” (The Economist)

Over recent months, we’ve seen a number of recurring themes in the Bitcoin community. Whilst the heated block size debate appears high up on that list, I wanted to touch on another key narrative that’s continuing to provoke much discussion.

Driven by rising interest levels within the mainstream financial services industry, the issue revolves around a deceptively simple question.

Can We Have Blockchains without Bitcoin?

I’ve had a number of conversations recently within traditional financial services organisations who are starting to explore the potential of blockchain technology. But one thing is clear. Whilst the technology is becoming increasingly attractive to them, Bitcoin itself is still sometimes seen as a weird crypto-anarchic-libertarian concept that must first be set aside in order to let the real work commence.

Now for those who understand Bitcoin in depth, this is surely a deeply flawed concept. When it comes to the blockchain and Bitcoin, each needs the other for survival right? And any attempt to split the two will kill the value of both.

But innovation should be welcomed in whatever form it takes – whether incremental or ‘Big Bang’ disruptive. Whilst the second option is clearly more attractive, I’d argue smaller steps are still worth taking when the true enemy is actually the status quo. The value of innovation is often subjective in any event, viewed differently according to each person’s unique mix of perspective and timescale.

So, to draw a parallel from my past life, a law firm that develops automated document assembly plans for the future will be dismissed without a second thought by those who are seeking to develop smart contracts that automate those firms themselves out of existence and vice versa. But then the technologists fail to develop an understanding of the all-too-real existing constraints and the incumbents fail to set their sights high enough.

The lesson? If opposing sides simply dismiss competing arguments without taking the time to genuinely explore their merits, both risk missing out on the subtleties. So with this post, I intend to do just that. Not to provide answers per se but to collect some key arguments from both side of the debate. Of course, it’s merely scraping the surface. But it’s worth starting with an overview.

Let’s Start With The Easy Bit – A Misunderstanding

Many who repeat the “blockchain good, bitcoin bad” refrain have overlooked one key fact. Many businesses that the press are now championing as being pioneers of blockchain technology are actually reliant on Bitcoin.

Whether that’s NASDAQ’s decision to use ‘a blockchain ledger’ on its Private Market platform, the use of ‘distributed ledgers’ to streamline financial settlement by Digital Assets Holdings (check out this recent talk by the CEO, ex-JP Morgan senior exec Blythe Masters) or even Overstock’s CryptoBond, any mention of Bitcoin itself is often abstracted by the press in favour of a focus on blockchain technology.

So it’s important to remember that each of these not-insubstantial bets have been made by people and organisations who believe that Bitcoin’s blockchain will succeed (check out the diplomatically-worded blog post by Peter at CoinCenter for a similar perspective).

For the rest of this post, I’ll call the (permissionless) blockchain that underpins Bitcoin the ‘Satoshi Blockchain’ for clarity.

The Argument for No

Secure. Public. Permanent. Immutable. Massively resilient. Just a few of the features that describe the Satoshi Blockchain. Yet all of these features cannot exist without one thing: Bitcoin miners who believe that they will earn more money than they spend if they turn on their computers to take part.

Incentive is crucial to the Satoshi Blockchain. If it will cost a miner £151 to get one Bitcoin by turning on his computer to mine when he could simply buy one for £150, which option do you think he will choose?

This mechanism beautifully balanced. And the result? Take away the bitcoins and the Satoshi Blockchain can’t function.

Unless the value of bitcoins to be won by miners exceeds the money they spend in burning electricity and investing in the capital infrastructure necessary to compete for dominance in the current mining arms race, the Satoshi Blockchain would be far less secure – and all of those fascinating projects listed above would be going nowhere.

The mining game might be preparing to undergo significant change with the resolution of the current block size impasse, just as 21 turns on its mining capacity (and BitFury their lightbulbs!). But irrespective, under the current model, it’s clear that the miners need to be paid for the system to work.

So if it’s clear that those who secure the network need a financial reward, why can’t we simply pay them with fiat currency, instead of bitcoins?

