# Allodial Title: Real Assets On-Chain/ January 15, 2019by Chris Walker

Blockchains allow us to exercise ownership of real-world assets in a form that has not been common for a millennium. This form of ownership is called allodial title, and it works where traditional legal title does not.

Over the last decade, the dominant blockchain narrative has moved from digital cash to digital gold to code as contract law. Bitcoin has proven its mettle as an (extremely volatile) global store of value. Various companies have begun anchoring real-world assets like company stocks to the Ethereum blockchain using a combination of blockchain smart contracts and traditional legal contracts.

### INTRINSIC VS. EXTRINSIC ASSETS

In the blockchain economy, there are two types of assets: (1) intrinsic assets which maintain value because of the protocol and (2) extrinsic assets which have value absent the protocol.

First, public blockchain protocols such as Bitcoin and Grin define an intrinsic asset (a coin or token) that may be transferred between users of a blockchain network. The tokens are digital bearer instruments analogous to cash or gold coins, but possession is determined by knowledge of a secret cryptographic key rather than physical access. The properties of these tokens, such as BTC for the Bitcoin network, are entirely determined by the rules of the blockchain protocol and the history of the network. Once one decides to participate in the Bitcoin network by running a full node, one need not trust any third parties. A blockchain coin has value (today 1 BTC = 3600 USD) because participants in the blockchain network believe it to be a good game of money, recognizing that other participants are likely to value these properties as well.

Second, tokens referring to extrinsic assets can be placed on a blockchain, commonly Ethereum. Like intrinsic assets, the token itself is controlled by the blockchain protocol. But unlike intrinsic assets, the extrinsic asset the token represents is not. A blockchain token may represent a digital asset like a Fortnite skin, a physical asset like a Walmart avocado, or a financial asset like a Tesla share. The ownership of these real-world assets cannot be directly verified or controlled by any blockchain, so these references to extrinsic assets require trust in the controller of the asset which in most cases is one or more third parties. Consider a blockchain token that represents ownership of a gold bar. If the actual gold bar is stolen, the blockchain token’s reference has lost its referent, even though nothing intrinsic to the blockchain has changed. Extrinsic assets can only be anchored to a blockchain token by trusting something outside the blockchain to refer back to the blockchain token to adjudicate claims to the extrinsic asset.

### LEGAL ANCHORS

Extrinsic assets can be anchored to a blockchain through legal mechanisms, like articles of incorporation and investment agreements that specify that a particular blockchain token represents a share of the company analogous to a paper stock certificate. For example, the 22X fund offered investors ownership in a group of startups as ERC20 tokens on Ethereum. Regardless of whether a company share is represented on Ethereum or in a Carta capitalization table, its value depends on the actions of the company, the company’s willingness to fulfill its obligations to investors, and a court’s ability to uphold investors’ rights. Thus, trust must be extended to such third parties. While this approach to anchoring assets to a blockchain may be effective for large holdings of company shares, it may not be suitable where legal expenses would outweigh an asset’s value. Many are skeptical that this is the future of on-chain real-world assets.

﻿Thankfully, legal anchors are not the only option. One can enforce property rights by extrajudicial means.

### ALLODIUM & ALLODIAL TITLE

In Commentaries on the Law of England, Sir William Blackstone contrasts two forms of title to assets that were first articulated in the Medieval period:

The true meaning of the word fee (feodum) is the same with that of feud or fief, and in its original sense it is taken in contradistinction to allodium; which latter the writers on this subject define to be every man’s own land, which he possesseth merely in his own right, without owing any rent or service to any superior.

By this standard, a property title in Montana is feodum because it is qualified, i.e. secured and taxed by the Montana and United States governments. In contrast, BTC is allodium because possession is defined by knowledge of a private key alone: there is no superior party that can extract rent or seize the asset. Several clarifications on allodial title are important.

