The Tendency Towards Entropy

The curious task of economics is to demonstrate to men how little they really know about what they imagine they can design.

Hayek said this about economics but I think it's applicable to everything we design. As an architect I have a front row seat to the limits of our ability to understand the consequences of our decisions, and I think it's fundamental. We get better at controlling for these unintended consequences, but we also seem to consume any advances by making our designs more complex (i.e. buildings used to be built with two or three materials, now it's literally thousands).

A theme I've noticed this year across multiple disciplines is the tendency of systems to be exploited in ways their designers didn't foresee. Informal agreements go by the wayside and once exposed become part of its regular use. I've noticed this in politics, legal contracts, games, the news, internet forums, etc. Historically a popular example of this is the gentleman's agreement that US presidents abided by - the two term limit that George Washington imposed upon himself - then FDR was elected to a fourth term; something no president had ever done. After his death Congress passed the 22nd amendment which limits presidents to two terms. Examples from 2016 might include: news organizations inability to convince readers of their legitimacy in contrast to fake news, the US Congress's unwillingness to review the president's Supreme Court nominee or Federal Reserve appointees, reddit's inability to deflect gaming of their algorithm, and internet service providers willingness to become 21st century robber barons. It almost appears that it's en vogue to break with convention and subvert anything that isn't specifically stated or enforceable legally.

If there's a take-away from me writing this it seems to be that as a culture we seem to be more accepting of this than previously. The spirit and intent of our contracts, systems, etc. seems diminished in relation to our ability to bend these systems to our will using their semantic deficiencies.

The tendency of a system towards entropy is also a law of thermodynamics - which may also be why life exists, so maybe this is actually more typical than the reality we've been living in?

Having a Kid - Notes at One Year

Today is my first child's first birthday. I thought it'd be nice to document some of the changes in my attitudes and views over the last year while I'm still mentally straddling the line somewhere between not having a child and being a parent. Soon I won't reliably remember what not having a kid felt like in the same way that I can't really recall what being unmarried felt like.

The love thing - it's taken time to develop. I'm skeptical of the claims of others to the contrary.

My concern for children's safety and empathy towards parents who have kids who have been harmed or killed is greatly magnified. I'm surprised more parents don't retaliate violently against those who harm their children.

Having a kid is surprisingly similar to having a pet - not a polite topic of conversation. The key differentiator is the ceaseless rapid progression with no plateau in sight.

Kids are less homogeneous than previously thought.

There's a strong possibility that we work too much.

Irreducible Knowledge

I've been noticing lately that I spend a sizable portion of my workday conveying information that gets misunderstood. Not just misunderstood like one of us didn't read the entire email, but literally the party I'm trying to convey information to doesn't have the requisite knowledge to really understand what I'm trying to get across. It sounds condescending - it's not meant to be. We all exhibit this, just in different knowledge areas. 

Below is a clip of the physicist Richard Feyman explaining the issue to a journalist who's just asked him a question the journalist can't understand the answer to because they don't understand basic physics, so instead of answering the question Mr. Feyman (a bit pedantically) explains the difficulty of conveying answers to someone who doesn't posses the requisite building blocks of knowledge in that area.

Most of the articles I read everyday are on subjects of which I'm not an expert, but often they touch on knowledge areas that I've studied and have a degree in - which isn't that important but it's to say I've taken all the foundational classes. What strikes me is that often these articles that deal with technical subjects, the two most notable offenders are economics and nutrition articles, will deal with fairly high level concepts - the stuff of 400 level classes, but they often do so while misunderstanding the basics, the irreducible facts. You can't write a piece on a low carb diet if you don't deeply understand macronutrients. It'd be disingenuous. It's like me writing an opinion piece in a law case when I barely understand the difference between criminal and civil law.

The internet is a boon for the intellectually curious but I often wonder if it gives us a false sense of knowledge. The traditional model of schooling is largely outmoded. Yet, it's difficult to accumulate the basic building blocks of a subject purely through our own research. That's changing with outlets like the Khan Academy and similar. It still leaves me wondering why we're so prone to this this type of error and how we can convey ideas to those outside of our knowledge areas.

