subjectivity as algorithmic impenetrability

Technology may be automating more and more things away from us, but it keeps bouncing off from our semantic core. The subjective realm is impenetrable to algorithms, and will always remain that way.

Technology enthusiasts in the West usually take this as a challenge, as if the goal of technology is complete replacement of humanity. But they are mistaken. As pointed out in an earlier post, they misunderstand how evolution works, and this misunderstanding parallels a paranoia that too has cultural origins.

• Western cultures break the mind-body duality in favor of body, and value the objective over the subjective. Since technology automates what is objective, it creates an insecurity. People naturally feel threatened and act defensive to protect what is meaningful for them.

• Western cultures value what humanity creates more than what creates humanity. Nature is viewed as an object that should be dominated, manipulated and subjected to the human will. Since technology is emerging through humanity, people naturally worry that it will treat its creators with disregard, the same way they treat their own creators.

Going back to the topic of this post, here are two great examples of how the subjective realm has proved to be impenetrable to algorithmic infiltrations.

Failure of Algorithmic Seduction: Amazon Case

When I think of creating desire, I think of my last and only visit to Milan, when a woman at an Italian luxury brand store talked me into buying a sportcoat I had no idea I wanted when I walked into the store. In fact, it wasn't even on display, so minimal was the inventory when I walked in.

She looked at me, asked me some questions, then went to the back and walked back out with a single option. She talked me into trying it on, then flattered me with how it made me look, as well as pointing out some of its most distinctive qualities. Slowly, I began to nod in agreement, and eventually I knew I had to be the man this sportcoat would turn me into when it sat on my shoulders.

This challenge isn't unique to Amazon. Tech companies in general have been mining the scalable ROI of machine learning and algorithms for many years now. More data, better recommendations, better matching of customer to goods, or so the story goes. But what I appreciate about luxury retail, or even Hollywood, is its skill for making you believe that something is the right thing for you, absent previous data. Seduction is a gift, and most people in technology vastly overestimate how much of customer happiness is solvable by data-driven algorithms while underestimating the ROI of seduction.

Eugene Wei - Invisible Asymptotes

Seduction is built on the mystique of the unfamiliar. That is why it is much easier to be captivated by someone whom you have just met. Data-driven algorithms on the other hand behave like people who know you for years.

Also seduction is a two-way process that unfolds dynamically over time. It involves tailoring a physical form around innate desires which are revealed through interactions. Advertisements on the other hand are unspontaneous one-way interactions.

Failure of Algorithmic Aesthetics: Netflix Case

Netflix came to a similar conclusion for improving its recommendation algorithm. Decoding movies’ traits to figure out what you like was very complex and less accurate than simply analogizing you to many other customers with similar viewing histories. Instead of predicting what you might like, they examine who you are like, and the complexity is captured within.

David Epstein - Range (Pages 111-112)

Algorithms can analyze only the explicit syntactic interactions between humans and make indirect inferences about the implicit semantic processes going on within. Since aesthetic judgment is a heavily semantic (subjective) affair, algorithms are better off trying to understand whose aesthetic taste is closer to whom, rather than directly making the judgment calls themselves. In other words, we discover great new songs and movies through each other. Of course, user interfaces hide away this relational complexity and we end up feeling as if the algorithms are making recommendations on their own.

truth as status quo

We now have the science that argues how you're supposed to go about building something that doesn't have these echo chamber problems, these fads and madnesses. We're beginning to experiment with that as a way of curing some of the ills that we see in society today. Open data from all sources, and this notion of having a fair representation of the things that people are actually choosing, in this curated mathematical framework that we know stamps out echoes and fake news.

The Human Strategy

Fads and echo chambers provide the means to break positive feedback loops (by helping us counter them with virtual positive feedback loops) and get out of bad equilibriums (by helping us cross the critical thresholds necessary to initiate change). Preventing illusion is akin to preventing progress. Every new truth starts with untruth. Future will be in conflict with today. Today’s new reality is yesterday’s false belief.

It is startling to realize how much unbelief is necessary to make belief possible.

Eric Hoffer - The True Believer (Page 79)

We are constructors of our social world as well as receivers.That is why companies like Facebook should never be involved in this war against “fake news”. Truth is inherently political. Algorithms for sniffing it out will inevitably end up defending the status quo.

spectrum of scalability

There are only two types of businesses that really matter, namely those that scale perfectly (totally inhuman) and those that do not scale at all (totally human).

