Everything is getting better and worse at the same time

“The truth will set you free, but first it will make you miserable.”

It’s late at night. You shuffle to the kitchen for a snack. Your hand fumbles briefly for the light switch, and… roaches! They quickly scatter, but now you’ve seen them. You know they’re there. From now on, you can’t not think about the roaches in the kitchen. It’s a shame, too, because you always thought of yourself as a neat person with a clean kitchen. But now that image has been ruined. The good news is, now you have better information about the world. You DO have roaches in your kitchen, and you can start to do something about it.

The Roach Reveal story is how I think about one of our modern patterns of experience: everything is getting better and worse at the same time. These days we’re getting ever more powerful sensors that show us the virtual vermin that lurk beneath the veneers of civilization. We’re seeing roaches we never knew were there. It feels like we’re being hammered endlessly by terrible news. But look closer: the roaches were always there. And now we can start to do something about it. And often it’s only when we realize just how bad it is that we finally decide to take action.

Related to this topic, I want to convince you to read The Alignment Problem by Brian Christian. This book addresses the dangers associated with machine learning, smart robots, and clever algorithms of all kinds. The title refers to a subtle and disturbing fact: it’s strangely difficult to tell a computer what you want it to do. You can tell it to do something that is what you THINK you want it to do. And then it will, with dizzying speed and precision, set about making you miserable, all while doing exactly what you asked it to do. This is often described in terms of deal-with-the-devil jokes: tell the computer to stop people from getting malaria, and it will murder everyone before they can catch malaria. “I solved your problem!” says the robot. “You killed my family!” says the programmer. That’s the alignment problem. It’s not a joke.

One of the topics in the book is algorithmic bias. Suppose you want to teach a robot to hire good employees. Or decide who should get a loan. Or maybe decide who should be paroled from jail. After you implement some pretty straightforward machine learning, you are almost certain to be disturbed by the results. The computer has learned from a horribly biased society. What else could it learn from? Naturally it mirrors back to us racism, sexism, and xenophobia.

Machine learning is the kitchen light. It’s illuminating the roaches that have been crawling through our brains and institutions for hundreds of years. Switching on these algorithms feels like a massive step backwards. We are in danger of encoding extraordinarily efficient prejudice. But the book comes with good news too. No matter how unhappy the kitchen light first makes you, it will also help you solve the problem. Seeing how biased our algorithms are, we can set about attacking the root cause. The root cause is not the kitchen light.

Our computers can teach us to be better humans.

How Do You Spell Vaccine?

This post is a follow-up to one that I did back in February: A Modern Magical Spell. There I was ruminating about the fact that, whereas old-school vaccines included little chunks of the virus itself, the Pfizer and Moderna mRNA vaccines are instructions, blueprints, that tell your body how to make those little viral chunks. They are, in effect, cellular DIY projects.

It sounds like a minor point, but the difference is huge. Sending messages is easier and more flexible than sending the thing that the message encodes. Why send cookies if you can just send the recipe? Why send a string quartet if you can just send the sheet music? Blueprints are cheaper than bricks.

Anyway, I’m writing about the same thing again because we now know exactly what text is in the vaccine. Text messages are convenient for a number of reasons, and one is that they’re easy for us to read. Think about this: the syringe that pokes you in the arm is a bottle with a message. Uncork the bottle, unroll the message, and you can see, you can just read off, exactly what protein sequence the vaccine codes for. Some folks at Stanford have done exactly that: Stanford Scientists Post mRNA Sequence for Moderna Vaccine on Github. They didn’t want to get in trouble for stealing anyone’s shot, so they did the equivalent of sifting through the trash for valuable documents: “RNAs were obtained as discards from the small portions of vaccine doses that remained in vials after immunization.”

