Liter of Cola

August 18, 2022

Without thinking that hard about it (and probably moderating my red hot take as a result), this is one of the best scenes from any movie ever made. (More than a few f-bombs in here, if you’ve got a kid around.)

I’ve obviously watched this a million times, but I’ve never really felt obligated to use it in an analogy, until today.

It’s just a quarter, and look how much more you get

I was just reading this Verge article on some of the concepts behind WeWork fraud kingpin Adam Neumann’s latest, wildly overcapitalized venture. The central thrust of the piece — and the critique of Neumann’s logic, really — is that while he’s talking about doing a lot to increase the tradability of a certain asset, he’s not actually doing anything to deal with the fact that the assets are often garbage.

When we started Efficiency Exchange ten years ago (holy crap), a major inspiration for the company was my wife’s work with auditors, who were tasked with assessing extraordinarily qualitative things (conditions and output quality of a factory) and turning that into data that could be consumed by people halfway around the world, some of whom had never been to a factory in their life. There was a TON of interest from VCs and other tech-world investors in making that data accessible, sortable, and identifying patterns and detecting efficiencies from it using things like (yes, groan) machine learning.

But there was very, very little interest in whether that data was, fundamentally, any good. There were people who worked on that aspect of things, but they were of a very different mindset, and essentially wanted to build better audits, align companies on standards, and do things like that. Basically, while the VC people only thought about scaling things — even if what they were scaling was fundamentally useless — the sustainability people only thought about that real-world applicability, and never about scale. That’s one of the reasons why progress on that front is so slow, whereas ten years later I can ask a web app to draw me a picture of Kermit the Frog eating a toaster and it will give me this in 90 seconds.

Basically, we can fix one piece of shit, or we can scale a big box of shit, but we’re just absolutely terrible at scaling a box of fixed shit.

Over the last decade, the scale people have absolutely, totally won, at least from the perspective of who spends, invests, and talks about what. EVERYTHING is at scale. Content, finance, assets, audiences, it’s all scaled. We make a lot of references to the micro-unit — i.e., some random YouTube influencer being interviewed — but the business of YouTube and all the surrounding services and industries are all about scale. After all, YouTube really is a scaled box of shit — it’s just SO big that it has lots of awesome stuff on it. That’s the answer. So much of everything that eventually something good will happen; the trick is just to expose you to as many things as possible in the quickest amount of time (swipe, swipe, swipe) so that eventually you get the dopamine hit. We never really did figure out curation, and it kinda feels like we’ve given up.

It’s not just consumers. The same piles exist at work — people who are focused on some specific thing without the greater context around it, and people who are focused on nothing BUT context. The former will plug away on their work until layoffs are suddenly and shockingly announced, and the latter is often living in a complete fantasy world because they shout queries at Salesforce and endlessly debate other high-level people about what the answers mean.

Does this look like spit?

Which brings me back to Farva and his liter of cola. Here we have a guy who has absolutely no idea what is behind the counter, or how anything back there works, interacting with a system that is only interested in how things actually work. Farva has to conform to the standards of the system to get anything useful, but he can’t — in fact, he actively refuses to — and since he has no understanding of the standards or the system, he gets crappy answers that frustrate him. Even reassurances from the system that it understands him (“that just means make it extra good”) just antagonize him further, because again, he has no fundamental understanding of how things work behind the counter, and he doesn’t want to acquire it. To be fair, the system doesn’t care either, going so far as to take a query it knows is nonsensical (“Do we make Liter Cola?”) and just throw it over the wall and see what comes back.

The best is when Ramathorn implores Farva to sort of “normalize” his query to the system — “just order a large, Farva”, and Farva responds with the legendary “I don’t want a large Farva, I want a goddamn liter of cola”.

Adam Neumann doesn’t want to talk about auditing, dammit! He wants to talk about the blockchain, and a world of frictionlessly trading assets. Are those assets nonsense? Maybe! Is that a problem? Of course not! We’ll push the “vet assets” button and hire a bunch of college grads in a boiler room to read them or something, I dunno.

Oh good, more thoroughly! Why didn’t I think of that? No one would invest in an auditing company who claimed they could solve the problems of auditing by auditing more carefully, because that’s not even an idea, even if it’s pointed at an actual problem. Instead, here’s something that is definitely an idea, but isn’t pointed at the problem at all. Perfect! This increasingly dense layer between reality and strategy is starting to drive me absolutely bananas. Those two concepts are different things, of course, but they’re obviously related, no? It’s like we’ve said “the plural of anecdotes is not data” so many times that we’ve decided data or other high-level concepts don’t need to be traced down to an actual real-world use case at all.

I really do think — and this is not even a warm take — that the internet and cloud computing have made solving a lot of unthinkable problems with brute scale actually feasible, and that really is an awesome, mind-opening achievement for technology. I don’t want to lose sight of that as I age. But this has gotten ridiculous, and now we’ve got a generation of people who are simply scaling things not just before they’ve considered the impact of scale, but before they even really understand the thing they’re scaling. The scaled version of the thing is so different, understanding of the micro-unit is a waste of time, or so I assume because I actually have very little experience looking at a micro-unit in any detail. You people are so petty… and tiny!

Here’s the thing. With any fixed amount of time and energy, you can do deeper analysis of fewer things, or more shallow analysis of more things. The cheapest thing to do in the age of communities, connectivity, and cloud computing is to do the latter, and so that’s rapidly becoming the answer, even when it’s the answer to a question no one asked.

At some point, though, you’re going to find yourself building a billion dollar asset swapping system with meaningless assets, or leaping over the counter to assault a food service employee, which I think is probably some tortured equivalent of investing $350 million in the WeWork guy.