You are here

Albert Wenger

Content Written by Author

Thursday, August 13, 2020 - 5:42pm

I have been meaning to write a blog post about Marxism following an exchange in the comments to one of my posts about Trump’s dictatorial tendencies. Essentially the thrust of the comment was that Marxism is a bigger threat in the US today than fascism. I disagree with this assertion, but I do think that the extremes to which we have taken capitalism have opened the door for a resurgence of beliefs that it needs to be toppled entirely rather than shrunken dramatically, as I propose in The World After Capital.

I want to start by pointing out a few things that should be obvious but maybe aren’t. First, there is a huge body of Marxist thinking that has evolved over more than a century and entire books have been written about narrow subfields, such as say Marxist critiques of modern cinema. I find it somewhat comical to think that anyone would find this a threat — it is a valid mode of criticism, which one can debate on its merits, but which in no way is going to give rise to a revolution. When people say that the liberal arts are overrun with Marxist thinking, it is useful to keep this in mind.

Second, there are policies, such as the Green New Deal or Single Payer Healthcare that I disagree with for a variety of reasons (mostly related to their approaches to labor and innovation) but which are not Marxist per se. Applying the Marxist label to them is often an attempt to smear them and avoid a debate on the merits. Canada and the UK have national health systems and last I checked neither of them is a Marxist country. Here it is worth keeping in mind that healthcare is but one sector of the economy and we have other heavily regulated or government owned sectors (e.g. water and sewage).

So the central idea that matters and is worth discussing is that of class struggle between labor and capital that can ultimately get resolved only through worker control of the means of production (and by extension the abolishment of capitalists). The first thing to note is that Marx was perceptive and right in understanding that there is a conflict here — that the interests of those providing labor often diverge from the interest of those providing capital. The second thing to note is that this conflict lay somewhat dormant for many decades as capitalism produced material progress that was widely shared. And the third thing to note is that with the advent of digital technology, the role of capital has changed (again this is the central theme of my book The World After Capital).

What then is wrong with the Marxist idea? The key problem is one of scale. It is entirely possible to have small worker owned companies and there are tons of successful examples for that. The question is how do you do implement worker ownership for something that requires thousands or tens of thousands of people, such as say a car company? Or at even bigger scale, how does worker ownership apply to the economy as a whole? One very quickly runs into governance issues which defy easy solution. That has been the key source of the problems with the attempted implementations of the Marxist model. So far the result has inevitably been a bureaucracy that wields great power and becomes deeply entrenched often abusing the very workers it was meant to represent. This is especially true in the model of Marxism where the means of production are owned outright by the state as a proxy for workers (the idea being that the state *IS* the workers, but the state inevitably winds up being its own entity).

Now I should be quick to point out that the same governance problem also exists in capitalism, but in theory bureaucratic excess is checked there by the functioning of markets. I say in theory, because in practice, especially over the last few decades of the rise of managerial capitalism combined with ever more concentrated markets and regulatory capture, the bureaucratic hierarchy has in fact become largely unaccountable (as have large concentrations of financial capital). We see this is in many forms including the extraordinary rise of managerial compensation as well as various abuses of market power.

So where does all of this leave us? I believe that the idea that all means of production should be owned by the state is a genuinely dangerous one due to the power that it vests in what becomes an unaccountable bureaucracy.  The ideas that we have well-functioning capitalism today or for that matter that capitalism can solve all problems are, however, equally dangerous. As long as we promote these (which unfortunately most of the political establishment in the US does, including much of the Democratic Party) we will have more and more people flocking back to ideas, such as Marxism, which argue that we should overthrow capitalism entirely.

My book The World After Capital, is an explicit attempt to point to an alternative path. A path in which over time capitalist activity will shrink as a part of human affairs, much as agriculture has gone from being the defining aspect of societies to being one of many endeavors.

Tuesday, July 28, 2020 - 8:04am

I wrote a post last week that incensed some people who took issue both with my use of the term fascism and with the apparent focus on the deployment of DHS agents in Portland. Based on the comments here is a new post.

I believe there is a clear and present danger of Trump attempting to become a dictator this fall. I am putting this at less than 10% probability but significantly above 0%. Enough so that I believe now is the time to push back hard against this possibility and not after it has happened.

For reference, I had a hard time in 2016 convincing people – including people I knew on the Hilary Clinton campaign – that Trump had a good chance of winning. It is exactly because people simply couldn’t imagine it happening that they made a lot of bad choices, such as not campaigning in some states. Similarly, if you cannot imagine the possibility of a Trump dictatorship you will not take steps to prevent it.

Let me start by saying that I deem it unlikely that Trump has a masterplan for becoming a dictator. Then again he probably didn’t have a masterplan for becoming president but he pulled it off nonetheless and being a reality show host worked well in that regard (whether planned or not), as did his embrace of Twitter and his ability to inoculate himself against rational criticism. Time and again Trump has shown the ability to seize opportunities that present themselves and part of the opportunity has always been that opponents underestimate what he is capable of.

So what about dictatorship? Well for starters it is clear that Trump publicly admires dictators and that he revels in power. This is obviously not new but something he has announced for a long time in many different forms, including the design of his homes in dictator style.

