The Virus and the Algorithm

It's (not) alive - flat and fragmented - antivirals

Photo by engin akyurt on Unsplash

It’s (not) alive

Viruses are bare-boned biology. So bare-boned that we don’t even know if they’re alive.

Basically, a virus is genetic material that is stuffed inside a protein shell (which occasionally wears a nice fatty coat). A virus doesn’t have a metabolism. It’s inert until it meets a host cell.

In a previous post about viruses and how they made us human, I ran a poll asking whether or not they’re alive. The answers mirrored the lack of scientific consensus: 45% of respondents said yes, 55% said no.

We don’t know if viruses are alive, and the answer will depend as much on science as on philosophy, as much on data as on semantics. The question of aliveness, however, is not the only mystery about viruses. Their origin is just as much a question mark.

We don’t know where viruses came from.

Most of our current ideas about viral origins fit within one of three broad hypotheses:

• Virus-first: Viruses are relics of an earlier form of life, pre-cells. But… viruses need host cells, so we’re stuck with a cellular chicken and viral egg problem.

• Reduction: Viruses are derived from single-celled organisms and simplified through evolution, as often happens with parasites. But… we don’t know of any transitional forms between cells and viruses (yet), and the parasites that we know have evolved from single-celled ancestors retain some cellular components that viruses lack.

• Escape: Viruses are escaped mobile genetic elements (plasmids, jumping genes…). But… how did they find or develop their protein coat? How did they develop the tools to sneak back into host cells?

We don’t know if viruses are alive and we don’t know where they came from. What we do know is their modus operandi. It’s a simple algorithm, really: Break into host cell, usurp host cell, replicate, and repeat.

From BioRender.

Metaphor in 3…

2…

1…

Flat and fragmented

In our modern online content ecosystem, a snippet of said content (such a boring, ubiquitous, blandifying term) is inert until someone interacts with it.

Sounds familiar.

Content that goes viral also requires an algorithm — a recommendation algorithm.

An interaction - like, comment, share - wakes up the piece of content by bringing it to the ravenous attention of a recommendation algorithm. The aspiring viral content gets pushed into individual feeds (or ‘host cells’?) of whatever platform it’s on. There, it pushes away other content (or ‘usurps the host cell’?) and clears space for other, similar pieces of content (or ‘semi-replicates’?)

I’m simplifying to the point of snark (an artisanal specialty of this newsletter). Different platforms use different algorithms, recommendation algorithms get tweaked all the time, and they take into account a lot of factors beyond simply ‘interaction’. But in the end, engagement is key.

Welcome to the broken land of the common denominator.

In his (recommended) book Filterworld, journalist Kyle Chayka tracks the rise of the now-ubiquitous recommendation algorithms and how they have reshaped culture and the way in which we consume it.

Because the algorithms tend to prioritize content that performs well across a broad audience, we’re exposed to similar-looking art, music, fashion, and interior design — an aesthetic monoculture, Chayka calls it. To make matters worse, artists and creators start optimizing their work for ‘algorithmic success’, leading to formulaic structures like songs with hooky intros in the first fifteen seconds for Spotify’s skip-happy algorithms. Chayka puts it nicely when he writes,

The feed structure also discourages users from spending too much time with any one piece of content. If you find something boring, perhaps too subtle, you just keep scrolling, and there’s no time for a greater sense of admiration to develop—one is increasingly encouraged to lean into impatience and superficiality in all things.

Since the platforms that dominate our online ecosystem are global, cultural differences and flavors are flattened. A Gen Z user in Seoul and one in São Paulo dance to the same TikTok trends, watch the same Netflix series, or buy the same Amazon-recommended gadgets.

At the same time, algorithms fragment culture by personalizing feeds to individuals, so that everyone sees a different version of the internet, curated just for them. People may think they’re plugged into ‘mainstream culture’, but they’re really in their own filter bubble. The prioritization of engagement also leads to echo chambers that reinforce existing beliefs, which deepens polarization¹ by continually feeding users content that aligns with their preferences or biases, isolating them from other perspectives. Cultural cohesion cracks. In the words of Chayka,

All kinds of cultural experiences have been reduced to the homogenous category of digital content and made to obey the law of engagement, the algorithms’ primary variable. Any piece of content, whether image, video, sound, or text, must compel an immediate, albeit often superficial, response from the viewer.

