What happens to work in our AI future?
Those who can adapt will find that AI complements, and doesn't replace, their work. Others won't be so lucky
Every CEO I know complains about their fixed costs1. These are expenses which don’t vary with revenues. What if AI can make many, and possibly, all, of these expenses variable? Fixed costs are, well, fixed: if your company has a bad quarter, your fixed costs don’t decline in proportion with your sales. Rather, your margins get compressed. But if fixed costs could be converted to variable costs via the magic of AI, and so rise and fall with the success or problems of the business, that would be much better.
Well, here’s an interesting post from John Kennedy and Will Manidis which argues that AI may well allow us to convert many of these fixed costs to variable costs. Here’s their thesis:
With all of these [AI-related] changes, nirvana is near: software as a variable cost business. No longer does a new software business need to default to developers, sales reps, or customer success managers to reliably generate cash flow at scale. Every cost is a choice, and founders can now choose to avoid the pitfalls and problems of the past.
If generative AI can reach its potential, it represents a new frontier for software, where organizational structures, cost structures, and growth strategies can and should be re-imagined. These new opportunities empower creative business development: dealmaking, marketing, elephant hunting. Imagine services businesses being able to spin up entire software lines, or the rigorous operating procedures of private equity funds being applied to building—not buying—software companies.
The word if at the start of the second paragraph of this quote is doing a lot of work here: if our AI dreams are realized, then we’ll reach nirvana. Their argument is essentially that if AI develops sufficiently, then we’ll be able to encode a lot more business processes into software. This would allow CEOs to convert fixed costs to variable costs. If sales and marketing functions for a software company can be encoded in AI-based processes, then the expenses inherent in those processes will rise and fall in proportion with sales. The costs will no longer be fixed, but rather variable, and, well, nirvana is achieved. Obviously AI has not yet developed to the point where this is possible. But all indications are that we’ll get there eventually.
What this means for white collar labor in general is an open question. My view is that white collar workers (and probably blue collar workers2, too) need to become conversant with AI. They need to learn how to operate AI tools and agents, and they need to understand the rate at which the technology is improving. Adaptability and agility are key. Unfortunately for many white collar workers, their training and deportment militates against adaptability and agility. But for those who are able to adapt to radically different work styles, I see advanced AI as being a complement to, and not a replacement for, white collar work.
My suspicion is that institutionalists will fare worse. I use the word ‘institutionalist’ broadly here: any person whose job entails ensuring the stability of the institution for which they work is someone whose job may be displaced by sophisticated AI agents. It’s no coincidence that these people are among the least adaptable and agile. The types of jobs that institutionalists do include: corporate middle management, a lot of legal work, accounting, regulatory stuff, teachers and professors, etc. If you work in one of these professions, and are not close to retirement age, I would be spending a lot of time figuring out how to position yourself on the complementary side of AI. How can you modify your work and your career such that you use AI as a complement to your job? The answer to this question will vary, greatly, upon one’s circumstance and ability. But the answer is definitely not to stick one’s head in the sand and pretend that AI is just vaporware.
I asked ChatGPT to come up with a way to relate the principle of fixed and variable costs to a real-world example with which most people would be familiar. It came up with the notion of a car. You can read its output here.
Many people argue that blue collar workers (i.e., “the trades”) are safe from AI disruption, since “AI” is software, and, well, plumbers don’t use Excel. I think this conclusion is wrong, simply due to the rapid progress we’re seeing in robotics…due to marrying it with AI.