The piece that inspired this article, which I missed the boat in commenting on, is an article on economic inequality by Paul Graham. I admire Mr. Graham very much. His essay How To Disagree is essential reading, but I think he misses some basic points of macroeconomics in his inequality article. That someone who is so clearly highly intelligent and deeply introspective troubles me because it suggests this is something we are all prone to.

Two Predictions: Self Driving Cars and Solar

The other day I was attempting to find photos and videos of me playing paintball. I played my last pro tournament at the end of 2007 and it's surprisingly difficult to find much evidence of it. I think it's largely because smart phones didn't exist then. Their adoption has been relatively recent and it's difficult to remember a time before them; the first iPhone was release midway through 2007 with the first Android phone, the G1, trailing at the end of 2008. This isn't news to anyone. My point in bringing this up is that two more shifts like this are underfoot, and much like smart phones everyone knew they were coming several years in advance. Once they happen it'll seem as obvious as the arrival of smart phones. The shifts I see are self driving cars and the dominance of solar electric energy sources coupled with a more distributed power grid. This is in opposition to other renewable sources and our current centralized generation system.

First, self driving cars. The current zeitgeist is that this is happening. So much so that writing about it is redundant. The reason I am is that most of the people I've discussed this with don't seem to grasp how this will fundamentally change our world. It's difficult to understate and the incentives for change in this area are absolutely massive considering that transportation is 10% of GDP. Here's some of what will change:

  1. No one will need to own a car. Cars sit idle 90% of the time. The greater the utilization, like any asset, the cheaper its use every time it's used. Related, this is Uber's actual plan. They're not a taxi company. They're a data company and they're building the ride sharing infrastructure for self driving cars. This is why their valuation is many times higher than the entire revenue of the taxi industry.
  2. Self driving cars will almost never crash. Currently about 30,000 people a year die in collisions in the US. Collisions will likely be treated more akin to airline crashes. They rarely occur and when they do they receive lot of media attention and the investigation is thorough. This will be an issue since people tend to have cognitive issues with ceding control and perceived danger.
  3. And no, there won't be a steering wheel. At some point people won't be allowed to drive on public roads because it's unnecessarily dangerous. The right to do so, while legitimate, will prove too costly to continue. Driving a car will be relegated to closed tracks only.
  4. Cars will travel closer together, think inches/centimeters, which means roads will get smaller and traffic will disappear. What will we do with all the excess space? It also means less road maintenance and traffic signals as we know them will go away.
  5. Gas stations, auto body shops, car dealerships, personal car insurance, lawsuits related to car accidents, traditional car companies, the number of mechanics will be reduced, medical personnel responding to crashes, etc.
  6. Cars will travel faster. On the face of it this seems like it'd make outlying suburbs more attractive, but telecommuting was supposed to do the same thing while in fact the reverse happened. Cities became more populated and suburban and rural areas less so. Telecommuting has made it possible to live a more nomadic life and non-privately owned self driving cars will further that.
  7. Cars will be electric. Why? There's a plethora of reasons, but mostly because it's quickly becoming the cheaper option. They'll also use less energy even though they travel faster.
  8.  Parking lots and garages will largely go away. This may not seem like a big deal but they're expensive and take up lots of real estate that can be used for higher value uses. Much as just-in-time-shipping revolutionized logistics by removing warehousing from the equation why would a self driving car company want to idle an asset and pay to do so? Maybe during periods of low usage, but more realistically the cars will charge themselves during non-peak hours. Maybe they'll even flip the equation by acting as storage for conventional power, soaking up cheap excess power generation during off peak hours and flooding the grid during peak hours when it's financially attractive to do so.
  9. Public transportation will become largely defunct. I haven't read anything about this but I don't see how it won't be affected. Trains are massively expensive and public transit in general is a money loser (it's a political issue endemic to public transit). I don't currently see a scenario that makes fiscal sense. It'll be interesting/frustrating to see how entrenched interests fight this, but the coffin is already designed.
  10. Long commutes are proven to reduce life satisfaction. Since self driving cars will travel faster and you won't have to pay attention this should be mitigated to a large degree, so as a whole it'll make people who endure long commutes happier.

