Why Irrationality May Be Our Best Form of Job Security

As Sam Altman and others have said, there have been three great technological revolutions in human history: the agricultural revolution, about 10,000 years ago, the industrial revolution, which started about 200 years ago, and the software revolution, which is happening right now.

We went from foragers to farmers to workers, and now we’re being phased out of the worker category. The service industry will likely remain after the post-software revolution (we will all still want cocktails and street tacos), but the status of all other professions is unclear, because sooner or later, a computer will do your job better than you can.

The software revolution has already started in my field, the law. In the last year, I started integrating automation technology into my business processes. Any documents that are repetitive or require little modification or customization now go through software. I suspect I might be fairly cutting edge in this way. I would guess that fewer than 2% of attorneys use any form of automation yet.

But they’ll get there. Or they’ll be replaced by those who will.

Still, something happened to me last week that made me feel secure in my future.

I was helping a client with a trademark search and I came across a result that was totally contradictory.

I ran the result by a couple of other attorneys and we all agreed. It made no sense. There was no way to explain the result except to say that an attorney at the USPTO made an irrational decision ten years ago, and now there is a whole body of law built around it.

In the United States, we’ve created a legal framework through a combination of legislative actions and judicial precedents that interpret those legislative actions. Most of them make sense when viewed together, but sometimes they do not.

In those instances, you have to speak with your clients and say, “hey, in this instance, the law doesn’t make a whole lot of sense.” And that creates an inherent uncertainty, because you cannot make predictive decisions based on contradictory information, no matter how smart you are or how quickly you can make calculations.

But still, I think if we were to get an adverse ruling in this trademark situation, I might be able to call up the attorney at the USPTO and have a chat. And if I were nice and polite and said all the right things, maybe I could get that attorney to see things my way.

Or maybe not.

But it certainly strikes me that the hardest processes to automate are the ones that are inherently irrational or contradictory. I am not much of a programmer, but it is not an accident that some of the biggest names in artificial intelligence are obsessed with Ludwig Wittgenstein, a philosopher whose lasting legacy centers around the inherent fuzziness of language (Blue book is the even the name of the company in Ex Machina).

Figure out how to program the fuzzy and contradictory nature of our existing forms of usage, and you have a true AI.

Until then, you and I will probably still have jobs.