What has Trump done in office that substantiates any risk? Here are some of the actions that I am aware of that are part of an overall pattern that demonstrates the potential of a flip to dictatorship. These are all actions that have historic precedent as part of dictatorial power grabs:

  • Trump has declared the press an “enemy of the people” and a source of “fake news”

  • Trump has held continued rallies throughout his presidency

  • Trump has agencies run by “actings” who serve at his digression and have not gone through a congressional vetting process which gives him direct power

  • Trump has undermined the independence and effectiveness of the judiciary through his appointments of unqualified judges, verbal attacks on judges that have ruled against him, and by commuting the sentence of someone convicted of obstructing Congress

  • Trump is tweeting frequently about how the election is being rigged as a way of putting the legitimacy of the results in question

  • Trump has been torpedoing the United States Postal Service as a way of interfering with mail-in ballots

  • Trump is moving federal agents into predominantly Democratic cities, nominally under a “law and order” agenda

  • Trump has pardoned soldiers implicated in war crimes and has publicly belittle military leadership. [*]

  • Trump has interfered with the functioning of the Federal Election Commission (FEC) by not nominating commissioners depriving the FEC repeatedly of a quorum. [*]

I am sure there are other actions that fit the pattern that I am missing and I would love for people to contribute more examples (update: the ones marked with * above are additions based on comments).

My longstanding opposition to Trump, going back to his 2016 campaign, rests on my view that he represents a meaningful threat to the workings of democracy and the principles of critical inquiry and science, which together have accounted for much of the progress that we have achieved.

To be clear, as this was also brought up in comments, I do believe that we require dramatic changes and that recent Presidents, including Obama, were incrementalists which has been completely inadequate fo the challenges we are facing (this is the subject of my book World After Capital). And yes Hilary Clinton would have been an incrementalist as well.

Finally, let me also add, lest someone bring it up as a distraction from the risk discussed here, that I am against violence, including damage to buildings — both because I believe it to be wrong, but also because I think it is ineffective and worse than that plays into the hands of someone who is a potential dictator as it provides an excuse.

PS For one commenter in particular, I will write a post about Marxism also. I do believe its resurgence is deeply problematic but doesn’t pose the same kind of clear and present danger (I will explain both of these in the post).

Monday, July 27, 2020 - 5:08pm

By now, if you are in tech and haven’t been on an extended Techmeme and Twitter break you have heard about GPT-3 a new and massive language model developed by OpenAI. I have played around with it a bit myself and its capabilities are impressive. Here are some thoughts.

The model perfectly illustrates the ever moving bar on what we consider artificial intelligence. Not that long ago facial recognition didn’t work at all, now computers routinely outperform humans. Putting together more than a couple of new sentences that actually made sense was a really challenging problem, GPT-3 routinely cranks out multiple paragraphs. “But it doesn’t understand the sentences” is the immediate objection often heard. Of course “understanding” is as poorly defined a term as “intelligence” so that objection really doesn’t do much.

Instead, a better way of thinking about these capabilities is to consider where and how they can be deployed instead of using humans. One obvious example would be customer support. A customer writes in and someone needs to write an answer. I am not suggesting that anyone should send out GPT-3 produced answers without checking them first, but the number of customer support requests that someone might be able to handle using GPT-3 could go up tremendously. Another great example is producing UI code. This is often a labor intensive part of projects and a lot of engineers dread it as it’s not exactly solving fun problems but rather wrestling with the idiosyncrasies of different platforms. There are already several early demos of how GPT-3 or a model like it could be used for that. And yes, this does and should change how we think about the longterm demand for human software engineers (something I pointed out 6 years ago in a post titled “Software Is Eating Software”).

A similarly faulty line of thinking has been that only humans can be creative. Again this is enabled by a completely underspecified definition of what it means to be creative. Here is an example of a poem that GPT-3 cranked out. Admittedly not exactly a masterwork but if a young student would turn this in the teacher or parents might say “that’s so creative!” But it doesn’t stop here. After a bit of back and forth over Twitter about submitting GPT-3 work to a poetry contest, Joshua Schachter prompted the model for a story about using AI to submit a poem to a contest. And the resulting short story is really quite impressive.

All of this is to say that objections around intelligence and creativity are rooted in definitional problems and also obscure the extraordinary potential of this technology to change the need for human labor. Of course, this is one of the foundational premises of my book World After Capital.

Of course it is also clear how this type of model can be used for all sorts of bad things, such as automating high quality bot attacks on social media or even creating content that can be passed off as having come from a particular author (but was not in fact written by them). This will put a premium on attribution, something I have written about in the past in a post called “Sign All Things.” One key reason for having self sovereign identity, with some probability of that identity being a human attached, is to mitigate against these types of attacks (btw, humans even under their real names say plenty of terrible things, so it doesn’t help with that as people sometimes think).

Finally a brief thought on the question of bias which always arises. Of course models are biased by the data on which they have been trained (at this point this is well established). The same is of course true for humans who are biased based on how they have been trained. But there are some key differences that are worth keeping in mind. The bias in a model is more measurable than for a human – the will produce text after text after text (and it will not be strategic about its answers, well, at least not intentionally strategic, it might be implicitly strategic in as much as it has picked those strategies up from the training data). There is also a much clearer hope of reducing the bias in models compared to humans.

Where does all of that leave us? GPT-3 is a major step forward in the capability set. It shows great new powers and as we know from Spiderman, with that comes great responsibility. There is the responsibility of OpenAI to monitor how this model is used and to measure and reduce its biases. But as importantly is our collective responsibility to create a future that lets humanity benefit broadly from these emerging powers. That is the very subject of World After Capital.

Friday, July 24, 2020 - 5:47pm

Earlier this week I wrote a post about Trump’s fascist actions which need to be vigorously opposed. In the post I predicted that people would show up with excuses. The comment section delivered on this in spades. I will save anyone the pain of wading through the mess and provide a summary here.

Excuse #1: The actions are legal

This is presented along with some copy/pasted section of law or some assertion as to the powers of the President. While we absolutely have given the Feds too much power via laws such as the Patriot Act, this excuse falls down on multiple grounds. First, we have the judiciary to establish the legality of an action and the actions in question here are being contested in court. Second, there are no clear cut laws here. For everything people cited as incontrovertible there are countervailing considerations, including the First Amendment and the Tenth Amendment (which nobody brought up). Third, just because a law currently exists doesn’t mean it is constitutional.