First, allodial title is not a measure of protection. An enemy army may storm a poorly defended allodial castle, or an attacker with a 5 wrench may torture someone for his Bitcoin key. The distinction between feodum and allodium rests simply on whether one possesses a thing by defending it with one’s own force or by trusting another who defends it with his or her force. Second, allodial title is not a legal title in the same sense as a property title or vehicle title. It is a more fundamental form of ownership. A typical title to Montana real estate is a declaration that one acknowledges the jurisdictions of Montana and the United States and that within this game, the rules, referees, and pieces have aligned such that one holds the property. An allodial title to the same property is a declaration is that one is playing a different game, and that the jurisdictions of Montana and the United States are irrelevant. Of course, those who make claims like these may quickly lose their allodium as more powerful parties reassert jurisdiction. Several states including Nevada actually toyed with a form of allodial title. But these titles were allodial in name only, as they included limitations and were administered by the states rather than the property owners who held this title. Those who assert allodial title do not need to reference supporting legal codes. They have armaments and fortifications. Finally, Blackstone refers to allodial title only in the context of real property rather than personal property. Applying the concept of allodium to personal property may be a legally tenuous step. But, in the spirit of allodial title, we will proceed anyway. In the fragmented Medieval landscape, allodial claims were common (perhaps in the simple “your property or your life” form), and today the clearest example of an allodial title is the defended territory of a sovereign state. But until very recently, there were no other good examples of allodial title because defense against adversaries such as sovereign states is a monumental task. But allodial title is making a comeback from an unexpected direction. ### ALLODIUM IN PRACTICE The iPhone is allodium. Without the correct key in the form of a passphrase, face, or fingerprint, an iPhone can protect its contents from almost every adversary, including petty thieves and hostile states. The photos, messages, and contacts inside an iPhone are part of another realm where cryptography is the law of the land. Unlike other attempts to assert allodial title over the past millennium, these defenses may hold. What makes this possible? Apple has taken a strong pro-privacy stand and engineered their devices carefully. But the iPhone is part of a broader trend. Modern allodium looks like sophisticated, affordable, and practically impenetrable electronic agents whose loyalty is determined by possession of a cryptographic key. This form of allodial title has only become practical recently due to the digitization of controls, commodification of strong cryptography, and pervasive internet connectivity. To attackers, stolen allodium is only a blank and lifeless form, with the real value protected by unknowable and unbreakable enchantments. Perhaps the allodium can be destroyed, but it cannot be used. The Bird scooter is another example of allodial title in practice. Forget the unicorn valuations and frustrated cities. The most interesting thing about these scooters is how they are defended. The scooters can be trivially stolen, and they have no title document or VIN. Prosecuting thieves and recovering stolen scooters would be laughably uneconomical (the manufacturing cost is perhaps250). Bird cannot assert legal title to these scooters practically or economically.

Instead, the real defense is that the wheels lock and it’s hard to unlock them until one pays. The scooters’ allegiance is to Bird, but the scooters fend off would-be assailants themselves, without resorting to Bird, the courts, or any other superior power: that is the essence of allodium.

Of course, these defenses are penetrable. Within a few days of their release in San Francisco, enterprising attackers were gutting scooters in search of parts or free rides, and online sellers now offer conversion kits that replace the scooters’ original circuitry with more compliant alternatives. Effective exploits probably exist for all allodium, including iPhones. But the field of battle is always changing, and the real question is whether modern allodium is sufficiently strong for the task at hand.

Our most important communication devices are already allodium. The electronic systems that make allodium possible will likely get less expensive and more secure. Allodial title will be used for larger and more important items, perhaps only as a first line of defense in the beginning, with a fallback to legal mechanisms when necessary. Imagine a house where every single functional element is allodium, variously loyal to the tenant, landlord, and utility companies. This future is already here, if unevenly distributed.

### Put a Blockchain on It

What happens when the rights to allodium are placed on a public blockchain?

Modern allodium can be the trusted agent outside a blockchain that automatically refers back to a blockchain token to adjudicate claims to the allodial asset itself. If access rights to an allodial device are determined by ownership of a specific blockchain token, then market participants can buy or sell the token with confidence that it represents a real-world asset. The allodium will change loyalties as the ownership of the blockchain token changes.

Some trust is necessary: one must assume that a particular allodium has not been hacked or compromised by another party, and is truthfully representing its capabilities and loyalties. The on-chain history of the allodium may provide evidence that these assumptions are sound. If one extends this limited trust to the allodium itself, a blockchain ecosystem enables one to transact in ownership rights to the allodium with other wholly untrusted parties using standardized mechanisms: smart contracts, exchanges, wallets, block explorers, etc.

This will increase the liquidity and automation of allodium transactions. Imagine Bird scooters without Bird, autonomous delivery robots, or communications satellites whose resources may be anonymously purchased via the blockchain. It is impossible to be certain where this liquidity and automation will be most valued, but given the colorful history of cryptocurrencies to date, we can be confident that many unexpected applications will be found and tried.

Final Notes

• Thanks to Lawson Baker, Anthony Sanderson, and TJ Kastning for reviewing this post.

• I discovered that Nick Szabo, as usual, thought of many of these considerations many years ago, in this case all the way back in 1997. He has a great article called Formalizing and Securing Relationships on Public Networks that includes an example of a vending machine as embodied smart contract (and, as it turns out, allodial title).

# Smart contracts and real assets/ July 7, 2018by Chris Walker

This post is a response to a Medium article by Jimmy Song called "The Truth About Smart Contracts."

Something about that article didn't seem quite right, so I'm using this post to define my thoughts. First, let's summarize Song's argument. Song is an experienced Bitcoin developer who has worked at Paxos and Armory, so I want to take his perspective seriously.

### The good

Smart contracts are executed by many computers running the same known algorithm and checking each other's calculation: no third party needs to be trusted to perform the calculation correctly. Trustless contract execution is a new, powerful property.