Pfizer didn’t want to publish this information… only they did publish it. They published millions of copies into little vials and distributed them across the country. For my previous piece, I made a guess as to the sequence that was used for encoding the spike protein. No surprise, my guess was wrong. But the good news is that the actual answer is posted on GitHub. Check out the fancy title: Assemblies of putative SARS CoV2 spike encoding mRNA sequences for vaccines BNT-162b2 and mRNA-1273.

Do you want to know the actual recipe for that BNT-162b2 vaccine? The actual text that would be injected into your bloodstream? Here it is.


There it is! That’s the payload. That’s what the fuss is all about. It gives me a thrill to look at it. It may look like a mess, but the ribosomes in your cells can read it like a recipe. So we might say that you can’t read it, but “you” can. Because you can. And you will. And it might save your life.

As biology becomes more and more of an information science, many strange and wonderful things will become possible. We’re still at the “Mr. Watson, come here, I want to see you” stage of communication.

Outsourcing Humanity

A cooking pot is an outsourced stomach. It shifts much of the burden of digesting food from inside to outside your body. Because of this, your real stomach doesn’t have to work so hard. This was a big deal for early humans, allowing them to make more effective use of available food.

Consider the shift that followed. Over millions of years, animals slowly evolved stomachs that could digest their preferred diet. This internal stomach is encoded in DNA, in the genome. Then, one fine day, fire-savvy, cook-capable humans came along. The job of the stomach could now be shared between the pot and the belly, which means that the combined “big stomach” is now a joint venture encoded by both genetic and cultural DNA. Across many generations, the biological stomach can now “relax”. Not only does it do less work, it can even give up some its ability to do the hard work of digesting raw food. Why bother, as long as you have a microwave handy? In this sense, you become a hybrid biological-cultural construct, as dependent on the cultural knowledge of cooking and pot-making as on your own self-constructing DNA. How long would you survive naked in the wilderness?

What’s true for cooking is true for many things. Knives outsource teeth. Clothes outsource skin. Glasses outsource vision. We make our tools and our tools make us. It’s an obvious statement and a profound one, particularly when you consider medicine.

Across the millennia, pathogens have sculpted our genome. Traits that have helped previous generations survive the ravages of malaria, plague, and tuberculosis get built into our genes. We have the tracks of thousands of pandemics etched into our genetic memory. But now, just as with our stomachs and our teeth, we are outsourcing, externalizing our immune system. This modern plague is reshaping not so much our genetics as our global cultural medical apparatus. Pfizer, Moderna, and AstraZeneca can do in one year what a deadly plague might have needed a hundred years and fifty million lives to impart to our genes. Our genes don’t need to improve so long as our medical care does. Put another way, our biological immune system is now free to get weaker, so long as biotechnology can carry its load. It doesn’t take a great leap of imagination to see us, across several hundred years, fading into our sustaining machinery. With it, we are gods. Without it, we are shrieking infants.

But the coming world will also give us the ability to upgrade not only our machines, but also our genes. Here’s one simple example of what that might look like.

As noted, traits that are not needed for survival tend to fade. This is why you can’t smell as well as a dog. You share with dogs a lot of the same DNA for sensing specific odors, but in humans, much of this DNA has been damaged and rendered inert. If a few bulbs burn out in your smell-o-nator, who cares? You can still have healthy kids. Over time, humans have lost the cunning of the canine nose. But that doesn’t have to be true forever. We can not only build outsourced spectrometers and mechanical noses, we will also be able to retrofit our DNA. This will be a dangerous and subtle skill to learn, but be certain it will come. In the future, wealthy parents will be able to endow their offspring with, among other things, the smelling capabilities of dog. I don’t know if that’s a good idea, but I’m telling you, it’s coming.

Would you pay to give your children a superhuman sense of smell? Or the ability to see colors no human ever saw?

Strava Heatmaps on the Cheap

Strava has been wildly successful at getting people to surrender their privacy in the name of bragging about fitness. You log your activity on the Strava website, and from this you can make personalized maps. You can even share this information up into a global heatmap database, from which Strava can show you where EVERYBODY has been running and biking and snowshoeing and paragliding and dolphin-riding and so on.

