This is the first of an intended series of articles on animacy.ai that is explicitly aimed at mere mortals - people outside of the “tech” industry altogether.  Stay tuned for more.

So, I am writing this while on a liveaboard dive boat in the Red Sea.  I admit that the timing for a vacation in the Middle East is a little weird, but we planned this long ago, so hopefully all will be well.

Though I hadn’t intended on writing during this trip, I’m having the same conversation with many of the folks I am meeting on the boat, so I thought it might be worth sharing.

The repeated script is something like this:

Them: What do you do for work?

Me:  Well, I’m a longtime software guy, so now I’m an AI guy.

Them:  Yeah, I’ve tried AI, but it is never gonna replace what people do in my job.  It might be helpful, but it makes mistakes and even if it is right, the answers it gives can be misinterpreted by people, so they still need an expert to explain things to them.

Me:  Hmm.  I find that all of the frontier chatbots are better at reexplaining things I don’t understand than experts are.

Them:  Well, yeah, but you’re a computer guy.

Me:  Hmm.  Actually, that is the whole thing about these agents.  They are speaking natural language and responding with the full context of what they know about the user, the project, and in many cases, the specific domain.  There are lots of studies showing human expert proficiency in many domains.

Them:  Well, I don’t think that will ever be true in what I do.

Me:  What do you do?

Them:  I’m a CPA (for example) and when ChatGPT answers a question about someone’s taxes, it gives the correct answer, but most people would misread the answer so it isn’t very useful.

Me: How does a human do better?

Them: We follow up with questions to make sure the client really understood the answer.

Me:  And you don’t think the bot will ever be able to do that?

Them:  Maybe for people like you, but not, say, for the average small business owner.

Me:  Haven’t you seen the bots improve dramatically in the last two years and especially in the last 6 months?

Them:  Sort of.  But they will never get there.  I’m not worried.

At this point, I have to make a decision whether to try to convince them to be worried, or just be polite and mentally log the conversation.

So, if you can see yourself on the “Them” side of the above conversation, or can imagine empathizing or agreeing with “Them”, here is my guidance.

Be worried.

The agents are coming, if they aren’t already here.

If you are in a knowledge work domain, you need to be mastering how to use AI, to get on top of these minions, so that aren’t replaced by them.  There are incredible productivity gains to be had, and you can become seemingly “superhuman” quite quickly, but this promise is against the certainty that deep domain knowledge and skills that humans have spent years to learn and perfect, will be raidly replaced by autonomous or interactive agents which can operate at a far lower cost than the human labor they are replacing.

Please do not think that I am saying this is good.  I am just saying this “is”.  The deficiencies of today’s agents, which in most cases are actually alleged deficiencies of older models, are likely already addressed by models that have been released or which are “in the labs”.  This is a matter of weeks or months, not years, in most cases.

To maintain your ecological niche in our disrupting economy, you have to become a superuser, a superbot, a bot apologist, a bot bringer…agentify, or be commoditized.

It will happen far faster than most of us think.

And don’t believe for a second that you can’t learn this.  I’ve been hearing BS about “I’m not technical” my whole career as an easy discount for not learning.  In some cases, the learning burden might not have been worth it.  Until now.

These things speak you language.  Not just your spoken/written language like English or Mandarin Chinese, but your domain language.  They “speak” CPA.  If you are an expert, you can tell them to treat you like one, and you can coach them until they are as good as the best human assistant (or at least an above average one).

You do not need to learn anything to “use” these agents.  They will coach you through using them.  They will build what you need to use them.  They will even build what you need to check their work and keep them honest.

You do not need to be technical.  You need to know your stuff.  You need to be able to check their work, the same as you would with anyone you delegate to.  I am simplifying this quite a bit, but hopefully you catch my drift.

Bottom line:  The agents are coming for your job, figure out how to use them, before everyone else does.

In a future article, I’ll write about success strategies for employment during AI disruption.