For instance, both content production and information-heavy technology businesses scale perfectly. (It is costless to reproduce films and software.) Spotting and nurturing promising artists and technology entrepreneurs on the other hand do not scale at all.

Greatest returns in business come from highly scalable (and defensible) businesses, but the most vital ingredient in building such businesses is talent. In this sense, talent management is only one-step away from scalability, and that is exactly why its return profile is extremely nonlinear, mimicking that of scalable businesses.

---

Media and technology worlds are structurally quite similar in the sense pointed above. But then how could media companies have been so slow and inept at crafting a legitimate response to the tech companies creeping into their domain?

Answer is very simple. Although media content scales perfectly as software does, it does not evolve after it is born. For instance, once a film is produced, it is done. Software on the other hand is born immature and goes through an evolutionary design process which slowly settles into an equilibrium. Media companies do not know how to guide this evolution. That is why they are prudently waiting for the equilibrium to emerge before making a move. (Think of Disney’s late response to Netflix.)

---

Investing only in totally scalable and totally unscalable businesses is an example of the barbell strategy popularized by Nassim Nicholas Taleb.

Universities are embodiments of this strategy. They pool their resources into two buckets: facilitating research and teaching students. Research is a form of content production and teaching is a form of talent nurturing. (They are nurturing future researchers.)

order out of noise

Firehose of falsehoods and diverse contradicting statements distributed through many different channels create completely noise environment. Through this noise everyone weaves their own story. That of course happens to be the one that they want to believe in the most. Since most of these facts are pro government. The end result is a diverse array of personalised strongly held pro government stories!

- Networked Propaganda and Counter Propaganda (Jonathan Stray)

Noise contains all possible structures as subsets of itself. Since we are naturally predisposed to recognizing patterns (i.e. compressing information), it can act as a "mirror" reflecting our complex web of pre-existing cognitive short-cuts and biases back to ourselves.

For instance, we can not help but hallucinate ordered structures out of visual and auditory noise. (e.g. pareidolia) I myself experienced a few such bizarre moments, involuntarily scalping out complex musical pieces out of random environmental noise. These episodes were inspiring but also quite intimidating. (I felt as if my unconsciousness accidentally leaked into my consciousness. Is this why some people enjoy listening to noise music?)

If you are willing to experiment with drugs like LSD, you can simulate such experiences using the brain’s own background noise:

Sometimes patterns can arise spontaneously from the random firing of neurons in the cortex — internal background noise, as opposed to external stimuli — or when a psychoactive drug or other influencing factor disrupts normal brain function and boosts the random firing of neurons. This is believed to be what happens when we hallucinate.

- A Math Theory for Why People Hallucinate (Jennifer Ouellette)

---

Orwell feared those who would deprive us of information. Huxley feared those who would give us so much that we would be reduced to passivity and egoism. Orwell feared that the truth would be concealed from us. Huxley feared the truth would be drowned in a sea of irrelevance. Orwell feared we would become a captive culture. Huxley feared we would become a trivial culture.

- Amusing Ourselves to Death (Neil Postman)

Jonathan's observation above is enlightening in the sense that it situates Orwell and Huxley as two ends of a single spectrum for controlling public opinion.

• Orwell: Dictate people by spoon feeding them your point of view

• Huxley: Incapacitate people by drowning them in complete noise

Both strategies backfire in the long run:

• Orwellian regimes can quickly unfold when information from the outside world starts leaking in.

• Huxleyan regimes shut people completely away from serious matters and this can set in motion a cultural backlash.

The best strategy is a mixed one: Drown people in variations of your point of view so that

• the information environment stays rich enough to make people feel as if there is an open public discourse

• the proportion of trivia among the narrations in circulation stays below the critical threshold that can set in motion a cultural backlash.

This way, all the stories people weave out of the pseudo random environment you created will stay close to your point of view.

Note that, as technology progresses and information flows at greater speeds, it becomes harder and harder to maintain an Orwellian regime as opposed to a Huxleyan one.

thoughts on shopping systems

A system with one seller and many buyers is called a shop (or an auction if only one good is being sold). When many shops are brought together, the resulting system is called an aggregator (or an auction house). If sellers can become buyers too, then it becomes a marketplace.

If seller side is closed to new entrants (which is always the case in the physical world), then what you have is something like a shopping mall. (Openness creates discoverability problems for sellers who are always willing to pay a rent for compensating the costs of maintaining a closure.)