It’s a lot of fun to play around with. Shown below, we’re zoomed in on Cambridge and Boston. It’s no surprise that busy streets are busy, but I like how you can see exactly where the gaps are under the Harvard Bridge. The Charles River looks like a highway. Which, for the purposes of sport (and fish), it is.

I was jealous of these personal activity maps, but not jealous enough to spend the money on Strava. That was when I remembered something important, something that I try to remind myself on a regular basis: whenever you imagine a potential app or web service, you cause somebody to retroactively make it for you. That is to say, your idea isn’t original, so someone has already done it for you. All you have to do is say “All-seeing Google, show me the thing that does X.”

As in the sentence: “Show me the site that will let me make my own Strava-like heatmaps for free.”

I know this is kind of obvious, but sometimes I forget to actually make a pointed request for something that is vaguely floating around in my mind. I’m kind of thinking about it (“gee, that Strava heatmap is cool…”), but I don’t articulate the specific question. This means that the app never gets retroactively created by internet elves on my behalf. You see how it works? The magic never happens.

Anyway, I was not disappointed. I was led very quickly t0 dérive by Erik Price. It’s free! It’s open source (MIT license)! It’s very good, and it works like a charm. Huzzah! How I love the new millennium. And thank you, Erik.

You just bring your own GPX files to the dérive website, and then drag them onto the browser. All the work happens in the browser, so you’re not actually sending any of the files back up to the site.

This pandemic has given me a lot of time to walk around my hometown, and my completionist tendencies coupled with a nice route-planning app (Footpath) have given me the motivation to explore.

How it started:

How it’s going:

You’d be amazed how many roads are close to you that you’ve never been on.

Washing Machines Large and Small

Here is a picture of two different factories that do the same thing: turn atmospheric nitrogen into ammonia.

The one on the left is a gigantic Haber process plant. It requires fantastic amounts of heat and pressure to do its work. Plants like this consume 3-5% of the entire world’s natural gas output. The one on the right is an enzyme: nitrogenase. It can do its work in the roots of the clover growing in your front yard, all with very little risk of an explosive industrial accident.

This vignette shows one of the biggest lessons that biology has to teach us: how to do useful chemistry with much less effort. Which has all kinds of benefits, as you can imagine.

People hear the terms “biotech” and “synthetic biology” and it’s no surprise that the pharmaceutical industry is what springs to mind. After all, we live in the shadow of The Virus. I don’t want to diminish the miracles being worked by the mRNA hackers at Moderna and Pfizer, but I do want to point how much synthetic biology is already doing for us in non-medical settings. This article about Frances Arnold provides a couple of real world examples.

Here’s a laundry-related story.

Up to 80% of your washing machine’s energy use comes from heating the water. But cold water wash is the way to go these days, thanks to engineered enzymes. Your laundry detergent contains eight or more enzymes, all designed to do the work that hot water used to. So do yourself and the planet a favor: set your wash cycle to “cold” and let engineered enzymes save energy and expense with every load.

I like this example because it highlights the same big machine/small machine dichotomy I described above. You can do a lot of work at high heat with a big machine (the washer), or you can do much less work in cold water with many small machines (enzymatic molecules). It seems like a cute example, but it’s real and it matters. There are many more stories like it.

So set that cycle to cold! One day we’ll look back and laugh at how profligate we were with our energy. Or maybe we’ll cry. Or maybe we won’t be here because we played with fire for too long.

Oooh, that got dark. Let’s go with laugh.

The Virtues of Incrementalism

I came across a cartoon the other day. Maybe you’ve seen it too. This is my re-drawn version of it, to give you the general idea. A guy pulls up to a burger place in a big gas-guzzling car. Smoke pours from the tailpipe. He’s about to take his order, but then he draws back: “No plastic bags for me. I care about the environment!”

