Generative AI is over, in case that wasn’t clear

Man, I’ve been waiting to write this one for a while.

Generative AI is over.

Wait, what? No, not really. I like to get histrionic when I spit out these little proclamations that will be more concrete anywhere from three to nine months from now. Because I want you to have that sentence etched into your brain as you plan out what exactly it is you say you do here over the next three to nine months.

Because I don’t want to see you, three to nine months from now, still sending five-page reports full of AI slop when a single surgical paragraph will do. I don’t want to see you get silently downgraded when your sales deck has three “It’s not this, it’s that” phrases on three consecutive slides. And I don’t want to see you saving your SaaS company by bolting on a chatbot to replace your report interface or customer service agents and calling your product “AI native.”

Yeah, the bad news is that’s not Taco Tuesday for amphibian left-handers; it’s a critical mass understanding of how LLMs work.

The good news is the case for AI has never been stronger.

A brief history of me crapping on generative AI

Look, I’ve been doing this AI nonsense for a long time. Sixteen years ago, I started talking about the time and place for generative AI. I rambled on for seven years. No one listened.

A little over three years ago, I started making a contrarian bet against generative AI in my public and personal writing. No one listened.

But a funny thing happened. The bet paid off exponentially. Across reads, followers, subscribers and, most importantly, revenue, my numbers over the last three years have jumped anywhere from 10x to 100x and more.

I’m a 100x-er! Thanks, AI! Thanks for nothing!

Then I said SaaS was dead. No one listened.

Aw, man . . .

My dirty secret: I love AI!

But the thing is, I haven’t been anti-AI at all. Far from it. I’m an OG and still a fan.

I have been very singular, consistent, and vocal in my criticism. Actually, more of a warning:

“Don’t bet the farm on generative AI. It’s not new. The novelty will wear off. AI is still nascent.”

• So when “thought experts” started drowning us in generative AI slop, I took my shot.
• When people first started catching on to generative AI’s tics, I took my shot.
• When entrepreneurs started building businesses around generative AI chatbots, I took my shot.
• When CEOs started blaming AI for mass layoffs, I took my shot.
• When AI consultants started skipping over the necessity of data as the AI moat, I took my shot.

People started listening. About a year ago.

All along, I’ve been leaning into what’s coming with this awesome new evolution in processing power and compute packaging—which is really what we’re talking about when we say “AI.”

So let me lash out one more time. Because . . .

We’re entering the Age of AI brevity

I finally took the jump a couple years ago and got back into AI consulting as the calls coming in stopped being about chatbot creation and started being about data, intelligence, prediction, and action.

And now, a couple years later, this AI cycle is repeating in much the same way it did when we first unleashed our Generative AI platform 16 years ago, when our mantra was, “Put the most critical, insightful, actionable information at the top.”

The rules are still the rules.

Rule No. 1: AI is a tool, not a replacement. Integrating AI without experienced human control at the helm is like asking a spreadsheet to make you breakfast. We’re finally figuring out that 100x productivity concentrated in one link in the chain does not increase the strength of the chain.

Rule No. 2: Human ingenuity is your source material. A workflow that isn’t stress-tested in the real world is automation that will fail in a way that knocks the entire system over. The reason why Amazon hasn’t been taken out by some weekend-vibe-coding AI upstart is not because of AI. Amazon uses way more AI than any weekend warrior. It’s because Amazon has hundreds of thousands of people who know and live the problems and the solutions.

Unless they keep laying them off.

Rule No. 3: Data enrichment is your moat. When everyone has access to the same computing power and tools, and those tools can be used to replicate brute force, a company’s competitive moat becomes surgical and is contained in the proprietary workflows that generate the data that no one else has or can get.

Rule No. 4: We’re entering the age of AI brevity. Less is more. And as these chatbots continue to yammer on in what’s probably an uncontrollable cycle of training and bloat-begetting-bloat, we’re no longer impressed. It’s no longer novel. We’re drowning in fluff we’re no longer reading, features we never asked for, and productivity we can’t wrangle.

Passing along generative AI resultant sets without the filter of human ingenuity and decision-making to pare those results down to a known and verifiable set of the most critical, insightful, and actionable information is doing something anyone can do, doing it poorly, and not expecting anyone to notice.

That’s what I want etched in your brain. Because I know I’m going to get dragged for saying, “Generative AI is over,” but only for the next three to nine months.

If you like tech, AI, entrepreneurship, and the culture around it, please join my email list to get experienced, 100% human-written guidance peppered with jokes and obscure references.

—Joe Procopio