Excuse #2: Trump is protecting federal property

This goes along with a variety of criticism of local government and/or protestors. If the administration really cared about the building they would not have waited seven weeks before this action. Also of course, this excuse got destroyed later in the day by non other than Trump himself announcing the extension of federal agent activities to other cities that don’t have any threats to a federal building. This type of excuse is the simple repetition of propaganda.

Excuse #3: Agents were actually identified

For starters even if they truly were this wouldn’t change anything unless they also stayed in the immediate proximity of the Federal Building, arrested only people right there who were attacking them or the building and had been deployed to actually protect the property (see #2). But of course even in isolation this doesn’t hold any water. The only evidence presented for this excuse was the DHS press conference. Any review of the many images and footage of the agents operating shows that they were impossible to identify.  

Excuse #4: Trump is just grandstanding, playing to his base, campaigning

This excuse is a way of belittling a large transgression of political norms to make it appear small. The deployment of federal agents in large groups to states is a big deal. One easy way to see this is to consider how rare it has been throughout US history. The Tenth Amendment covers the division of labor between the federal and state level with policing clearly the role of the state. This matters a lot considering that the elections are coming up in November. The cities the administration has on their list are heavily Democrat, which is exactly where Trump has a strong interest in voter suppression which will be a lot easier to accomplish with federal agents already deployed.

Excuse #5: Calling it fascism is hysteria

This excuse comes in some harmless sounding form like “this does look bad, but calling it fascism is hysteria” where it appears that the commenter is sort of agreeing. The label “lame” isn’t quite appropriate here — “pernicious” would be a lot better. It deliberately or unwittingly removes the effectiveness of protest by trying to remove the label of “fascism.” A related excuse is to get into nitpicking over whether this is “fascism” or “authoritarianism.” Again these are ways of blunting the criticism to the point where it becomes ineffective.

Why am I calling these excuses instead of arguments? Because a situation of clear and present danger requires action. And when arguments have either the explicit purpose or the inadvertent side-effect of suppressing action then they become excuses.

What is the clear and present danger? The possibility of Trump becoming a dictator. To be clear, I am putting that at less than 10% but also significantly above 0%. Not because I think Trump has a masterplan but because he is an opportunist with an admiration of and desire for dictatorial power. And the combination of COVID19 and demonstrations provides the perfect opportunity for seizing emergency powers.

PS I am also banning a commenter for using schoolyard bullying tactics which I had previously explained to him were not acceptable here on Continuations. I gave him the opportunity to remove the specific comments which he chose not to. I will reconsider this, if he goes ahead and deletes those. 

Wednesday, July 22, 2020 - 9:17am

I was born in Germany in 1967. Growing up my friends and I used to wonder why our grandparents hadn’t resisted Hitler’s rise to power. We were young and full of bravado and oh so sure about how outraged and courageous we would have been. But the sad truth of course is that we most likely wouldn’t have and that we were simply talking tough with the double benefits of hindsight and zero actual consequences for ourselves. Most people didn’t protest for one of many possible reasons, including ignorance, disbelief, profiteering, apathy or fear (the latter especially as fascism picked up speed).

Now we are all given an opportunity to protest a slide into fascism right here in the US. Because that is exactly what the deployment of unidentified federal agents into Portland represents. I can’t wait for commenters to show up with anything from “it’s just an election stunt” to “this is just what Portland needs” and seventeen other lame excuses. So let me get this out of the way: if you don’t call this out as fascism, you would not have protested Hitler either. You would have been a supporter or a bystander.


“Albert, you are being overly dramatic.” No. If this is not your line, you don’t have a line, especially as Trump went ahead and publicly proclaimed his plans to roll this out to other cities. This goes against anything the constitution stands for, from due process to states’ rights. It is a direct attempt to intimidate opposition and fire up an ever more aggressive movement. Oh and let me save you the time to head over to Wikipedia for fascism. Here is the opening sentence “Fascism (/ˈfæʃɪzəm/) is a form of far-right, authoritarian ultranationalism characterized by dictatorial power, forcible suppression of opposition …”

What can you do? Well for starters, use whatever platform you have to call this out as fascism. Then call on your state representatives and senators to do the same. Your governors too. And most importantly get your District Attorneys and Attorney Generals to sue the DHS, as the Oregon has done. We have to use independent institutions as long as these are still functioning. This is crucial because once these go everything goes. And we have to use them to hold the line through the elections in November.

Related: Catalog of excuses.

A version informed by comments: The Threat of a Trump Dictatorship

Wednesday, July 8, 2020 - 6:41pm

My Twitter feed these days is full with back and forth on “cancel culture.” Much of it quickly reduces to bickering with each side assuming the worst possible interpretation of what the other meant or what their motives are. The latest exhibit here is the Harper’s Open Letter and the reactions to it. This is more than a bit annoying as there are some genuinely interesting issues here that are worthy of a more robust debate.

Central to this discussion ought to be an analysis of the power of speech to suppress speech. I think most people would agree that calls for physical violence against someone for speaking are a clear suppression of speech. Now that would be an example of incitement, which is notably one of the exceptions to free speech, i.e. it is not protected speech.

But what about calling for someone to be fired from their job? Consider the following two scenarios:  “<<name of CEO>> tweeted <<xyz>> and should be fired” versus “<<name of low level employee>> tweeted <<xyz>> and should be fired.” We intuitively feel that there is a difference between the two but where does it come from? The obvious answer would seem to be power, but it is worth making that more precise.