Big Car Burger Guy is such a dope! He’s an easy target for ridicule. But ultimately I think making fun of him is unhelpful. I want Big Car Burger Guy to swear off plastic. Good for you, Big Car Burger Guy! If that’s all he can do today, then that’s okay with me. He’s moving in the right direction. I hope he’ll do more tomorrow, but I guarantee you that if I make fun of him today, it won’t lead to a better outcome tomorrow.

Let’s go back to the cartoon. It’s funny because if Big Car Burger Guy really cared about the environment, he would… what? Well, let’s see. He wouldn’t drive a car powered with fossil fuels. Actually, he wouldn’t drive a car at all. He’d be on a bike. No, he’d be walking. And he wouldn’t eat burgers. He would eat only locally-sourced vegan food. That he grew himself. Without fertilizers.

Where does it stop? How much change is enough? Who gets to decide who is virtuous and who is ridiculous?

I think of this as all-or-nothing reasoning vs. incrementalism. I know we’re just talking about a cartoon here, but it’s important, because snarky hot-take ridicule is a serious cultural problem. All-or-nothing reasoning says that you either make a great big change or you’re a fraud. But it leads to an unstable situation. We know what lameness is, but it’s hard to say where virtue begins. If I do X, will I still be shamed? All-or-nothing reasoning stops people in their tracks. And it drives away the very people you need to recruit.

Incrementalism, by contrast, is a recipe for action. No matter who you are, no matter where you are, you can always do something. Incrementalism tells you that every little bit helps. Because manifestly, it does! Make a small improvement today. And tomorrow make another one. And then another one the day after that. I’m often mystified as to why we discount incrementalism so ferociously, given that it’s essentially the only way things ever do change. It seems like a flaw in the human operating system.

This tweet is a good comment on what I’m describing.

All I’m asking is that you give incrementalism a little bit of a chance. See if you like it. Maybe you’ll come back for more tomorrow.

Source: ncse.ngo

What Do You Do About Unhappiversaries?

My wife was diagnosed with pancreatic cancer on the day before Thanksgiving in 2018. It made for a somber meal the next day. In the two years since then, my stomach has tightened just before Thanksgiving. I call occasions like this unhappiversaries. Anniversaries of trauma. Maybe it’s the death of a parent. Or the motorcycle accident. Or the day the divorce became final.

We all have unhappiversaries. My question is: what do you do about them?

One answer might be to ignore them. If you can forget them entirely, my hat is off to you. But that’s generally not an option. The memory will push its way in, so the options become to either reject or acknowledge.

I once knew a woman who, along with a friend, was in a terrible motorcycle accident. She spent months in and out of hospitals, her face held in place with pins and plates of steel. We spoke after the anniversary of the crash. I asked her, had she wanted to revisit the site? There was no question. Yes. Something drew her back. She and her friend went to the spot where it happened, were there at the very moment that it happened. And they sat on the side of the road, and they just looked at it. After some time had passed, and without exchanging a word, they drove away. She couldn’t explain why it made sense to go there or why it helped. It hadn’t removed the pain, but it helped to see, to contemplate this spot in her personal geography, this mighty bend in the river of her life.

You might be tempted, instead, to spend the day thinking about anything else, crowding out the memory with noisy distractions. But I’m on the side of acknowledging the darkness. It won’t be ignored. I sometimes personify it with a question: Have I learned what you would teach me? Nothing teaches sympathy like suffering. But you have to listen.

Tragedy shapes us more than joy. Our scars are what distinguish us, after all. My father, a psychiatrist, once put it like this: we are the sum of our imperfections. If you are to be a compassionate friend to your own lonely self, then you must appreciate the landscape after the storm has passed. It tells the story of the storm. It tells the story of you.

How do you observe your unhappiversaries?

A Modern Magical Spell

Double, double toil and trouble;
Fire burn and cauldron bubble

Everyone knows what a magic spell is. Say the words and things will happen. With the right incantation, you may see the future. Harm your foe. Ward off evil.