So, while everything starts to look the same, we’re also less likely to see the same things. That’s the paradox at the center of Filterworld. Algorithms make culture feel both uniform and disconnected, flattened and fragmented.

And, because I like to tickle metaphors until they squirm, most viral content is like a cold; it makes you feel as if you’re face is stuffed with cotton balls, ruins your taste buds, and turns you into a sniffling mess.

There is no cure for the common cold.

Yet.

Antivirals

The common cold is not a single virus; it’s a term for a group of viruses (mostly rhinoviruses) that elicit similar symptoms.

The common cold viruses mutate rapidly, which is why it’s proven challenging to develop a vaccine. Beyond vaccines, the slippery little sneezers also evade antivirals² so far.

Yet, there are antiviral success stories. HIV is no longer a death sentence thanks to a combination treatment of antivirals, and Hepatitis B and C no longer turn your liver to mush and your blood to bile (poetic license there) because of antiviral medication.

Antivirals work by interfering with the viral algorithm; they either prevent the virus from infecting the host cells or they prevent the virus from replicating inside the host cell.

Just as the metaphor thinks it managed to slip away to safety, let’s pull it back in.

Can we prevent viral content from infecting our feeds? Not really, that’s the choice of the platform owners. As long as they choose to prioritize engagement above all else, virality will stay a target, a goal, and a mission for content creators. As watchers/readers, all we can do is choose which platforms (not) to engage with. We can isolate and ignore social media in an informational quarantine. But recommendation algorithms are no longer only a social media thing. Netflix. Spotify. Amazon. What we watch, listen to, read, and buy is increasingly mediated by the engagement metrics that inevitably lead to the primacy of virality.

But, perhaps, in a small way, we can try to prevent the replication of viral content in our feeds. Just like we can choose (not) to engage with specific platforms, we can choose (not) to engage with specific pieces of content that are pushed into our feeds. We can choose not to sneeze them along. This is not as easy as it sounds, of course. The platform recommendation features are designed and tweaked to appeal to the widespread, rapid-fire instinct for loudly asserting one’s opinion, as well as our hidden desire for a life free of friction.

“You don’t have to decide what to watch or read; the algorithm decides for you. Just sit back and stay on our platform a little longer…” [recommendation: ominous music]

The challenge is that the online content that doesn’t kaboom into virality all but disappears from our feeds in twenty seconds. It’s there, sure, but it’s covered by an endless torrent of formulaic, shallow bait that hopes to ride a viral wave. The very nature of endless, incessant scrolling leaves little time for nuance, ambiguity, and subtlety. Yet, these aspects of a text, song, or video can challenge your mind and inspire you to think deeper or differently about topics.

But we have to go looking for the nuance, ambiguity, and subtlety triplets; they’re rarely tailored for virality. Our feeds should represent our evolving experiments in taste and interests, not a torrent of clickbait with the sole purpose of hooking us. Or, as Chayka puts it,

We should talk even more about the things we like, experience them together, and build up our own careful collections of likes and dislikes. Not for the sake of fine-tuning an algorithm, but for our collective satisfaction.

Recommendation algorithms are not evil, but today, platforms use them to encourage mindless, shallow, and frictionless consumption by focusing on metrics of engagement. That way leads to a loop where we’ll get stuck in an endless repetition of new stuff that’s all the same.

Find the works and words that challenge you, because those are the last thing the recommendation algorithms want to give you.

Atchoo!

---

Thanks for exploring viral algorithms with me today. Is it hypocritical to want this post to go viral and let you know that clicking buttons helps with the algorithm? I non-algorithmically recommend clicking all of them.

1. The spread of misinformation in echo chambers can be modeled as… a viral contagion! Biology is awesome.

2. Antibiotics (which are effective against bacterial, but not viral, infections) do not - can not - work against the common cold. If your doctor pushes this, find a new doctor. Non-steroidal anti-inflammatory drugs can ease some of the burden and zinc may help you recover faster, but that’s about it in terms of dealing with a cold. Also, rest and hydration.