Bioengineering the World’s Tiniest Cows

Let’s start today’s show with a trivia question. What Lou Reed song is most likely to increase milk production in cows? Anyone?

If you guessed “Walk on the Wild Side”, I’m sorry, but that is incorrect. But thank you for playing on today’s episode of Dairy Island Discs. If instead you voted for “Perfect Day”, then congratulations! The song (and, to be fair, other slow-moving songs like it) increased milk production by something like 3%. If it makes you feel any better, REM’s “Everybody Hurts” would also have been accepted by the judges.

I came across this tidbit while reading about Perfect Day Foods. Even though the Lou Reed anecdote inspired the name for their company, their actual goal is not to make dairy cows more efficient. It’s to do away with them completely. Perfect Day uses a bioengineered fungus to produce milk protein (casein) in the same way that we currently make alcohol: fermentation. These milk proteins can then be used by other companies, like Betterland, to make high quality cow-free dairy products.

Side-stepping the cow is good if you’re a vegan: no animal is involved. It’s also good if you’re lactose intolerant, since Perfect Day makes milk proteins but not milk sugars. But more generally it’s good if you want to decrease the impact that food production has on the planet. Bioreactor tanks require much less acreage (and fart considerably less) than cows.

When we think about moving away from animal food products, it’s often in terms of plant substitutes. Can you make a soybean patty that looks and tastes like a burger? Can you process oats in such a way that you get something resembling milk? Sort of. But your options are pretty limited. Ultimately there’s only such much you can do. Vegetable matter just isn’t the same as animal matter.

In the last few years this story has changed. You can now “teach” microorganisms to produce the exact same proteins made by animal cells. These aren’t sad look-alikes that taste wrong. These are molecular facsimiles. What is meat? What is milk? The important part is not where it comes from. The important part is what it is. And that is, by and large, a collection of proteins. If you can get the right proteins in place, you have a no-compromise replacement food. If you can do it cheaply and efficiently, you’ll disrupt the global food system.

We’re just seeing the first glimmers of this now, but it will only accelerate. We’ll get better at protein engineering, and the environmental costs of animal food production will become more apparent. In short, animals are an inefficient way to make the proteins we care about. They use too much land, they take too long, and they generate too much waste. Up until now, there was no alternative pathway to create these proteins. You want casein? You need animal milk. But now a door has opened, a lower energy pathway to achieve the same end. The first examples won’t be perfect. Incumbent industries will protest and lobby to protect themselves. Picky or nostalgic eaters will initially be unimpressed. But the change is coming. It’s coming because it’s possible and because it’s necessary.

Can you hear that sound? It’s the sound of a tiny herd of fungal cows coming to a dairy aisle near you. No word yet on how they feel about Lou Reed.

Parameterized Design on Demand

Here’s the scene. You’re at a friend’s house doing a little innocent bookshelf-snooping, when you spot something special. Oh my! It’s Wittgentstein’s Tractatus Logico-Philosophicus! At that moment, you realize that you too must have this book. Assuming you don’t want to steal or borrow your friend’s copy, you just need to note the title and author, and you can order it when you get home. But we live in a modern and impatient age. There’s a faster option: Open up your Amazon app, scan the barcode, and order the book instantly. It’ll probably be at your house before you’ve finished tomorrow’s lunch.

So the sequence goes like this: see, scan, order.

This routine is old hat for books. But it doesn’t take a genius to see that everything is going in this direction. Amazon is already using image recognition to help you identify things you might want to order. You can take a picture of your friend’s stapler, his egg timer, or his melon baller, and the Amazon app will do a credible job of helping you buy an exact copy.

What comes next? It seems pretty clear that before too long, everything will come with a unique digital identifier in the same way that every book comes with a unique bar-coded ISBN number. People don’t necessarily want visible bar codes hanging off their staplers and melon ballers, but near field communication and embedded IDs should make that unsightly problem go away.

So now you’re at your friend’s house, and you spot a side table that that would look great in your house. Your phone’s NFC chip reader tells you it’s a Tor side table from Model No. furniture. You can even order one right away, though you won’t get it instantly, since it’s generated on demand. So far, so good. But this is just the book story again, only about a table. But wait! Since it’s parameterized and generated on demand, you get some flexibility. It turns out your friend’s living room is larger than yours. If you want that table, it needs to be smaller. And you would prefer it in mauve. Conveniently, the Tor design, being parameterized, has a configurator that lets you make the necessary changes to its dimensions. You make it a little smaller in width and depth and place the order. The parameterization is built right into the unique model number.

What size table do you want?

The sequence now goes like this: see, scan, tweak design, order.

And in fact, modifying the Tor side table design is a real thing that Model No. furniture does, according to this article: Parametric Design and 3D Printing Deliver Custom, Sustainably Manufactured Furniture. Although, newsflash, this article is a year and a half old, and it appears that Model No. doesn’t do design customization anymore. But still, this is an idea whose time has come.

Drawing on ideas from the software world, it’s not hard to imagine an ecosystem of open-source furniture. Scanning a coffee table might take me to a GitHub-style page where I can find a list of suppliers who can print it, along with a configurator that lets me modify the design. Maybe I want to change the parameterization. Let’s say I want to make it so I can modify the height in addition to the depth and width. I can fork and refactor the configurator, verify the suppliers can still print it, and bam! I’ve added value for all subsequent purchasers of the table. Maybe I’ll even throw in an augmented reality model for simulation in my virtual living room model. This lets me anticipate how the final product will look in my house at every hour of the day.

Once design and manufacturing become amenable to consumer-driven digital inputs, a lot of fun things can happen. Refactorable parameterizers, procedural parameterizers, collaborative decorators, open repositories, renderers, simulators, AR walkthroughs, digital twins. All these things are on the way.

It’s going to get a lot easier to benefit from the good taste of your friends. But you’re never going to read that Wittgenstein book. Trust me.