CEOs tend to be highly compensated (not necessarily true of the founder of a startup) and so CEOs are assumed not to be immediately destitute if they lose their job. But the real reason there is a power differential is that the decision to fire a CEO belongs to a board of directors. And generally we would assume a board would deliberate on the question what the right consequence should be, depending on what <<xyz>> is which might simply be a reprimand or potentially nothing at all (note: there are and have been exceptions to this also, which I will get to shortly).

This is in stark contrast to a low level employee who is both likely to actually get into financial trouble quickly and is often seen as entirely dispensable. If there are enough people calling for them to be fired, a company might conclude that it is simply bad for business to keep this employee around and summarily let them go. It is easy to remark, “well they shouldn’t have tweeted <<xyz>>” but that clearly implies a curtailment of speech, which is what this debate should be about.

There are effectively two things that are new and have come together that require we examine more deeply what is going. The first is the ability for speech of the “fire them” type to be massively and rapidly amplified in a way never before possible. This is reasonably well understood although the remedies being proposed for information cascades are largely rooted in an outdated Industrial Age regulatory mindset.

The second is the weakness of institutions in defending their constituents. This needs much more attention than it is getting right now. Why, for example, are universities so quick to give in to pressure? Because their model is super fragile today. They have taken up their price to ridiculous levels while at the same time facing competition from emerging online alternatives. Sounds familiar? Yes, because the same has been true for publishers for quite some time. Many companies find themselves in a similar position.

So where does all of this leave us? For starters nothing is helped by trying to dismiss these changes as irrelevant when they are so clearly evident all around us. Then we need to be willing to explore new and more fundamental solutions rather than trying to patch the existing systems. I have long proposed making systems such as Twitter and Facebook programmable as one digitally native way of reducing their power including defanging the risk of information cascades. It is the basis of what in my book World After Capital, I refer to as informational freedom. By itself it won’t be enough though, which is why I am also a strong proponent of economic freedom (some form of universal basic income) and psychological freedom (through having a mindfulness practice).

If we want to restore our ability as a society to have productive critical discourse, then patching our Industrial Age systems won’t do. Instead we need to fundamentally reinvent how we do things in the digital age (which I hope can turn into the knowledge age).

Tuesday, July 7, 2020 - 5:39pm

ICE announced yesterday that foreign students at schools that only instruct online this fall (which includes my alma mater Harvard) must leave the US. I was a foreign student on an F1 visa and would have been distressed by this. More importantly there are quite a few foreign students for whom there is not enough time, or not enough money, or even worse a threat to their lives in their home countries. This is yet another one of those recent US policies where the only plausible assessment is that “the cruelty is the point.”

The ability to attract foreign students to our universities has been wonderful for the United States and the world. A great many international friendships that have lasted a lifetime are formed among students. The presence of foreign students helps expand the cultural context for American students and vice versa. Many talented foreign students wind up staying here and starting companies or contributing to existing companies. And so on.

I hope this nonsense gets challenged in court right away and if anyone is aware of such a challenge I will happily contribute to it. In the meantime, I have signed (well tried to sign, subject to email verification which seems to be down) this White House petition.

Friday, July 3, 2020 - 9:33am

In my book World After Capital, I propose a theory of history in which technology changes the binding constraint for humanity. After hundreds of thousands of years of the Forager Age, constrained by the availability of food in the natural environment, humanity invented agriculture. With that set of technologies (planting, irrigation, domestication of animals, etc.), invented roughly 10,000 years ago, the constraint shifted from food to land in what became the Agrarian Age. Then only a couple hundred years ago a series of new technologies (steam, electricity, chemistry, mining, etc.) shifted the constraint again, away from land to capital. By capital I mean the physical capital of the Industrial Age, such as factories, buildings, and roads. More recently with the advent of digital technology (computers, packet-switched networks) the constraint has shifted yet again, away from capital to attention.

In the book I note that each of the two prior transitions came with dramatic changes to how humanity lives. In the transition from the Forager Age to the Agrarian Age we went from being nomadic to sedentary, from flat tribal societies to extremely hierarchical feudal societies, from promiscuity to monogamy, from animistic religions to theistic ones. In the transition from the Agrarian Age to the Industrial Age we went from living in the country to living in the city, from large extended families to nuclear families or no family at all, from commons to private property (including private intellectual property) and from great-chain-of-being theologies to the Protestant work ethic. I then go on to make the argument that we need a similarly dramatic set of changes to get from the Industrial Age to the Knowledge Age, which explains why the incrementalist changes pursued in most developed societies have fallen far short. I then propose three increases to freedom — economic, informational and psychological — to help us with the adjustment that’s needed.

I have been happy with this characterization of the role of technology in human affairs. Big technological shifts change the binding constraint on humanity, which results in large scale reorganization of how humans live. What I have been struggling with is to what degree we can understand the features of those reorganizations. While they make a lot of sense ex-post, we are now finding ourselves in the midst of one and would therefore ideally learn from history. This is particularly important for two reasons. First, in the past transitions we often wandered in the dark for long periods before finding a successful model (those periods were often marked by extreme violence through revolutions, wars as well as mass death from disease and starvation). Second, in the current transition we don’t have a lot of time as we are faced with the ever accelerating climate crisis.

This is where I now feel I have had another breakthrough in thinking, which resulted in a “Duh” forehead slap. The first part of the answer is of course “incentives.” That much I had thought all along but the missing piece was considering why incentives had to change and that’s where measurement comes in. So here we go.

In the Forager Age, when the constraint was food, the measurement problem was almost trivial: everyone in a tribe sees how much food the hunters and gatherers bring back! It is either enough to feed everyone or not. In so-called immediate return societies (which had no storage) that’s literally all there is to it. With a bit of storage the story gets slightly more complicated but really not by much. I believe this explains many of the features of successful foraging tribal societies including the flat hierarchy and the equality of sharing.