Related to all the above is energy production and usage. Solar won the small scale renewable energy contest years ago and few have taken notice because it's happening slowly and there's a stigma attached to solar (PV, photovoltaics) from its early years. It's also been strung along due to backlash from entrenched publicly regulated energy company's outdated pricing models (charging mostly for usage instead of connectivity which is becoming the real product) and the fact that batteries haven't improved fast enough. The competition was and currently is: distributed natural gas, oil, gas digesters (waste to methane), wind power, hydroelectric, and to some extent conventionally distributed electricity but I see that as part of the same system. True geothermal and wave power generation has never been anything beyond niche. All of these systems will remain but solar's growth will eclipse them all by orders of magnitude. Why this is happening is due to a confluence of fundamental physics and cost issues.

On the physics side of things conventional electric power plants have a theoretical maximum efficiency of 42% due to the Rankine cycle and most operate in the low to mid 30s, wind power has a maximum efficiency of 59% due to Betz's Law and current technology is already near that. Distributed natural gas is more viable but poses a myriad of health issues that are becoming less tolerated. Heat pumps are making conventional heating systems obsolete while induction cook tops are doing the same thing to cooking ranges both of which negate the need for running gas lines to buildings. Hydroelectric poses environmental issues which is why they're almost never built anymore in the US. The list goes on, but the point is that solar efficiency is still in its infancy. There are huge gains to be made while contending technologies are already near their theoretical limits.

The real breakthrough in recent years has been the cost of solar. It's plummeted to the extent that the barrier now is the labor to install them and the regulations surrounding attachment to the grid. A real world example of this is Tesla offering its customers free charging by building stations that run on PV. The brilliance of this is two fold. First, a car bought from Tesla includes in its price all the associated energy costs for the life of the car up front. That's not possible with a gasoline car. Second, PV's costs are almost entirely up front since there's almost no maintenance. Energy is a ruthless finance game and PV is a known quantity. Other systems currently have lower costs up front but unknown variable costs during their useful life. This makes them more difficult to finance.

Energy is a much more nuanced discussion than self driving cars, so I'm leaving a lot unsaid since this is already a tome. The time frame for self driving cars is unknown but has already started. They should really start to come online in the next few years. The total dissolution of our current system may take multiple decades. I have no idea. There are a lot of interests that will be made obsolete and they won't go down without a drawn out fight. Solar is harder to predict but that should take far longer. I wouldn't expect anything too noticeable to happen in the next few years. I'll consider this tipped when the majority of conventional homes are being built with PV and net meters as standard. That could take more than a decade but the writing is on the wall.

Willpower - Ain't Nobody Got Time

My friend Kevin sent me an article, written by a fellow tech entrepreneur, that describes his difficulty deciding how to divide his limited time and attention. His thesis is that we often tell ourselves and others that we don't have time for something when what we actually mean, even though we may not acknowledge it, is that we don't have the attention to give. And I agree - it's a common sentiment.

Attention is just another way of saying our ability to focus cognitive resources on a task, and doing so is a matter of willpower. The last decade has seen the birth of research concerning this - a.k.a. ego depletion (APA 1APA 2WiredYou Are Not So Smart / Podcast). The gist of said research is that willpower is both finite and variable. Finite meaning that our willpower is limited and variable in that the amount of willpower we possess can be altered by behavior.

As to "having time," in research surveys people report having less leisure time than ever, but in fact the opposite is true (thanks washing machine!). So what's going on? All else equal, earning higher wages makes us feel more anxious about our free time and we feel we have less of it even though the opposite is true. Researchers have termed this effect time poverty.