Unfortunately, smart contracts aren't intelligent, they are simply encoded as precise software instructions. This means they suffer from the same problems as regular computer code: inevitable bugs and absolute blindness to the entire universe except for their specified inputs.

In particular, Ethereum's smart contract language, Solidity, is Turing-complete: it's a computational trump card. Smart contracts don't need this power (Bitcoin's Script is not Turing-complete), and this extra power is hard to wield correctly, caused many expensive bugs.

One of the main attractions of smart contracts is the idea that "code is law" -- with sufficiently complex smart contracts, the need for legal arbitration disappears because the code runs itself. Unfortunately, this idea hasn't always been true: the Ethereum network has already rolled back transactions that followed the rules of the code because they were seen as bugs and theft. Rather than trusting the smart contract, we ended up falling back on human intervention.

### The ugly

Smart contracts work well for digital bearer instruments, where possession equals ownership, and parties have the ability to swap digital assets as they desire.

But most assets, e.g. real estate, stocks, and gold bullion, are made of atoms and law. Those can't be wrapped up nicely into digital bearer instruments within a blockchain (they aren't code), so their representation within a blockchain is merely a sign of the asset, not the thing itself. Blockchains are blind to the external world, so they need to resort to third parties to describe the state of the actual assets: the gold really is in this vault, the house really is mine to sell. Once smart contracts resort to trusted third parties, they lose their central advantage: trustlessness.

### starting with why

I generally agree with the facts as presented by Song, but have a different view on valid motivations for those exploring blockchains and smart contracts.

Blockchain was started by people like Hal Finney, Wei Dai, Nick Szabo, and the mysterious Satoshi Nakamoto, who shared an obsession with decentralization, privacy, trust minimization, and sound money. I suspect that only this obsession could create a system as brilliant and simple as Bitcoin, and I can personally relate to these interests ever since I read The Creature from Jekyll Island describing the creation of the Federal Reserve in 1913.

Trusted third parties are security holes: on some level fiat currency is just paper.

They have a valid point. But the philosophy that created blockchain may not be the philosophy that capitalizes on it most effectively. The absolutism of trust minimization obscures pragmatic opportunities.

Trust / Trustless. Liquid / Illiquid. Automated / Manual. These aren't binary properties: they exist on a continuum. One extreme is a smart contract transferring digital bearer instruments in cross-chain atomic swaps. Awesome! Another extreme is a loan founded on nothing more than a handshake and a promise. Great!

Most economic activity takes place in the vast space in between extremes, and I see tremendous potential in distributed ledgers and smart contracts because they let us explore regions of this space that were previously inaccessible.

Consider a smart contract representing ownership of a real world asset like a small Delaware C Corporation. Delaware altered the Delaware General Corporation Law to support stock ledgers held on the blockchain, so this is legally reasonable.

Clearly this asset requires trust in the directors of the company no matter how it is recorded: if they abuse that trust, the company will fail and the asset will probably be worthless. Also, it's readily apparent that the digital record (or stock certificate) isn't the final authority: if the company loses a Delaware court case, a judge may ultimately decide the fate of the asset.

Despite these limitations, this digital asset may still be more liquid and more trustworthy than alternative options, like a stock ledger held in an excel document. It wins, not by eliminating law and trust, but by streamlining and minimizing them. If the smart contract follows a standard pattern, it requires less auditing by potential investors. If it covers basic corporate transactions, fewer aspects of running a corporation need legal or accounting input.

A mature blockchain ecosystem will make ownership of real assets much better by reducing friction. Want to sell your stock to another investor who meets the smart-contract standards? Of course! Collect dividends automatically? Certainly! Use the stock as collateral for a house loan? Why not?

All of these features require some amount of trust in third parties. But that's OK. So does the rest of life.

# Architectural Musings/ June 30, 2018by Chris Walker

A recent trip through Italy made me consider why European houses can be so much more charming than US houses, and how easily we could replicate some of that feeling here.

Part of the attraction can certainly be explained by the excitement of exploration,  selection bias (buildings that last tend to be appreciated, and tourists focus on the best of the appreciated buildings), and the shape of cities themselves (organic growth vs 20th century central planning for automobiles), but I believe that the buildings themselves have significant differences.

Great European buildings have a glorious texture. They're made of materials that wear in rather than out, so they look better over time, even when that time is measured in centuries. We've all heard the term patina, but what does it really mean here? I think there are a couple of key elements to graceful architectural aging.

### Beautiful oxidation

Sun, rain, mildew, ice, sand, and above all oxygen attack every structure. Graceful materials have either reached a fairly stable energetic state, like limestone, or the process of reaching that state is slow (e.g. due to protective passivation) and produces byproducts in pleasing natural hues, like bronze and zinc.

Paris is famous for its zinc roofs, which weather into a gorgeous patina and can last for centuries. Photo by VB Fetter, Unsplash.