In the Agrarian Age, when the constraint was land, the measurement problem got significantly harder: you can really only tell at harvest time (thus only once per year in many regions of the world) how well off a society will be. Again, I believe that this explains many of the features of successful agrarian societies, in particular the need for a lot of structure and strict rule following. It is crucial to keep in mind that these societies were essentially pre-scientific. So they had to find what works by trial and error. When they found a rule that seemed to work they wanted to stick with it and hard code it (much of this happened via the theistic religions).

In the Industrial Age, when the constraint was capital, the measurement problem got even harder. How do you know where a factory should be built and what it should produce? It might take years of process and product innovation to put physical capital together that is actually truly productive. I believe this explains much of the success of the market-based model, especially when contrasted with planned economies. Effectively the solution to the incentive problem moved from static rules to a dynamic process which allows for many experiments to take place and only a few of those to succeed.

So far we have seen how the prior shifts from food to land and then from land to capital corresponded to massive increases in the difficulty of the measurement problem. We went from nearly immediate measurement (Forager Age) to yearly (Agrarian Age) to multi-year (Industrial Age). This brings us to the current transition from capital to attention as the binding constraint.

While I don’t claim to know what the features of successful Knowledge Age societies will be, it is clear now that the measurement problem for attention exists on a decadal or potentially even hundred year scale. Take the current coronavirus pandemic. The last prior pandemic of similar scale, including economic impact, was the Spanish Flu about one hundred years ago. So as a society you may only find out if you have paid enough attention to pandemic preparedness every hundred years. It gets even worse when you think about a big asteroid strike on earth. Those happen roughly every million of years.

This is also true from the perspective of the individual. If you commit your attention to some artistic endeavor or research program. When will you know if your attention was well spent? Often not within your own lifetime. Some art that we today recognize as magnificent as well as science which we regard as transformative was dismissed, considered fringe or even actively fought for decades. And conversely, much of what was popular in the moment has not withstood the test of time.

The allocation of attention cannot and therefore should not take place on the basis of near term measurement. Markets are good for the multi-year measurement problem of capital but terrible for the much longer timescale one of attention. Much of World After Capital is about how to free humans up to allocate their attention as they see fit outside of any near term measurement system. Free to allocate their attention to what they believe will stand the test of time. One super important question that flows immediately from this, which World After Capital does not (yet) answer, is how that can result in enough attention on topics such as pandemic disease or the climate crisis.

PS History and society have tons of detailed events and features that are not explained by any of this – my concern is strictly with what I perceive as a very large scale pattern.

Wednesday, June 10, 2020 - 5:40pm

Today Algorand and Blockstack are announcing that they are collaborating on the Clarity smart contract language. Clarity is a decidable (i.e. non-Turing complete) language inspired by lisp. I am extremely excited about this development and have spent a few hours this week writing a smart contract in Clarity.

I have long argued that Turing complete smart contract languages are problematic because the only way to really understand the system as a whole is to run it and see what happens. Still this is the route most projects have chosen and even more odd to me has been the embrace of WASM as an execution environment which optimizes for speed over safety and analyzability. Given that the key point of decentralized systems is censorship resistance that has long struck me as the wrong order of priorities. After all, there is no-one you can call to just unwind a mistake (modulo the DAO hack reset, I suppose).

Clarity instead is all about safety. Avoiding costly mistakes long before they can cause millions of dollars in damage. Decidability is one key aspect of that. To give just one example of the power of decidability: with Clarity you can know the precise gas fee of a contract in advance. Another key aspect is making actions such as issuing non-fungible tokens, which are a critical component of many applications, completely straightforward. There are many more aspects of Clarity that are a delight, such as the simple and clear type system. And being inspired by Lisp, Clarity naturally lends itself to building up code out of short functions (if you have never seen Lisp before, the language may strike you as odd at first – but I promise that learning it will make you a better programmer).

Now some people may think that you lose a lot by giving up Turing completeness, so I wanted to write a non-trivial example. One came to mind easily: an all-or-none funding mechanism. There are many possible use cases for such a contract, including a system such as Kickstarter, or my personal favorite, the second coming of Kitchensurfing (I hope). For readers who may not remember Kitchensurfing, it was a wonderful service where a chef would come to a home and cook there. The key missing feature though was the ability to get everyone to split the payment upfront, so that meant the host had to put up the entire bill and then collect from guests which is more than a bit awkward.

I won’t post the (nearly) complete contract here, you can find it on github. There is a bit more argument checking I need to add in one of the functions and I also have been remiss in writing a test harness (yes, I know I should have written that first, but I was too damn excited).

Here are some key things I would like to call out. How do you define a non-fungible token in Clarity?

(define-non-fungible-token event-pass uint)

That’s all there’s to it. Later, when you want to issue an event-pass to someone who has paid, all you need to do is

(unwrap! (nft-mint? event-pass event-pass-id tx-sender) (err "unable to issue event pass"))

Again, that’s it. Three lines of code and that includes the error handling!

Let’s look at a different part of the code to show another elegant feature of Clarity. With asserts! I can easily check if a condition is met and if it is not met issue an error. Doing so will abort execution and leave all blockchain state untouched

(asserts! (< (get expires-at event) block-height) (err "event still funding"))

There is a lot more to discover and I encourage everyone who is interested to check out Clarity. The tooling is still growing but there is already a Clarity LSP for VS Code (which helped me find a number of mistakes). A full blown REPL running against a chain instance is in the works and will be a great way to interact with Clarity contracts during design.