If buyer side is closed to new entrants, then what you have is something like a shopping club. (Sellers will be willing to offer discounts to have access to such closed-off demand sources.) Shopping clubs collapse if the number of sellers is increased over time and sellers are essentially left competing as in an open system. This sooner or later happens due to competition among shopping clubs who feel pressurized to hunt for more customers and flex their closures. Of course, the irony is that when a shopping club is truly successful and every buyer and seller is in it, then it becomes the whole economy and paying a cut to the club becomes pointless for both buyers and sellers. 

Groupon tried to circumvent this fate by concentrating the demand and spacing sellers across time in return for a deep discount from each seller. It tried to achieve the paradoxical goal of having seller variety without seller competition. Since Groupon buyers moved from deal to deal and had very little loyalty, the time dimension got behaviorally collapsed and sellers again ended up competing with each other.

Groupon could eventually be brought down to its knees (by the sellers who understood what was going on) because it did not appeal to all potential buyers. (A lot of people want to be able choose from a wide variety of options rather than being dictated to a few randomly fluctuating options.) In other words, sellers had something to fall back on.

Amazon on the other hand can only be brought down by regulators because it appeals to all potential buyers. Sellers have nothing to fall back on. (Amazon does not demand a discount from sellers but exposes them to intense competition, which at the end of the day amounts to a not-as-deep-but-a-more-permanent form of discount.)

Some further thoughts:

• Only oligopolistic markets will be able to resist Amazon which started off by killing the competitive bookstore market. By becoming literally the only bookstore in town, Amazon now faces a strong resistance from a coordinating set of publishers from whom it now wants to buy directly. It can either buy its way up the supply chain or become a publisher itself like Netflix did. Former strategy will be expensive since publishers selling their businesses will be asking for very high premiums for defecting and will make Amazon pay the present value of all their future fat profit margins. Latter strategy is tough because of the same reason why Spotify will eventually fail: Creating a music star (similarly a bestselling author) is a much more complicated process then creating a hit movie. (Spotify’s profit margin is tiny compared to that of music producers and it does not seem to scale neither.) Talent discovery is stochastic and talent nurturing is difficult, and neither scales well. (Also there is a lot of consumption of old songs and books - which is not true for movies. So the incumbent producers with deep archives have a competitive advantage.)
• It is easier to sustain a shopping club in a physical environment where one can maintain a locally acceptable level of competition among the club-member sellers by sampling across all shopping malls. (In a virtual environment, there are no distances and competition is always a click away.) That is why shopping clubs usually end up housing only non-virtualizable experience goods like dinners in fancy restaurants etc.
• In a physical marketplace, buyers can not hide. A carpet seller in a bazaar can grab you by the arm and drag you to his store. In a virtual marketplace, on the other hand, all he can do is to harass you with the cooperation of the marketplace itself by using the allowed advertisement mechanisms. So, in some sense, you forfeit your right to be protected from harassment to the marketplace in return for the promise to build some norms and predictability.

linearity of podcasts

I can not stand podcasts. Linearity of the medium is intolerable.

• Talking is OK since it is at least a two-way interaction.
• Watching is OK since each time slice is enriched with an image. (Images allow fast and non-linear transfer of information.)
• Reading is best since one can scan a text in any direction and at any (varying) speed.

Of course, the reason behind my impatience is that most information is either trivia or noise.

korku ve duyarsızlaşma

Kelimelere kıyasla imajlara karşı çok daha hızlı duyarsızlaşıyoruz. Mesela cehennem kelimesi herkeste kendine özgü bir korkunçluk yaratabiliyor, fakat herhangi bir görsel cehennem tasviri çok hızlı şekilde etkisini kaybedebiliyor. Hatta sadece bir jenerasyon sonra komik bile bulunabiliyor.

Bu sebepten dolayı dinler görsellikten uzaklaşıp metne ve soyutluğa sığınmışlardır. Dinlerin temeli korkuya dayanır, fakat insanoğlu süreklilik arz eden korkularını normalleştirerek yenebilme özelliğine sahiptir. Tek yenemediği korku dinamik, kendi kendine yarattığı korkulardır. Metin bu korkuları çekip çıkartabilen ve sürekli canlı tutabilen basit ama çok güçlü bir teknolojidir.

büyüme hırsı ve karaktersizlik

Yatırımcılar büyüme görmek isterler. Vizyon çok da umurlarında değildir. Rakibiniz büyüyor mu, şu an sizin de hemen büyüyüp ona yetişmenizi beklerler.