If the structure includes thin surface layers, it's critical that the material underneath also ages gracefully, or a few decades will reveal that the building's core is unsound when the outer layer is inevitably compromised.

Everything here starts oxidized or looks great even while crumbling to pieces (yes, those bricks could be repointed).

Beams are necessary for most structures, but economical wooden beams seem to sag within a few decades, and drooping horizontal lines stand out like a sore thumb.

One of Salvador Dali's lesser known works, "The persistence of RM-30 zoning."

Many classical buildings use arches simply because the arch was the only way to build cost-effectively at a large scale, but a great side effect is longevity. Within a single building, compare these openings: the arch isn't noticeably sagging.

### Complex texture

I think that it can sometimes be a mistake to create "texture deserts" at various architectural scales. In other words, a building may have features at the 1 meter scale and at the 0.001 meter scale, but very little in between. Well, as the ground shifts and the building ages, nature has a habit of adding texture at every scale. If the added texture sticks out as the sole signal on that textural wavelength, we see it as a glaring flaw. But if it blends into a broad spectrum of textures, we see beauty.

For example, compare stonework with large, smooth concrete faces. Both will crack over time, but only one will look good afterwards.

### Honesty of form

This is the most important thing to me. Through artifice, we can fake just about any material or form, but it's a mistake to think that this artifice will stand the test of time. There's a sort of uncanny valley between the thing itself and something else paying homage to the original. For example, the Corinthian column pays homage to the acanthus leaf, but nobody is confused that the column is pretending to be a plant. Great.

Source: JD Mason, Unsplash

Source: Acanthus article, Wikipedia

What is not so great is a design that pays homage to a predecessor by slavishly imitating its nonessential parts. Examples are limitless, but here's are a few I saw today.

Stone facade and stucco over plywood? Who, me? Never. This is a proper stone house. You can see that those two stone pillars are holding up half of the house (we ran the numbers and were able to make one a little thinner), and the mortar is so strong that we didn't even need to build an arch of voussoirs above the garage door. Progress!

Colonial brick house. Lots of curb appeal. So long as you don't walk more than 10 feet down said curb. Good luck using those storm shutters! This house is as precisely artificial as the grass out front.

### ApplicAtion

This is all a nice theory until somebody has to write a check to build a building; when that occurs, economics tends to stifle everyone's architectural preferences. But I wonder: are there cost-effective ways to achieve the goals above when building a house today? Perhaps a combination of new construction techniques like 3D printing and traditional materials will get us closer. I am toying with the idea of 3D printing a masonry house some time in the next decade. Are building codes close enough to make this feasible? I don't yet know. Stay tuned. I'll end with a shot of a house I found today that was genuinely charming.

# Singapore/ May 10, 2018by Chris Walker

Chris: I spent last week setting up a factory in Singapore with Skyforge, the 3D printing platform that's now owned and developed by Jabil.

Initial reaction to Singapore: far too humid. Felt like a giant immaculate shopping mall. Still wonderful to visit, esp. the gardens by the bay.

After more thought, Singapore strikes me as a modern Venice: a surprisingly powerful city-state founded on excellence in international trade. Besides their physical location, I am guessing that the Commonwealth / Christian heritage has a lot to do with their success. The government seems really intrusive even by Californian standards, but my concern about government overreach is proportional to the size of the nation and the distance between politicians and electorate. The factory I visited was extremely clean and locals were friendly, proud of their jobs, and very competent.

# Peak SaaS/ July 22, 2017by Chris Walker

In the software world, SaaS is the air we breathe. It's everywhere, but so are the huge, unresolved problems that come with it. We need solutions, and cryptoeconomics may help us build them.

We all know and love SaaS: the \$1-100 per month web apps that do useful things like send emails, stream movies, back up our photos, manage schedules, and so on.

Sometimes SaaS is exactly what customers need. But normally businesses choose to build SaaS because they know it lets them:

• Earn revenue proportional to value by charging for continued usage
• Protect their digital assets from theft by placing them on a secure server
• Simplify service delivery by using a standard browser interface

Those are great things, and without any better alternatives, we will continue to build and buy SaaS (I've done both). Unfortunately, most traditional SaaS products come with serious side effects.

### Fragmentation

First, SaaS fragments data for users. SaaS protects digital assets like source code and user data by placing them on a secure server. As a result, each SaaS product tends to create its own data fiefdom, with its own data types and storage technologies. There is no general language which these fiefdoms use to securely and intelligently communicate with each other. That means every SaaS product has incomplete, redundant, and inconsistent versions of customer data.

Sometimes we attempt to fix this by gluing together different SaaS products with software integrations, but this is hard to do correctly. Even worse, the complexity of integrating multiple SaaS products, each with its own quirky API and data types, increases rapidly as the number of systems increases. The high cost of each integration means that most potential integrations are economically unviable.