For one very good reason. Do that and you destroy one of the characteristics that has enabled Bitcoin to thrive over the years.

Censorship resistance.

In other words, if you build a permissionless system (that is open to anyone to join) that requires money, it must use a native currency. You cannot use money that is issued by a third party because there is always a risk that that third party might then censor (e.g. delay or withhold) payments. Instead, the money must be created within the system itself and it must function as a bearer instrument itself (like cash) – in other words, if you hold it, it’s yours (with no third party rights cutting across the ownership).

And this is why it is impossible to separate Bitcoin from the Satoshi Blockchain. There is simply no other way to incentivise the miners. It needs native digital money that is valuable and a guarantee that this will only be distributed according to the rules of the protocol.

The Argument for Yes

So it’s agreed then. If you want to reach decentralised consensus under adversarial conditions, you need an incentive structure.

But – hold on a second. Proof of Work (a security model that intentionally makes it very expensive to attack the network because it requires miners to burn all this electricity) is surely only necessary under certain conditions.

What if you’re a financial organisation? You don’t actually want the system to be open to everyone. You want to restrict those who have the ability to update the ledger to an agreed number of known people – because if it all goes wrong, you’ve then got someone to sue. You’re perfectly happy with the security systems that you currently use within your organisation to date and provided you control the small number of people who can update the ledger, you don’t need Proof of Work to protect the integrity of the network.

And I believe that it’s here that you see the key point emerge. Bitcoin’s model works amazingly well because it is a close-to-perfect solution for people for whom the anonymous verification of transactions carried out on a P2P basis (meaning that no third party is required) is one of the key attractions.

But within financial services, a heavily-regulated sector that views anonymity as a weakness and collectively has no desire to avoid censorship by third parties, a demand for a related type of blockchain-technology with a different feature-set appears to have emerged.

Here, there’s no need for miners. Save the money and replace them with ‘permissioned’ blockchains. This simply means that those using this private blockchain can ensure that someone who wishes to update the ledger must be authorised in advance and their identity known. Suddenly we have a very different beast. Now we have a database tool that no longer needs tokens that command a market value – because they have no need to act as an incentive. And of course, it avoids any uncomfortable questions about bitcoins within an environment that demands a high level of compliance and box-ticking.

Different Use-cases – Or A Fundamentally Different Technology?

For many, this closing of the doors and restricting access goes against the very essence of the (Satoshi) Blockchain. But there seems to be a growing understanding that as the area continues to be explored, people are most likely talking about (at least) two very different concepts. There’s an argument that the term ‘blockchain’ cannot cover both concepts.

But is a permissioned blockchain not simply a database?

In some ways, yes. But this is a database on steroids. One that contains rules about who can update the ledger and how. One that gives you conclusive evidence of when the records were changed and by whom. One that can be designed to reject any types of transactions that are deemed to be unacceptable at the outset. And one that will be reconciled globally in a fraction of the time that it currently takes and therefore can be automatically audited.

When it comes to the banking industry, it’s hard to believe that back office settlement platforms won’t move soon to decentralised ledgers. Think about the current system for a minute. Currently every bank has to spend huge amounts of time recording and continually reconciling complete records of every single transaction that takes place between them. If A sends money to B, both A and B have to record that transaction in their respective books. One transaction, two records. But what if that transaction was instead recorded automatically on one single ledger?

Clearly the financial services industry will have privacy concerns.
However, as Richard Gendall Brown explains in writing about this concept that he has called the replicated, shared ledger, these should not now be an issue. The system works because there is a single record of all transactions, a copy of which (complete, updated and protected) is held by each financial institution. Each would only be allowed to update the records that relate directly to its own dealings – and so the master record is now collectively maintained by all, allowing everyone to enjoy the benefits of a system with vastly improved efficiency.


To me, it’s clear there is a use-case for both types – and no doubt many variations in between (and beyond). Undoubtedly permissioned block chains are a significant step forwards for the financial services industry (amongst others). However, as progressive as such a change would be, it still doesn’t provide the oxygen that is so readily available within a permissionless system such as Bitcoin to nurture the truly ground-breaking innovations.