Over time Clarity will become an independent project. It would be wonderful to see other chains support Clarity as a development language (possibly even the only one). In the meantime there are many amazing opportunities to contribute, from just writing contracts in Clarity (and providing feedback) to working on the virtual machine or on development tooling.

Sunday, June 7, 2020 - 5:37pm

“If you can’t measure it, you can’t manage it” goes an adage that is often recited as if it were some kind of profound truth (misattributed to Deming). And of course there is some logic here. Let’s say you own a website and you want to grow it. Measuring traffic to the site will help you figure out whether you are succeeding with that. But even this seemingly trivial example helps us understand a crucial point which could be summarized as “You manage to what you measure.” 

Suppose that page view is your traffic measurement, now you can immediately see that there are many different ways of growing that which will have widely different implications. You could buy random traffic. You could split up articles into multiple sections each requiring a click and hence a separate page view. You could publish salacious nonsense to increase social sharing. And so on.

We of course recognize many of these as the very aberrations that plague so much of the web. This is just one example for a general problem: quantity is easy to measure, quality is difficult to measure. Pretty much all the other metrics that people have introduced in the web world (visits, visitors, time, etc.) are all attempts to capture some aspect of quality.

There are situations in which quality is so difficult to measure and simultaneously so important that any measurement of quantity becomes dangerous. Fundamental research is one example. In a misguided application of the adage from above, funding for fundamental research has been increasingly allocated based on quantity metrics, such as number of publications and citations. The results from this approach have been horrendous. For example in theoretical physics an area such as string theory has received massive funding (based on these metrics), while disruptive new approaches such as constructor theory struggle to attract resources.

Venture capital is a lot closer to fundamental research than it is to web traffic. Assessing the deep quality of a startup is hard. As with research often a great deal of time has to elapse (years, possibly a decade) before you can tell which companies are truly important. It is therefore at best a distraction to try and build a successful portfolio by keeping track of such statistics as how many startups we met with last quarter. At worst over time all the resource allocation of the firm will go towards those metrics and none of it towards thinking, which is crucial for actually figuring out new things.

At USV we spend a lot of our time thinking and none of our time tracking quantity or other superficial aspects of deal flow. Of course we do track the outcome of our process over time, but knowing that a lot of time needs to expire for those outcomes to mean something. As we have written recently we will add diversity of the founders we back to that.

Now you might say but how will you ever improve racial diversity of founders in your portfolio if you don’t measure it in your process? The answer is by changing the process. So much of the venture capital process revolves around networks. As a result you can have a big impact on outcomes by changing which networks you are connected to. We did this successfully a few years back when we recognized that we had not backed enough female founders. We will be doing this now.

Finally, measurement also relates to the question whether USV is doing enough about the climate crisis. As I have stated on several occasions, including in this interview with Jason Jacobs, I believe that the climate crisis cannot be solved without getting out of the industrial age. I have written a whole book about how we might accomplish getting to the knowledge age, see World After Capital.  It is a mistake to believe that there is some easy measurement that can be applied to USV’s portfolio that would – quarter by quarter – tell us whether or not we are helping with that transition. Only time will tell.

Friday, June 5, 2020 - 6:38pm

I watched a livestream organized by BLCK VC yesterday. It was incredibly powerful and should be widely seen.

Tuesday, June 2, 2020 - 5:36pm

Watching what is happening in the United States, my adopted home country, is incredibly painful. I have seen the last few days coming for years based on the ever more untenable income and wealth distribution combined with militarization of the police and their ongoing unchecked brutality, especially against black people. You can browse back through Continuations and find many posts speaking to these topics, such as this one from 2014.

I have a lot more thoughts about this moment in history and where it might go, but here are some unequivocal statements I want to make right now.

1. George Floyd was murdered by police officers who need to be prosecuted accordingly.

2. Black Americans continue to suffer from systemic racism that has its roots in slavery.

3. Americans of all races are living increasingly precarious lives, while a small group has accumulated vast wealth.

4. Police forces all around the country are overly militarized, structurally corrupt and unaccountable.

5. Demonstrations against #1 - 4 above are entirely justified and must continue

6. Police violence and escalation have injured and arrested citizens who were engaged in peaceful protest in #5 in violation of their 1st amendment rights

7. Looting and burning businesses is wrong and counterproductive (while yet also being entirely understandable)

8. Some of #7 is caused by groups intent on escalating violence

9. Trump has been inciting violence from the beginning of his candidacy and is reveling in the use of force

10. We must do everything we can to remain peaceful even in the face of state and agitator violence (peaceful does include civil disobedience, such as sitting and blocking an intersection).

Saturday, May 30, 2020 - 11:40am

“Two out of three ain’t bad” is the title of a Meat Loaf song but is also the idea behind triangles: situations in which you can achieve only two of three objectives. The classic example is the idea that in manufacturing you can have any two, but not all three from fast, cheap, good. (*)

Much of the current debate about social media and how to regulate it is people shouting loudly past each other because they are pursuing different objectives in what I believe is a social media triangle:

Freedom: there isn’t a central authority that can exert power over individual expression or appropriate rents generated by contributors to the system.

Openness: anyone can join a globally connected network and express themselves without being trolled or harassed.

Criticism: there is a mechanism by which people get exposed to opposing viewpoints and relevant facts and by which information cascades (especially spread of misinformation) are curtailed.

All the existing big systems such as Twitter, Facebook and YouTube fail on the first objective, as they are controlled by for-profit corporations, which are further subject to regulation and intervention by governments.

This of course has many people, myself included, arguing for decentralized alternatives. I think we have to be clear though that this will make accomplishing objective #2 harder (although to date the bar set here by the centralized players is extremely low). More importantly though accomplishing #3 will not be possible in a decentralized system.

