More Efficient

Reading or hearing or having to consider even one more thing about LLMs or so-called "genAI" feels like the straw that will break the very sad and deeply tired camel's back. All the promises of being left behind if you don't start using LLMs to do the things you actually love to do remain, unfortunately, unfulfilled. Rather than being left to our own primitive devices, we instead endure a constant flood of proselytizing and performed inevitability, apps and peers begging you (sometimes subtly, sometimes not) to wake up, become modern, come to your senses, and most importantly, become more efficient.

For the sake of the camel, though, let's set aside the buzzwords and this specific manifestation of hypercapitalism and instead consider efficiency itself for a moment. What is elided in that consideration, by way of euphemism, bounded assessments, and decontextualization? Because that tells us a lot not just about what we are wading through today, but also about how capitalism and markets have disproportionately depended on efficiency, especially since the industrial revolution.

Efficiency is important. It is the backbone of capitalism and markets. Market performance is about an equation: profit equals gains minus losses. Efficiency is about increasing your gains, and decreasing your losses. Whoever increases their gains the most, and decreases their losses the most, whoever is the most efficient, is the winner. And being the winner is a big deal.

The winners rule the world - they decide whether or not your drones work, who gets to be on the front page of the New York Times, who gets sacrificed, who doesn't, and so on. So there's a great, powerful incentive to be efficient, to increase your gains and to decrease your losses. After all, even if you break a few rules on the way, you could always make it so those rules don't apply to you anymore.

This is the most straightforward and broadly understood narrative of the market. Everyone is competing to increase their gains and decrease their losses. And of course there is corruption and "cheating" and so on, so there's a kind of meta-game around rules and regulations, to try to cap or otherwise control or govern feedback loops and so on. And yes, any game designer can tell you that if the winners get to make the rules, then your game balance is going to eventually break and be very hard to recover, but that's not what the camel problem is about.

The camel problem, the reason the beast is so tired, is about a third-order market. So let's say that gains and losses, in the most basic sense, comprise a first-order market. A direct market. You could say then that the rules and regulations for the first market constitute a kind of indirect, second-order market.

The third-order market is about the way you define gains and losses in the first place. Especially losses. Gains are hard work. Gains require investment, strategy, maybe inherited monopolies, bribes, takeovers, they're an active thing, they have to be maintained, defended even. But the market still demands that you be more competitive, more efficient, that you demonstrate growth and value, that the number goes up, and so on.

So the easiest way to be more efficient is to simply not account for your losses at all. This is easier said than done. You can lie and break the law, of course, but that's risky. Fortunately, there are two very popular strategies for hiding your losses legally.

The first strategy is to create an industry where the losses are not yet regulated. For example, if there are no regulations against dumping formeldahyde into the ground water around your factory, you could make a business around using formeldahyde to cure the leather for your workboots. You can create a profit margin, or an efficiency, by not wasting money on waste disposal. Higher quality workboots at a lower price! Local jobs and a good company culture! This is a great, and most importantly, legal business to operate.

Or you could create a business around operating unlicensed cabs, without paying any of the taxes or applying for the permits that account for their impact on the city and region. Or you could create a business around operating unlicensed hotels, without paying the taxes or applying for the permits that account for their impact on the city and region. You get the idea.

The second strategy is to move an industry somewhere with more lax regulations. For example, if India has fewer regulations around worker wages and waste disposal, you could relocate your garment manufacturing work to that region. Move it from an area where your losses would have to be tracked and accounted for, to an area where your losses no longer have to be tracked or accounted for. Economic development! Modernizing the third world! This is a great, and most importantly, legal business to operate.

You can even create enclaves and special export zones in "developing" nations, where you can use a mix of rules and regulations that maximize the way your losses can be hidden without ceding full control to the host country. In both cases, you have the added advantage of being able to conduct your business a very long way away from the people who consume its output.

Essentially, you can either hide your losses in plain sight, or you can stash them far away. Both strategies are legal, and both strategies create the illusion of efficiency by concealing the actual impact of the industry.

It is only through the deliberate and laborious obfuscation of losses that industries like these can pretend to be anything resembling efficient. It is only through a tightly constrained and narrow assessment of gains and losses that you can claim the losses to be as low as they are, and thus the profits to be as high as they are. If your losses are invisible, then they do not need accounting, and you are efficient. You win.

This is relevant to the aforementioned camel-in-poor-condition because this is exactly how LLMs and "genAI" work. This third-order market is exactly how and why "AI" is a growth industry and a hotbed for investment.

Yes, your output will be lower quality. Yes, you can only do this if you adopt liberally from work done by others, without compensating them for their contributions. Yes, you will lose vital skills relevant to your profession by outsourcing the actual labor. Yes, you will devour massive amounts of underregulated natural resources. Yes, you will produce huge amounts of underregulated waste in impoverished areas unable to combat it.

But at least you will be more efficient.