With the evidence to date, the answer to me seems to come down to this. Large financial services will inevitably focus on incremental change – because when you’re dealing with such significant sums of money and the strictures of regulation, any improvements in efficiency, no matter how small, will have massive financial benefits. Permissioned systems will therefore fill this role successfully in the shorter term.

Yet the true innovation, the moon shots and the long-term disruption must, in my view still be discovered within the Bitcoin ecosystem. It’s almost certain that the business models that will truly disrupt the existing financial sector have not yet been found. But I do believe they’re coming. So if you’re in a bank and you think that there’s even a 1% chance of Bitcoin being successful, you surely need to be thinking extremely carefully about the future and position yourself accordingly because the effects of waking up and seeingt such a day arrive are so significant.

I’ll leave you with a quote from one of the most powerful women on Wall Street, Blythe Masters:-

“How seriously should you take this? I would take it about as seriously as you should have taken the concept of the internet in the early 1990’s. It’s a big deal. And it is going to change the way that our financial world operates”.

Medical Apps and Regulatory Challenges

The growth of the internet has provided us with unparalleled access to information. Technology has enabled us to distribute this information on a scale that has never previously been possible and for a fraction of the cost.

However, as the barriers come down, we’re still learning how society reacts when individuals act upon such new and increasingly accessible sources of information. Healthcare apps are an obvious example. The growth of fitness tracking obviously has legal implications but some medical professionals argue that they also run the risk of being detrimental to your health in certain situations as well.

It seems to me that innovation will always present challenges. The question is whether finding a route around existing hurdles (such as regulations put in place to protect the public) provides a result for the world that is genuinely net positive or negative – or whether we still need protection as such innovations are trialled. The risks of making the wrong choices are clearly much greater when dealing with health matters than with other sectors but the principle, in my mind at least, remains the same. If forced to choose a ‘side’, technology itself must always be viewed as neutral and the overall effects of open innovation viewed as being more beneficial than harmful.

That’s a tougher argument to make if an app on your phone incorrectly informs you that your blood pressure is fine and you delay potentially life-saving medical check-ups as a result. But there doesn’t appear to be another route forwards if you truly want to evolve.

The Falling Price of Technology

Short post tonight as I’m involved in the Creative Currencies Chiasma event. Given the high quality discussions I’ve already had with people, I have no doubt whatsoever that I’ll have plenty to write about once the event wraps up on Thursday evening. But in the meantime, here’s a couple of quick facts.

Progress continues apace, as does the falling price of technology. It’s worth looking at some of the numbers I think, laid out as they are helpfully in this post to get a sense of perspective. For example, I found it interesting to see that:

  1. Since 1980, the price of computers has dropped 99.9%.
  2. Since 1980, the price of software has dropped 99.3% (unsurprising when you consider how much is free).
  3. Since 1980, the price of TV’s has dropped 97%.
  4. Since 2000, the price of cameras has now dropped by 75%.

Obviously there are a variety of factors that have helped to influence how costs have been reduced (or perceived to have fallen, in the case of inflation) but there’s some handy numbers there if you’re looking for a few solid examples of how innovation can drive down costs.


Fred Wilson On 2014’s Key Themes

Fred Wilson put up a couple of posts around the turn of the year which are worth taking a look through. He starts by setting down some of the key themes that he saw during the course of the year and followed it up the next day with a few thoughts on what might be on the cards in 2015.

It’s worth reading the posts in full but if you want a quick summary, the key themes in 2014 were set out as being:-

I’m interested to see a big overlap between these themes and many of my posts here. I don’t for a second claim to have anything like the insight that Fred has into what’s going on but I think most people would agree that uncovering basic general trends is not rocket science per se. Where people like Fred excel is in being able to absorb all of this information, analyse it and then actually manage to pick a crop of companies for each fund to invest in from which the global “winners’ of each sector could ultimately emerge.

I wonder what will be beside the set of bulletpoints this time next year.