You generally need more than just words to make the magic happen. You may need a bubbling cauldron and some special ingredients.

Eye of newt and toe of frog,
wool of bat and tongue of dog.

I want to tell you about a modern charm that is, in every particular, a real and true spell of protection. It is written on a parchment small beyond seeing, rolled in fat, sprinkled with sugar and salt, then doused in an icy bath. Fragile and delicate, it must be handled with extreme care. Even gentle warmth will warp its power. But in the hands of an adept, it will shape the course of nations.

Cool it with a baboon’s blood,
Then the charm is firm and good.

When you are ready, the text is drawn from its freezing cask, streaming fog. An experienced acolyte sits near you and administers the charm, whispering to the very cells inside your body.

And this is how the spell begins: “M F V F L V L L P L V …”

Here is the rough translation:

Know this evil. Mark it well.
It comes for thee, in thee to dwell.
It comes to choke thee in thy sleep.
Choke it first, thy soul to keep.

Please hear me when I say that all of this is true. The spell is an mRNA vaccine, such as the ones created by Pfizer and Moderna to ward off the SARS-CoV-2 coronavirus, better known as COVID. It’s a description (literally “spelled out”) of part of the virus. In magical terms, it’s the tooth of the dog that might bite you. The letters in the text above are from the protein sequence of the very spike protein that punches a hole in your throat. These are the fingers that prize open the windows of your lungs. “Corona” means crown, and these are the spikes on the crown. This is what the vaccine is instructing your immune system to beware of.

So far, I’ve been using the old language of magic. Here is a more modern description. It still has a lovely incantation-like bounce to it, don’t you think?

The Moderna COVID-19 Vaccine (mRNA-1273) is an mRNA vaccine candidate against the novel coronavirus SARS-CoV-2 encoding for a prefusion stabilized form of the spike (S) protein, a class I fusion glycoprotein analogous to influenza haemagglutinin, respiratory syncytial virus (RSV) fusion glycoprotein (F) and human immunodeficiency virus gp160 (Env), and which is the major surface protein on the coronavirus virion and the primary target for neutralizing antibodies.

I want to give you a glimpse of how the message is conveyed. Because it really is spelled out like text on tape. And incidentally, this is why the technique holds great promise for the future. We don’t need to grow our vaccines anymore. We can just type them on a tiny scroll. All you need is a nanoscopic typewriter with a keyboard four letters wide.

The starting sequence in mRNA looks like this.

atgtttgtttttcttgttttattgccactagtc ...

This message gets encoded by your own cells to build a protein out of amino acids.

atg = M = Methionine
ttt = F = Phenylalanine
gtt = V = Valine
ttt = F = Phenylalanine
ctt = L = Leucine
gtt = V = Valine
tta = L = Leucine
ttg = L = Leucine
cca = P = Proline
cta = L = Leucine
gtc = V = Valine

This is, of course, just the beginning of a longer sequence that eventually forms the dagger that punches into your chest. Here is the complete sequence. Or you may prefer the graphical version.

And finally, “rolled in fat and sprinkled with sugar and salt”? It’s true. Look at the ingredient list for the Pfizer vaccine: A Breakdown of the Ingredients in the COVID Vaccines. Everything is there for a reason. The mRNA fragments are rolled into little fatty lipid cells to keep them safe. Salts help match the pH, and sugar stabilizes the shape.

And now the charm is nearly ready…

Double, double toil and trouble;
Fire burn and cauldron bubble

UPDATE: I made a guess here as to exactly what sequence they would use for the vaccine. But the actual sequence has now been published. See How Do You Spell Vaccine?

Grow Your Wallpaper

Giraffes have a look: brown puzzle blocks edged with white. Zebras are even more iconic, with those bold contrasty stripes. I don’t know how much good it does them on the Serengeti, but the branding is magnificent.