To bring this back to the original article - the author mentions not wanting to take on another responsibility that divides his attention which reminded me of a reddit AMA (ask me anything) from earlier this year with Elon Musk. He replied to a comment asking how he managed to become proficient in so many technical areas:

I do kinda feel like my head is full! My context switching penalty is high and my process isolation is not what it used to be.

What Elon is saying is that changing mental tasks takes time. So much so that he can't focus on a given task as singularly as he was once able to, and when he does he spends a lot of time readjusting to said task.

This leads us to a TED talk about how those in creative fields need large chunks of uninterrupted time to be productive and how it conflicts with managers schedules which are typically broken into 15 minute increments. 

This is a nearly endless subject. I think a lot of it is driven by the fact that most other aspects of our lives are being completely altered, made faster, more efficient, etc. while time and its management has hard to move boundaries which all of us are increasingly bumping up against.

The Means of Production in Architecture: A Macro Explanation

As a follow-up to my last post, I just read "How software is changing the architecture industry." I don't agree with its conclusions (survivorship bias) but I thought it was interesting that others are noticing the same trends, so I'd like to attempt a macro explanation:

This is what pure competition predicts of successful technological innovations. When first invented a new technology offers a comparative advantage to early adopters, but over time the technology's adoption becomes ubiquitous and comparative advantage disappears. The technological innovation becomes necessary just to compete in the industry. Economic profit approaches zero, and if all goes well normal profit remains.

Architecture is either in the early or late part of this cycle. My bet is that we're in the second part - that is, there isn't much if any additional profit to be made off of adopting BIM. Clients and contractors already know this capability is out there and they expect it. Lawyers even more so. The margin of acceptable error in drawings has been much reduced in recent years. Basically, if you're still using CAD you're in a John Deere world with a horse and plow.

The Means of Production in Architecture

There seems to be a natural career arc in many fields whereby in your first few years you focus on production - in architecture this means drafting and becoming familiar with the current software which manifests itself as drawing copious amounts of bathrooms and stairs. The longer you're in the industry the more you're involved with clients and you tend to spend less time on actual production. At some point you're almost entirely disconnected from the means of production - and it makes sense. As you become more valuable you shouldn't do work that lesser compensated people can do, but abandoning what made you valuable in the first place poses issues. You're setting the stage for your own obsolescence. This isn't unique to architecture, but within architecture it leads to those above me not understanding how the software I'm using works. It sounds innocuous but it leads to things like hiring consultants that can't deliver what I need to design complex buildings. It's like asking for steel to make a car and being given wood. I'm really talented but I'm not a magician.

Regarding the means of production there are two types of advancement:  incremental change like updated software and faster computers, and are structural shifts such as moving from a hand drafting based practice to CAD. Again, this isn't unique to architecture, but our structural shifts have unfurled as follows:

Hand drafting: antiquity - late 1980's early 1990's.

CAD: 1980's - present, but in decline.

BIM: clear successor to CAD as of about 2008.

The frustrating part of living through this period of transition is the lag in adoption. Revit (a software made by Autodesk, BIM is the type of software) is to CAD what email is to hand written letters. I can literally do the work of several people using BIM yet almost a decade in and schools aren't teaching it with any rigor and firms are resisting its full adoption. BIMs use is a common topic of debate in both firms and schools, but this isn't an incremental change. This is a structural shift. Much like climate change, If you think there's a debate there's a good chance you have no idea what you're talking about.


I'm a terrible multitasker. Granted, research shows that almost no one is. None the less, my job entails some brute force drafting tasks that allow me to zone out, so I've been consuming a wide variety of podcasts. These are my favorites:

99% Invisible - stories about architecture and design.

You Are Not So Smart - a discussion of cognitive biases and heuristics.

Planet Money - economic reporting and storytelling.

Freakonomics - storytelling about counterintuitive research.

Dan Carlin's Hardcore History (Common Sense by him is pretty good too) - long, well researched historical content that focuses on wars.

The Truth - short fictional stories.

This American Life - investigative storytelling?