We've grown accustomed to the balkanization of our own data because SaaS has taught us that there isn't any other way. For example, my wife and I just moved to Palo Alto, and now I need to let the world know that we've changed addresses. Of course the Post Office and Amazon know already, but how many other records in other databases do I need to change? 100? Why isn't it simple for me to update my own address? The truth is that I cannot cleanly update my address, and SaaS creates the same class of problem for every single bit of your data that it stores.

It is self-evident that every person and every business would appreciate a good unified interface to their important data and functionality. But traditional SaaS can't deliver the goods.

### Trust

Second, SaaS centralizes data for bad actors, reducing the privacy and security of users. Somewhere in the unread Terms & Conditions of a SaaS product there generally lies a clause like this:

In exchange for your momentary and fleeting access to the Service, you agree to give us all of your data. We'll keep it forever and might sell it to anyone. How would you know, anyway? We promise to keep it secure. Trust us. You don't really have a choice.

The problem is bigger than the perverse incentive for a SaaS provider to abuse customer trust by taking customer data and quietly using it in ways that customers would not expect or approve. SaaS eliminates any structural guarantee that customer data is fairly secure and private. It's easy for me to audit access to the hard drive I am holding in my hand. But when I put my data on your server, it becomes your data, and I cannot audit what you do with it. When a million other customers put their data on the same server, it becomes an attractive target for many kinds of bad actors. We've created an ecosystem where a small number of keys unlock the doors to a lot of valuable data. Is it any wonder that bad actors get the keys?

### complexity

Finally, SaaS adds huge amounts of non-essential complexity to support the SaaS business model. Consider a typical SaaS product like a to-do list and scheduler for small businesses. Let's say the target market just needs a good way to make lists or schedules and share them with co-workers. The core functionality to do this might be 2500 lines of code that took a month for a single person to build. But to bring this to market as SaaS, this code must be transformed into a startup replete with:

• People: marketing, sales, operations, management
• Costs: personnel, facilities, legal, contractors
• Systems: cap tables, payment processors, unemployment tax reporting

Maybe the startup will do well and result in a billion-dollar acquisition. More likely, it will self-immolate and in a year or two the customers will move on to another solution. Either way, turning a relatively simple piece of code into a startup increases the stakes and complexity by orders of magnitude.

But fifty years ago, the customer might have solved their scheduling problem with a chalkboard. Even if the SaaS solution is ten times as convenient for the customer, it's easily a thousand times more complex and inflexible than just using a chalkboard. Is this tradeoff truly necessary? Perhaps we are missing big opportunities to create economic value with software defined in units smaller than the SaaS startup.

### KICKING THE SAAS HABIT

Traditional SaaS is complex, balkanized, and filled with security and privacy pitfalls, but there is reason to hope. Cryptoeconomics, a novel field of study that combines adversarial game theory, modern encryption, and networks of connected machines, is a promising avenue to explore.

Using the tools of cryptoeconomics like blockchains, peer-to-peer routing, and public-key encryption, we can create protocols that unify data for users and protect digital assets from bad actors. Software providers can use the same protocols to build simpler, more trustworthy units of software functionality, securely deliver them to customers, and receive compensation proportional to the value they create. That's all we need to start building a different kind of software ecosystem.

In short, I suspect that SaaS is not the end game for software.

# Introducing SST/ June 27, 2017by Chris Walker

Over the past few years, I've talked with a number of executives who run large software-driven businesses or heavily integrate software into their operations. A particular phrase has come up in almost every conversation, something both very important to the success of their business and difficult to achieve.

They want a single source of truth for their important business data: a unified, complete record of everything that matters to keep their people on the same page.

I don't think that this value proposition only appeals to the largest enterprises: everyone who uses digital systems spends a lot of time or money struggling to keep emails, computers, file systems, databases, websites, spreadsheets, etc. on the same page.

This sounds like an insurmountable problem: by necessity, a company uses different software products for different domains, each with its own proprietary format or database: email, cloud storage, and an industry-specific alphabet soup of platforms like CRM, CAD, PLM, MES, ERP, PCM, etc.

But I think we can solve it.

With a group of friends, I've been working on SST, an open-source universal data store enabling a single source of truth for collaborating on important data.

More specifically, SST enables participating nodes to create a universal key-value store which reaches consensus by blockchain, verifies storage transactions using commutative hashing, and secures values by public-key cryptography, forming a decentralized, highly available storage network.

SST aims to create a simple, low-cost way to store and share data that always works. Ethernet always works for transmitting data, but we want a similarly ubiquitous and trustworthy system for storing data.

SST enables software to do things that sound a bit like magic today:

• Traceability: want to track the complete history of every document? Done.
• Automatic updates: there's no need to attach the latest version of that presentation to the email chain. The intended recipients already have it -- in fact, you don't need the email either.
• Pre-integration: want your latest sales to be reflected on the company's website? The website and the CRM already talk the same language -- it's probably a one-line code change.
• Cross-company collaboration: need to share select data with business partners? SST keeps everyone on the same page without vendor lock-in or giving one company undue leverage over another.
• Simplicity: can you deploy software the same way you cache content, the same way you sync databases, the same way you back up critical assets, the same way you send emails? We think you can.