The interesting question is: how do you paint a zebra? How does a zebra come by its look? By going to the animal print store and searching the Zs? Visiting their favorite stripe salon for a new do? No, they get those stripes at the same store that sold you your fingerprints. The mom store. It’s in the genes. But saying it’s in the genes doesn’t really tell us much. How does it work?

Suppose you want your dining room done up in realistic zebra stripes. Your mom being human, you can’t do the zebra trick. So you go shopping for a quality zebra wallpaper print. That’s a challenge, but let’s assume you find it. Then there are other problems. First, it’s hard to get the paper to line up perfectly so the seams all fit. Next, you’ll notice repeating patterns in your wallpaper, because that’s how wallpaper works. Once you notice it, it will drive you crazy. But real zebras never have this problem. Their patterning is unique, and it shifts perfectly to suit the contours of their body. How?

The answer is that it’s a procedural pattern. There’s a program that paints zebra-ness in just the right way onto the canvas that is Mr. Zebra. And now, through the magic of neural networks, we can do it too. Check out this paper on self-organizing textures from some folks at Google.

Self-Organising Textures

Check it out! It’s mesmerizing, and there are dozens of patterns to play with, including, naturally, the iconic zebra. And since it’s algorithmic, we can give it different conditions. This, for example, is what zebra stripes look like at the south pole.

In the future, everything will be procedural. Architecture, silverware, clothing. If you want zebra stripes in your dining room, all you’ll need to do is specify exactly the area that you want to cover. Then you can just drop a zebra pattern seed in the lower left corner and wait. But we can do the zebra one better. If, next week, you decide it should be a giraffe-themed dining experience, just drop in a giraffe seed.

We’re starting to get a better idea how life cooks up, as Darwin says, “endless forms most beautiful.” I’ll leave you with a link to work by Sage Jenson, who might be one of the first biological artists. These simulations, originally inspired by slime mold self-organization, are essentially living canvases.

Visualizing Viral Mutation

This video is a beautiful visualization of how life finds a way to thrive in adverse circumstances. It provides some insight into what’s going on with the coronavirus as it mutates. The key point is that we think of vaccination as extinguishing a virus, when it’s really just one part of an endless back and forth tennis match. We’re continuously learning from each other and training each other. It never stops.

In the video, an E. coli culture has been placed at the ends of an enormous plate of agar. You can see some some stripes in it. Each stripe is more poisonous than the one before it. The bacterial colony starts in a happy zone with no poison. It grows quickly until it hits the first poison strip (antibiotic). It looks like it’s stalled out, but in a few locations, the bacteria figure out a way (they mutate) to defeat the effect of the poison, and you see another breakout. All the “children” who continue on are descended from the clever bacterium that cut the hole in the fence.

This process is repeated three more times. The progress of the colony is stymied… until somebody cuts through the next fence, and the relentless march continues. In other words, we may think we’re making a prison for bacteria, but it’s more useful to think of it as a training center for ninja commando bacteria!

The superimposed colored lines show branching trees of inheritance. The bacteria that reach the red “end zone” have had to adapt to life-threatening poison multiple times. Is their appearance in a sea of toxic sludge miraculous or inevitable? It’s hard to know ahead of time, but there they are. Life finds a way.

Here is the video. This is not a notional diagram or cartoon. You’re watching genuine evolution, the engine of life’s variety, happen before your eyes. There it is! If Darwin believed in heaven and was there right now, he’d be smiling at this. But I think he’s just dead. Anyway, press “play” already…

Here is a similar-looking tree diagram for the coronavirus, as maintained on Nextstrain.org. As before, each branch represents a successful mutation. Maybe it was in response to some stress, or maybe it was just random. But from just a few virus particles on the left side comes a fantastic woolly forest of viral variation on the right. It never stops.

We now have the magic goggles that let us watch evolution like a spectator sport. Nextstrain even has this fascinating “market share” diagram. Some strains do better. Some do worse. The ones that do better get to make more copies. It’s as simple, and as complicated, as that.

And it never ever stops.