Why OpenBazaar will Fail before it Succeeds

OpenBazaar is perhaps the most hotly anticipated app ever in the bitcoin space. Its promise of a truly censorship-free marketplace means it could be bigger than all the so-called "darknet" marketplaces combined and more - it could disrupt Etsy, Ebay, and other more traditional marketplaces.

So why do I think it will fail? Simple. At launch there will be no support for IP anonymity (i.e. Tor). You can't just boot up Tor when you use OpenBazaar because OpenBazaar requires you to make multiple network connections to multiple peers, and you can't simply route all those bits of traffic through Tor's obfuscation and routing protocol, which is not built for one-to-many peer connections. This means none of the darknet traffic will move to OpenBazaar, and OpenBazaar is way too obscure to seriously start chipping away at Etsy and Ebay in the near future. So when OpenBazaar launches, it will fail.

However, some brave and kind soul will patch OpenBazaar and build in anonymity support. Probably not over Tor, but rather over I2P or another anonymity protocol better suited for p2p network connections. Once this is done, OpenBazaar will quickly draw in all the sellers and buyers from the darknet markets and grow substantially. So eventually it will succeed, but that will likely not be until 2017, and it will be proclaimed a failure in between now and then.

Bitcoin's 10x Products

In two recent posts from Greylock Partner's Sarah Tavel, she argues that successful businesses don't just create a product that is 10x better than what exists. They create a 10x better product that is also cheaper. Uber is the quintessential example here - not only is it a better experience than a taxi, it's also half the cost. She argues the mistake many companies make is focusing on a 10x better product that ends up being the same cost or more expensive, and that this ends up not working.

This got me thinking about bitcoin. The real issue with bitcoin isn't that it's not cheaper than alternatives - it is cheaper. The issue is that the end products built around bitcoin aren't 10x better than what already exists. In fact much of the time products built around bitcoin are only half as good as Visa/Mastercard/ACH/wires.

Bitcoin e-commerce transactions are a great example of this. Receiving payments in bitcoin saves 2-3% for merchants and eliminates fraud and false declines (this remains true even if the merchant instantly cashes out to dollars via Coinbase/Bitpay). It is definitely cheaper than alternatives. However, is it a 10x better product for consumers? Absolutely not. Between pay-with-Paypal, credit card 1% kickbacks, airline miles, and slick UX provided by companies like Stripe, the bitcoin checkout process feels clunky by comparison, and definitely doesn't offer a 10x experience - if anything it's more like a 1/2 experience.

Bitcoin has cost-saving advantages built into it - cheaper, faster, un-censorable, and global payments. But it doesn't yet have a critical mass of 10x products.

IBM Selling Blockchains

As someone who's been involved in bitcoin since 2011, seeing this video come out of IBM is pretty mind boggling. It's a professional advertisement for bitcoin from a bluechip company. However in your head you have to replace every time they say "blockchain" with "bitcoin" for the video to make sense.

The p2p Economic Layer

Traditionally p2p network nodes have been "altruistic" - that is, Tor nodes and torrent seeders help the network not for personal gain, but because they ideologically believe in what the network stands for. This means that most p2p networks today support "sexy" causes like free speech, free information, and the like. But what about less sexy networks, like storing encrypted data (IPFS) or routing packets (Mesh networks)?

Cryptocurrency projects are emerging that add an economic layer to traditional p2p networks. Things like Filecoin, Joystream, Monestsu, and Althea are adding a monetary incentives to p2p networks for file storage, torrenting, and mesh networking, respectively. Adding a monetary layer means that some or all network nodes operate for economic gain instead of ideological beliefs. This means that less "sexy" p2p networks (like encrypted file storage, in which nodes bear non-neglible storage costs) can function even though the participants may not be ideologically motivated.

What types of monetized p2p networks will emerge that couldn't feasibly exist without economic incentives?

Decentralized Bailouts with Ethereum

The Dai credit system whitepaper is decentralized monetary policy, enforced through mathematical contracts in ethereum instead of legal contracts enforced by a central authority. In this case two tokens are issued on the ethereum blockchain, the Dai and the Makercoin. Dai is created by users locking up collateral such as bitcoin or another blockchain-based asset (using an ethereum script similar to bitcoin's OP_CHECKLOCKTIMEVERIFY). This means the value of all Dai is backed by the value of collateral obligations that are locked using an ethereum script. Thus intuitively, when the value of the underlying collateral assets goes to zero, the value of the Dai should go to zero.

However, the Dai is "insured" by a second token, the Makercoin. Makercoin programmatically increases its inflation rate to "bail out" the Dai if the Dai's value drops. Makercoin holders are incentivized because they earn interest when the Dai's value doesn't drop - they are betting on the stability of the Dai's underlying collateral assets.

Still wrapping my mind around how this works, but this is a pretty fascinating example of how ethereum contracts may decentralize complex financial products like ETFs and insurance.

Bitcoin Micropayments are Creating a Global Labor Market

Mechanical Turk (Mturk) is a global labor market which matches people looking for work with people who need labor. Each task pays anywhere from $0.01 to $100 and includes things like sorting photographs, answering survey questions, and organizing data. Anyone with an internet-connected mobile phone can participate in the Mturk economy, and given that India just became the second largest smartphone market in the world and has a number of proficient English speakers, Mturk has taken off there. But there's one massive complication: all payments occur in Amazon points, which are redeemable only through purchases on amazon.com.

Most Mturk workers stemming from India are under 35 and live in households which earn under $15k per year. These Mturk laborers are earning Amazon points and need Indian rupees - they're not exactly in an economic position to go 1-click shopping on Amazon, and many do not have bank accounts. So how do they cash out? In two really inconvenient ways. In the first method, they can buy something with relatively sticky value, like say an Ipad. They ship it to themselves and then resell it online for (nearly) the amount they bought it. They end up paying international shipping twice and lose a cut on the resell value. Plus they need to cash out in lump sums, saving up $100-$1000 in Amazon points at a time.

The second method is more complicated but saves them more money. They use a service like purse.io - which allows them to purchase goods on Amazon for westerners, in exchange for bitcoin. This way they don't have to worry about international shipping and reselling - they ship the item straight to someone who wants to pay in bitcoin. The western buyer is incentivized to pay in bitcoin because of discounts on goods Purse offers. From there the person in India cashes out the bitcoin for rupees, ostensibly through Local Bitcoins. So the person in India still has to give out a pretty decent cut (first to Purse, and second when they get burned on the street bitcoin/rupee exchange rate). However the percentage they give up is still smaller than the cost of Ipad reshipping.

It's pretty amazing that Indians are using Mturk at all given how difficult it is to get paid. What if there was a Mturk-style labor market based entirely on bitcoin payments? This allows people to get paid on a daily basis instead of in lump sums (once the lightning network is live, they'll get paid on a job-by-job penny-by-penny basis), and their only fee is on the street exchange rate to rupees. Not only would this solve a big problem for the Mturk laborers in India, but it would likely expand the number of total workers who use the service. (This effect is similar to the way that Uber hasn't just eaten the taxi industry's lunch, but actually vastly expanded the size of the taxi market because their product is so much simpler to use.) 

A couple cool projects working on a bitcoin-based Mturk are Jeff Garzik's implementation for the 21 computer, and the Rein Project. There is potential here to build a global work & labor market over the internet that is facilitated through bitcoin and much, much larger than Mechanical Turk is today.