This is a different way of thinking about software, so we're also building a proof-of-concept SST client called Concorde to demonstrate these ideas in action.

The project is just beginning: the defining whitepaper, code base, and reference client all exist and are privately available, but much work remains.

If this sounds interesting to you, then please contact me. We already have a community of folks interested in tackling this problem, but we want your input on potential applications for this technology. Our goal is to build SST into a simple, general purpose, open source protocol that solves data collaboration problems in every industry.

# Skyforge Acquisition/ July 7, 2016by Chris Walker

I'm very excited to share that Skyforge, the 3D printing software platform I co-founded, was just acquired by Jabil, Inc.

Some events in life are unplanned gifts: this is definitely one of them.

When John Feusi and I started our company, Element Robot, we thought Skyforge was a 3D print vending machine, but within a few pivots Skyforge had turned into the cloud-based 3D printing software platform we have today. But every pivot was based on our vision that 3D printing would change the way the world manufactured goods, and that we could help make that happen.

The past few years have been tumultuous for the 3D printing industry, to put it nicely. I'll save details of the Skyforge story for later, but many companies lost sight of the realistic potential of 3D printing, instead focusing on a limited consumer market or diving into delusions of 3D printed pizza.

That's precisely why the opportunity to use Skyforge at Jabil is so exciting.

Jabil has 180,000 employees focused on world-class manufacturing for the brands you know and love, like Apple. At this level of performance, Jabil has zero tolerance for fantasy. They've emerged as a thought leader in adopting 3D printing at scale -- a vision I can appreciate.

We initially discussed setting up Jabil as a pilot customer, but it became clear that it made better business sense for them to own the platform and use their manufacturing expertise to direct the platform's future growth. This is pretty much the ideal use case for Skyforge: Jabil has close relationships with some of the most innovative 3D printer OEMs and a global scale.

Every startup takes a tremendous amount of teamwork, so a few thank-you's are in order:

• Our customers, who took a chance on a new team, tolerated software bugs, provided feedback that hugely improved Skyforge, and 3D printed lots of unexpected things.
• Jonathan Staab, our first "employee" / co-owner and great software developer, who is now hooked on startups.
• Our investors, board members, and advisors: Andrew Crapuchettes, David Miller, Timothy van den Broek, and Mike Stoltzfus.
• Our competitors and friends in the 3D printing space: Tom Bielecki, John Dogru, Dylan Reid, Benoit Soete, and more. According to René Girard, competitors are probably more like mirrors than enemies, and I have a lot of respect for anyone who's dedicated time and energy to advancing 3D printing.
• Independent contractors, tolerant friends, sages, fellow entrepreneurs -- Josh Hartung, James Heilman, TJ Kastning, Joel Lohrmeyer, James Prado, JT Grauke, Nick Smoot, Ben Nieuwsma, Andrew Chorlton, Joe Harby, Mike Meehan.
• Most importantly, my family -- who put up with decreased time and increased stress -- and especially Kate, my amazing wife, who met me at the beginning of this journey and has had the patience to see it through until the end.

Thank you.

I can't provide too much detail on the upcoming changes, but we're relocating to San Jose, CA, and within a year we'll be deploying Skyforge at manufacturing facilities across the globe ... which is what I've wanted to do since day one.

# Arches National Park/ April 22, 2016by Chris Walker

Chris: James and I made a quick trip to Moab and the surrounding area this spring, and I just got back the developed film. The light meter on this camera has never worked, so I have to either use an iPhone light meter app or estimate it myself. So a few are a bit off.

The high desert is unexpectedly beautiful, with a huge variety of flora. You can camp inside the park in certain areas, so we found a hidden wash and set up camp under sandstone cliffs. The park was great, but I found the nearby Corona Arch area more exciting: you can see a powered paraglider flying through Corona Arch near the end of the photos.

It turns out that arches don't generally erode in the middle, where you might expect, but at the inside of the base on each side, where I think the compressive stress is highest. That's not always the case though: the very thin Landscape Arch lost a huge section near the middle in 1991 due to rain soaking into the sandstone and adding weight. Finally, if you are standing next to an arch and flying a drone, and you tell the drone to home to your location by GPS, bad things will probably happen (wasn't us).