At the current trajectory though there is a good chance that we will wind up with the worst of all outcomes, at least for a while: a regulatory environment in which massive players are perversely protected from new entrants but simultaneously hidebound because their conduct is subject to behavioral rules (which require consulting an army of lawyers for every product change). That is what stagnation looks like and we know it all too well from many other industries as well as some prior moments in information technology. In tech those moments were kept thankfully short by massive platform shifts (e.g., mainframe to PC) but there isn’t an obvious one of those on the horizon (except for blockchain but more on that in a moment).

There is a clear alternative to this which helps us accomplish objectives 1 and 2, while at the same time incentivizing competition. Give a Section 230 like protection to companies in return for providing a complete set of enduser APIs. In other words, require Twitter, Facebook, YouTube etc. to be fully programmable in order to have their liability limited.

How does this accomplish objective 1? By reducing the network effect lock in of the incumbents. In a fully programmable world new interfaces can be created that let me interact with multiple systems transparently, so I don’t need to keep track for example of which social network my friends are on. Some of these alternatives may be new centralized systems but others may be decentralized ones.

What about objective 2? In a fully programmable world, users can control what they see and what they don’t. So while that will not prevent trolls from getting on a system it allows endusers and coalitions of endusers to filter what they see.

Of course this immediately shows that objective #3 will be challenging. Again centralized players have done a horrendously bad job at this to date, largely because they are optimizing for total attention gathered. To get a glimpse of how hard that will be one need to look no further than information cascades on WhatsApp.

Why do I still think this is better? Because the misinformation problem is much larger than any one social media system and reaches all the way back to such fundamentals as how people learn and what value systems they internalize. In other words, it has been and will continue to be one of the central problems of human progress. For more thoughts on that I have an entire book which you can find at World After Capital.

(*) It turns out that you can achieve all three if you start with quality but only if you build a culture of quality. Similarly here I believe you can eventually accomplish all three if you build a culture of criticism.

Wednesday, May 20, 2020 - 11:38am

I have been getting a large number of requests for help with internships. Many students either never got one for this summer or had theirs canceled. My advice to them is: create your own. Figure out what it is you want to learn, start learning it, write about it online and find others to join your effort. 

These days there are amazing resources online on almost any topic plus you can social networks to find like minded DIYers. I propose that everyone who pursues this path use the same hashtag #DIYInternship so that folks can more easily find each other.

I realize that not everything can be done that way for example if you need access to machinery or to a lab. And yes initially you will start by yourself and you have to build your own contacts. Also, of course internships are part of the path towards finding employment. But here it is quite clear that a #DIYInternship will show more initiative than simply doing nothing this summer.

Finally, if this takes off, I am pretty sure there will be people with experience who will jump in and answer questions or provide advice. So stop waiting for someone to give you an internship and starting creating your own.

Friday, May 15, 2020 - 5:36pm

Towards the end of 2019, which feels a decade ago, I wrote a post on growth. It was an attempt to get away from the blanket term “growth” and its counterpart “degrowth,” which have become stand-ins for entire philosophies and move the discussion to specific categories such as knowledge (need more) and consumption (can do with less).

The COVID19 crisis, while causing immense pain, also provides a unique opportunity to reflect on what we need in our lives and what we can do without. And that examination should guide how we rebuild our economy and society.

The lockdown has resulted in the highest unemployment levels since the Great Depression. In one shocking statistic, it appears that almost 40% of people earning under $40K have lost their job. This is a massive disruption to what in my book World After Capital I call the job-loop: the vast majority of people spend most of their waking hours in a job and much of the rest on consumption of goods and services, which in turn are produced by other people having jobs.

Now lots of people are asking how will we bring these jobs back? But the more fundamental question to ask is: what can people spend their time and attention on if they don’t need to have that specific job? Or possibly any paid job? Or if they are empowered to start their own business? So far the government stimulus amounts to $3 trillion, which could easily pay for more than 1 year of Universal Basic Income (UBI).

What about the other side of the job loop, consumption? We are seeing more clearly than ever right now that health and freedom of movement are deep human needs. Having dozens of pairs of sneakers, by contrast, is not. It may be cool, hip or even constitute a valuable collection, but is clearly a want. If you are sick from COVID19 or locked up in a tiny apartment, this distinction between needs and wants will be crystal clear.

Way too many of our productive resources are aimed at making and selling stuff to people that they do not need. And we are seeing how hard it is and long it takes to retool those resources to instead produce masks and tests and everything else we truly need. The crucial reason for this is that we have become overly reliant on the market for allocation, especially the allocation of human attention. This is deeply problematic because prices do not and cannot exist for crucial needs, such as pandemic preparedness and finding purpose in life.

So here too our question shouldn’t be, how can we bail out all the businesses and keep making more stuff, but rather what is it that we need and won’t get from the market? We know some of these areas already and the crucial one of course is solving the climate crisis. But there is a lot more, such as the opioid crisis, which in turn is just part of the larger crisis of individual purpose. There is no shortage of issues and people deserving more of our attention than we are giving.

Susan and I have been incredibly fortunate (so far) to be healthy and have our loved ones be healthy. We are financially secure and can do our work remotely. And on top of that we are in a place with good internet connectivity where we can just walk outside whenever we want to and be in the woods.  But we know many others who are not nearly this lucky and much of our time has been focused on helping where we can. Being able to do so has been fulfilling and has made us shift our priorities towards doing more of that faster.

If ever there was a time to reset personal and collective priorities, this is it.

Sunday, May 3, 2020 - 9:38am

One key lesson from COVID19 is that we need a lot more decentralization. This is especially true when the center is as inept at managing the crisis as the US federal government has proven to be. For example, the power of agencies such as the CDC and the FDA has turned out to be problematic, e.g. in giving guidance on mask wearing or trying to increase the availability of testing (both central to the road back). This is not just a critique of current leadership but rather of the accretion of excessive federal power more generally.