Twitter'ın ölümü bu baskıdan ötürüdür. Facebook gibi herkese hitap etmek adına ürün zamanla karaktersizleştirildi ve esas kullanıcılar platforma küstürüldü. Oysa Twitter sosyal medyanın belki de en değerli ve eğitimli kitlesine hitap ediyordu.

Sosyal ağlar da insanlar gibidir. Herkesi mutlu etmeye çabalayan biri hakkında ne düşünürsünüz? Ben şahsen karaktersiz olduğunu düşünürüm.

Snapchat'in çizgisini bozmamasını, yetişkinlere de hitap edeceğim diye çırpınmamasını son derece vizyoner buluyorum. Uzun dönemde sadece bu şekilde kitleleriyle ve duruşlarıyla ayrışan platformlar ayakta kalacaklar.

facebook as a cultural hype

There are hypes of different periodicities. Cultural shifts are among the slowest. (Quantitative traders are well aware of the fractal nature of cycles.)

Mark Zuckerberg is a serious man. He does not like hype. He prefers utility over coolness.

General opinion is that Facebook owns "the social layer" and Mark is connecting the world. But this is complete bullshit for anyone who knows a little bit of sociology.

---

Facebook does not own the social layer. It is the social layer who owns Facebook.

Mark may detest coolness, but the early adopters of social platforms are always among the youngsters and the youngsters care a lot about coolness. In fact, Facebook owes its wildfire growth among university campuses to its cool beginnings.

Now the youngsters have shifted to other platforms and Mark is going crazy realising that he will not be able to buy all these platforms off. (Public markets serve an enormously important social role: For a company like Snapchat, going public is the only option for realising value for its investors without submitting itself to a greater behemoth.)

Youngsters disrupt the status quo set by their elders.  (It is their sociological role to do so.) They are catalysers of change. They challenge for the sake of challenging. They test which social structures deserve to survive by shaking them to their cores.

And make no mistake, Facebook will crumble too.

---

Mark is not connecting the world. It is the opposite. Mark is disconnecting the world by virtualising the already-existing real friendships. 

The very word "friend" has lost its meaning. We have become estranged from each other, turning into wanna-be celebrities broadcasting to our own friends and anxiously building fake images. We are now meeting up less often in the real world because we are meeting up more often in the virtual world.

Mark is riding a massive cultural wave and creating a positive feedback loop that is accelerating the death of this very wave. The hegemony of self-exhibitionist narcissistic openness will soon come to a halt. (Early adopters have already fled this trend.) The pendulum has shifted too much in the direction of openness and atomistic individualism. Now, along with the massive conservativeness wave in politics, it will shift back to reservedness and community-oriented holism.

True. The most tangible network is our friendship network. But that should not make you think that Facebook is built on better foundations than something like Twitter which is built on interest-based networks. In fact, the opposite is true: It is platforms like Twitter which serve a real need rather than Facebook.

Interest-based communities existed before Facebook and will continue to exist after Facebook. The true strength of the internet has lied in its ability bring out the long tail in everything. Thanks to the internet, a substantial number of people can gather around a very niche topic and buy stuff related to a very niche interest, stuff which would have never made it to the physical shelves due to the diffuse geographic distribution of the demand.

Soon people will realise that the only real value offered by Facebook lies in its ability to connect us to our long-lost friends, and Facebook will be used primarily as a catalogue of expired friendships and our long-lost friends will upgrade from a mass grave to a proper cemetery.

So Mark, why so serious? 

an information consumption guide

The amount of information created daily is huge. You need to set up some proper filters to preserve your sanity. Here is my personal advice:

- For your deep readings, follow a professional human curator who has similar interests as you do. (You need someone whose actual job is curation. Pay if necessary.) Algorithmic machine curators are horrible at spotting high-quality articles, because they use inputs generated by the general public and the general public has absolutely no taste or depth.

- For your news reading, follow an algorithmic curator. Machines are much better at quickly scanning large databases and bringing together a comprehensive and timely feed. You will not experience much downside neither since the importance of a piece of news is highly correlated with its popularity.

Note that, in both cases, other human readers are utilised to do the most mechanical part of the job for you. This means that if these advices are followed by everyone, then they will no longer work. Machine curators will have no signals to work with and human curators will wait for others to curate.