# NVIDIA GTC 2016/ April 11, 2016by Chris Walker

Chris: I spent last week in San Francisco on a short project that included a trip to NVIDIA's technology conference on GPUs and machine learning. A few quick takeaways:

1. Potrero Hill is beautiful. I can't figure out why the Mission is considered more hip.
2. NVIDIA has brilliantly cornered this market. When you need machine learning hardware, you buy NVIDIA, without exception. I'd estimate they're 2-5 years ahead of the competition (Intel, AMD, ATI?), given their DGX-1 machine learning supercomputer and other feats. I'm really curious what signals they saw that justified going all-in (to the tune of multiple billions of dollars) on machine learning 5-10 years ago.
3. It makes sense to go all-in on machine learning. Most people are simply unaware of how fast this field is progressing, and even the experts are continually being surprised by the pace of innovation. For instance, NVIDIA's DaveNet is a deep neural net that can drive a car (fairly well, I believe). This isn't a collection of human-designed classic robotics components like PID controllers, inverse kinematics, computer vision modules, pedestrian detectors, etc. It's a single network that learned how to drive.

# Anglosphere Power Ties/ March 21, 2016by Chris Walker

Chris: What does a presidential tie look like? I think Trump's ties are awful, but is that simply a reflection of my abhorrence for the man? Can you select a world leader based on tie alone? Let's find out -- but be warned, you won't be able to look at certain politicians the same way.

(There are larger issues at stake here, and I think this very frank video by Marco Rubio is a decent summary of current US problems. But this post isn't about politics, so on to the ties!)

## world's biggest diplomatic necktie analysis?

I've taken the top google image results for prominent Anglosphere political figures and cropped the images that include neck wear (or the lack thereof). To keep things tractable, I've skipped Australia, Ireland, and New Zealand.  Our contestants:

• US Presidents: Barack Obama, George W. Bush, Bill Clinton
• US Candidates: Donald Trump, Bernie Sanders (Hillary Clinton doesn't wear ties)
• UK Prime Ministers: David Cameron, Gordon Brown, Tony Blair
• Canadian Prime Ministers: Justin Trudeau, Stephen Harper, Paul Martin

### Guess the politician

Hint: they're each from a different person. By the end of the article, this will be a cinch (heh).

## The Anglosphere Collection

### Barack Obama

Sharp, with a variety of styles, knots, and dimples. Obama's ties show sprezzatura: tie dimples are crisp if present, but sometimes askew. He also wears a fair number of grey ties. Obama consistently wears a US flag pin after its absence caused a political scuffle (I kid you not) during his 2008 campaign.

### Bernie Sanders

Consistently casual. About five ties worn with low-key outfits. Sanders only wears button-down collars, and prefers solid blue or blue pinstripes to white. Check out his next interview, and you'll probably find him wearing one of the combinations below.

Sanders also always wears a gold pin, which is apparently issued to congressmen as some sort of low-key ID badge. If he were to get the Democratic nomination, this would likely become a serious issue during the general election. But I think we can all grant this: Sanders' dress is consistent, and it matches his unassuming personality and blue-collar message.

Edit: the week after I wrote this, Sanders started wearing shirts without a button-down collar. Coincidence?

### Bill Clinton

One cool cat. With the widest variety of bold colors, warm colors, patterns, and knot styles, Clinton definitely has the sharpest look of anyone on this list. He generally prefers patterns to solids, has relatively few dimples, possibly due to a larger and looser knot, and isn't afraid to wear really adventurous patterns. He eschews the US flag pin... and gets away with it.

### David Cameron

Predictable, mild-mannered, and friendly. Or at least that's what his ties say: I know very little of the man. I think most of these are the same ties worn on different occasions. He's tying the same small asymmetric knot every time, although I can't tell if it's a four-in-hand: the ties consistently veer to Cameron's left.

### Donald Trump

Aggressive, garish, and hardly self-aware. Trump's ties are frankly hideous. The knot has to be a Windsor or Pratt, but it's pulled so excessively tight that it wrinkles over itself and the tie comes boiling out underneath unevenly. Most ties have a dimple or they don't. With Trump, the only thing that's clear is that the tie is suffering. These knots look a bit like knotted small intestine (told you that you wouldn't be able to look at these politicians the same way). Maybe this style should be called the Trumpsor. At least the white collar is on-brand (a friend notes the sheen of these ties was probably on point in the '90s).

### George W. Bush

Humble confidence. Do you need to buy Bush a Christmas present? If so, a red tie with light spots will do nicely. These are power ties all right, but there's nothing about them that's ostentatious, and the patterns tend to be fairly small and subtle. I don't see any clear "tell" that a tie belongs to Bush, as he uses a variety of knots pretty nicely. I think he favors the Windsor, but I see a few in there that are narrow or asymmetric.

### Gordon Brown

Disheveled. That's unfortunately the single word that comes to mind. Brown's ties have no dimple, because they're simply too loose to hold a dimple in place. That looseness also causes the knots to sag or slip to the side. The untidiness extends to his collar points, which have rolled over in a few cases. The knots steal the show, and not in a good way, but at least his tie selection is more interesting than Cameron's.

### Justin Trudeau

What!? I had to look up some back story here to understand what's going on. He's the only contestant with multiple top-30 photos that are shirtless, the only contestant with wingtip jacket lapels, the only candidate to frequently unbutton two buttons, and the only candidate with pinstripes on his jacket.