The size of the economy of New York State is roughy $1.7 trillion as measured by the equivalent of GDP (an admittedly bad measure). That is about 150 times the GDP of the entire United States in 1800 (assuming I did my math right on that).  Or if New York were a country, it would rank 11th in the world, ahead of over 150 other countries. California is even bigger coming in 4th in the world (and ~275 times the size of the United States in 1800). It is completely unclear why outside a few crucial topics — that can only be regulated at the federal level — states of this size should not be making independent policy decisions. For example, why shouldn’t New York and California approve their own at home tests?

COVID19 may, however, turn out to be a catalyst for the ultimate decentralization, that of money. The dollar’s role as a global reserve currency has for many years put the US in a position of strength. But dollar dominance has proven to be a massive problem in this crisis — everyone who has dollar denominated debt, which includes not just US corporations and states, but also foreign sovereigns and corporates was relying on economic activity, including international trade, to produce the dollar necessary for debt service. With the COVID19 lockdowns that source of dollars has suddenly dried up which has forced the Federal Reserve to step in, producing an extraordinary 2.35 trillion dollars in the space of 6 weeks. For a super clear explanation of this see Jill Carlson’s great post.

The Fed is effectively making a last ditch attempt to prevent a massive global debt collapse. Even if we can stave that off in the near term, the crisis will make many entities around the world accelerate their search for an alternative to the dollar. This isn’t just idle thinking as the extraordinary speech by then Bank of England governor Mark Carney shows and is further illustrated by the massive freakout that central banks had over Libra’s plan for a stable coin based on a currency basket (the search for an alternative clearly does not include one that was feared could be controlled by Facebook).

One of the most interesting ways the decentralization of money could really pick up steam is with community currencies. US States cannot print money but will find themselves with massive budget holes from a combination of increased crisis response spending with a massive loss in tax revenues (footnote: there may be a way for states around this, but it is likely complicated and might result in an ugly fight). But there is a long history of community currencies in the US. And of course there is the famous “Miracle of Wörgl” in which a town helped lift itself out of economic depression by creating its own currency.

This is also an opportunity for crypto technology to really come into its own. For example we have been spending time upstate New York in Columbia County. It would be fantastic to have a local digital currency that is created on and settles via a blockchain. The county, or even a single city like Hudson, could issue it. Or better yet, citizens could create it for themselves. If anyone is aware of such experiments, I would love to learn more about them.

H/T to Tamar and Pete for getting my thinking on this going earlier today with an email exchange starting with this post by the Schumacher Center.

Saturday, April 18, 2020 - 5:38pm

Since I wrote my post on “The Road Back from COVID19″ we have made lots of progress on two out of the three pillars: masks and tracing. Masks went from a discussion of their merits to being mandatory to wear in public in several states, including New York. Tracing went from a hodgepodge of approaches to an API supported by both Google and Apple. On the third pillar of testing, however, we are massively behind.

It is by now well understood that the virus is transmissable for several days before the appearance of symptoms. To be able to reopen without immediately heading back into a steep infection curve that would once again overwhelm ICU capacity, we must ramp up testing dramatically with a target of many millions of *daily* tests. Ideally people could test themselves at home and/or at work several times a week with results in minutes.

Is that just crazy or can we get there? Yes because the required technology already exists, we just need to approve it and manufacture it at scale. Here are just two examples of machines, the MicrosensDx and the Accula. There are many more startups and established companies that have tests and there are fascinating proposals for super high throughput cheap testing.

We need to take three steps right now: (1) approve lots of these test immediately and (2) manufacture at scale and (3) monitor ongoing results. Here is how to accomplish this

Step 1: Approval

The regular FDA approval process needs to be completely sidestepped. Instead we either go a decentralized route, allowing states to approve their own tests, or we put together an approval task force recruited from the leading test scientists. The goals for that group should be to greenlight dozens of tests within the span of days and to then define follow-up reporting requirements to enable ongoing monitoring (step 3 below).

Step 2; Manufacture

There are lots of components to these tests and they are not easy to manufacture. But there is definitely manufacturing capacity that could be repurposed. To do so quickly I believe we will have to invoke the Defense Production Act as well as be willing to spend a lot of money. Every dollar spend on testing will unlock a multiple in economic activity so this is among the best money we can spend, even if some of the tests we buy don’t work well (or maybe not even at all). Having more different tests approved in Step 1 means we can lean into manufacturing much harder with a portfolio approach.

Step 3: Monitor

Finally we need to set up a reporting and stats infrastructure to monitor the performance of the tests as they are deployed so that we can hone in on the ones with the best sensitivity and specificity. With the right reporting protocols we will rapidly learn what works well.

Essentially this amounts to reversing the normal approval process which takes a long time with the goal of having only high quality tests in market. Here we want to optimize for speed and massively over allow, then pull back later. Because this is so essentially a 180 from the FDA’s normal operations we need a special one-time panel on this.

Governors should be exerting public pressure right now calling for this and if it has not been put in place by end of next week they should proceed on their own.

Albert Wenger is a partner at Union Square Ventures (USV), a New York-based early stage VC firm focused on investing in disruptive networks. USV portfolio companies include: Twitter, Tumblr, Foursquare, Etsy, Kickstarter and Shapeways. Before joining USV, Albert was the president of through the company’s sale to Yahoo. He previously founded or co-founded five companies, including a management consulting firm (in Germany), a hosted data analytics company, a technology subsidiary for Telebanc (now E*Tradebank), an early stage investment firm, and most recently (with his wife), DailyLit, a service for reading books by email or RSS.