Getting S^&t Done

Speed matters. Time is precious. Sometimes the decision is made by the passage of time rather than by a choice of action. No decision is a decision.

This turns out to be a crucial concept in enterprise business, where the levels of bureaucratic approval ensure that meaningful decisions require weeks, if not months. Practically, this means that if action is required in a shorter time frame, either that course of action is the proverbial “nonstarter,” or the entire decision-making apparatus needs to be circumvented. Generally, this leads to poor decisions.

And yet, technology continues to accelerate our productivity and our capacity to complete tasks. Moreover, this process is rarely mandated explicitly from management. Does your boss demand, via some company handbook, that you own a smartphone? Does your boss demand that you use some map app to navigate your vehicle to a location you’ve never visited? Does your boss demand that you use automobiles and airplanes?

Most likely not. But if you do not own a smartphone, it will be extremely difficult to manage the communication expectations of modern professional life. If you do not use some navigation app, when your boss tells you to be in a particular place at a particular time, you might find it difficult to manage without the modern efficiency provided by planes, trains, automobiles, and navigation systems. If you communicate via snail mail and transport yourself on foot, you will simply be unable to maintain the level of productivity enabled by technology. No specific mandate will be required - the productivity demands will suffice. Technology is mandated by the times.

Software

AE is in the business of software development. The entire industry surrounding development, data science, and other software products is a demonstration of the economic value generated by speed. LLMs can read and comprehend billions of times faster than human beings. Neural networks might be performing little more than the four basic mathematical operations, but they can execute those operations at speeds humans cannot fathom. Email is invaluable because it is instantaneous and the pony express was not.

And yet, human decision-making is still slow. Anyone in our sales department could regale you with tales of us discussing a client’s need for a product in a couple months. We estimate that we could build the product in a few weeks…and by the time the project is approved internally, the available time is exhausted, the need is moot, and all parties promise to try again the next month, quarter, year, etc.

The world of technology proceeds at an accelerating pace. This ensures that the time frame during which one can gain ground over competitors continues to narrow. Opportunities emerge and pass just as quickly.

We address this challenge by conditioning ourselves to work at the pace of technological development. We conceptualize, develop, and deliver software products every day. Corporate leaders are not going to demand the usage of Copilot and other large-language-model-assisted coding practices. But eventually, the disparity between the speed at which one can deliver with such aids will leave those coding without them akin to those who eschew email or automobiles.

Prose

Unless you’ve remained stunningly well-insulated from the news of the day, you’ve read/seen/experimented with GPT-4 and/or other large language models. We’ve seen the stories of algorithms passing bar exams or penning acceptable (if someone generic) college essays.

And I’m a guy who pens essays and parodies as some fraction of his professional responsibilities.

Currently, the ability to seed one’s creativity with the idea generation made possible by technology has not rendered writers obsolete. It has, however, meant that the content writer who takes two weeks to craft an eloquent 1,000-word article is a dinosaur. It means that the graphic artist who generates one image in a similar period of time is similarly past their expiration date.

Will the need for the linguistic tastes of a competent human being be eliminated by large language models? Eventually, perhaps. For now, I’ll be John Henry, pounding the keys in vain before the literary equivalent of cardiac arrest does me in.

Expectations

Technology enhances the capabilities of users. The industrial revolution ensured that the production capacity of an individual was no longer constrained by the brawn in their backs. The computation revolution ensured that the production of an individual was no longer constrained by the number of hours in a day (code runs overnight, even if we don’t). The LLM revolution guarantees that creative capacity is no longer burdened by the same limits associated with idea generation and iteration.

So maybe it helps to hire people who are accustomed to penning an essay in an afternoon, accelerating from 0-to-MVP in a single day, or writing a parody over the weekend.

We don’t always work “harder,” but we do let technology help us work faster and produce exponentially more value in so doing.

After all, anything else goes the way of the horse-drawn carriage and the pony express.