So he's got panache. Something about the full images reminds me of Christian Bale as Bruce Wayne. Trudeau is the son of Pierre Trudeau, a previous prime minister, and successfully boxed against a conservative senator a few years ago. Well, that's one way to solve problems.

The ties are bold, flowery, and friendly, and the knots are generally really large. That is, if a shirt is even present. I can't see any UK politicians pulling this stunt.

### Paul Martin

Low-key people's man. The only other contestant with button-down collars, Martin can be differentiated from Sanders by the lack of a lapel pin, the reliance on solid colored shirts, and the infrequency of a dimple. Otherwise the task would be impossible, as both Martin and Sanders have very similar tastes in ties: red and blue with small light patterns, generally set on point.

### Stephen Harper

Normal. A bit like Bush, nothing about Harper's ties really strikes me as odd. I see a lot of Windsors, a few ties that are a bit loose, and a decent variety in patterns. You'll note the two red remembrance day poppies. Inspired by In Flanders Fields, a war poem on WWI, these are pretty common to see on Canadian and UK politicians.The photo with white flecks is an outdoor photo of Harper -- the only photo in the entire list including any kind of precipitation.

### Tony Blair

Dignified. Blair makes adventurous tie selections, with a few ties lighter than the shirt he's wearing. I don't know that anyone can make that look good. But in general these ties are interesting and well tied. There's a strong contrast, in that Blair's tie knots are the narrowest and longest on this list, but he also uses very tall, wide collars. The other UK politicians favor tall collars as well, but Blair's combination of collar and tie is uniquely aristocratic. This probably matches the position well -- or at least far better than Brown's ties.

## Final thoughts

There are a few things you simply won't see these politicians wearing frequently:

• A shirt and tie with no jacket. Unless you're aiming to look like a groomsman in a wedding, it's probably good to follow this rule yourself.
• Trick knots like the eldredge or trinity. Just say no.
• Bowties. Dr. Who would differ, but the data is clear. Bowties are not cool for world leaders. Time Lords and professors may get a pass.
• Bold shirt patterns: they drop the formality really fast and are hard to pair with a tie.
• Novel tie patterns. Only Clinton gets close on this one. Everyone else sticks to solids, stripes, or small spots, checks, or floral patterns.
• Odd clothing colors. There's very little green or brown. Perhaps these colors are more appropriate for park rangers, military uniforms, professors, and biology conferences.

Finally, my favorite tie, photo, and Anglosphere leader:

P.S. Answers for the first tie list: Barack Obama, Bernie Sanders, Bill Clinton, David Cameron, Donald Trump, Gordon Brown, Justin Trudeau, Tony Blair.

# Functional Universes: No Man's Sky/ February 19, 2016by Chris Walker

Chris: The No Man's Sky video game is brilliant.

Most games are built by designers that hand craft characters, maps, and entire worlds. What else could they do?

In No Man's Sky, the designers create a set of algorithms and set them loose. The algorithms create a universe populated with planets and creatures new to everyone, including the designers themselves. This universe is huge, with approximately 18 billion billion planets.

That's an impossibly large amount of data to store, right?

No.

Just about everything in these worlds can be created by a deterministic calculation from an initial seed value and the algorithms. A world isn't "calculated" until someone visits it. And when they leave, the data is thrown away. If they return, it'll be there again, just as they remembered it (with the caveat that time can be an input to these functions, so things can move and change). This universe is made of functions, not data.

Can a small number of functions generate a universe of complexity? Yes -- but more on that in another post.

I'm excited for this game, but also concerned. First, because everything is a deterministic generation from a fixed seed, it's mostly impossible for players to leave their mark on the universe. Yes, they can log findings in a central repository and name discoveries, but "sandbox" games like Minecraft thrive because they enable creation, not just categorization.

Second, from public demos, it looks like the game AI (an entirely different species than deep learning AI) is pretty rudimentary. Since the likelihood of running into another human on one of these 18 billion billion planets seems pretty low, the game needs good AI for a rich experience. But because the AI are on generated worlds, I don't see how they can have complex canned plot lines of interaction with our character. Alliances and fighting? Sure. Cooperation for resource gathering? Sure. But is that enough? Can the AI develop more complex plots themselves? I doubt it.

So the promise of endless world discovery may turn into a very lonely and futile journey. I hope not, because this game is an original, incredible work of art. Either way, Sean Murray is definitely enjoying the adventure.

# Europe 2015/ October 4, 2015by Chris Walker

Chris: Business (representing Element Robot) took me to the Euromold conference in Düsseldorf and the TCT Show in Birmingham. Taking advantage of the opportunity, I brought another entrepreneur along and we also visited sights and friends in Utrecht, Rotterdam, London, and Stratford-upon-Avon. A whirlwind tour, but a very enjoyable one! Utrecht was magical and London was incredible. I'll be back.