The Flood Before the Flow

I have written a lot lately about jobs, labor shifts, and the longer arc of AI's disruption. That disruption is real. It is coming. And I have been upfront about the scale of what is ahead.

The short version? The train has left the station. But the journey is just beginning.

As I have argued, AI will reshape the workforce. First slowly, then all at once. What we are seeing now is not mass replacement, but a rebalancing. Productivity gains. Augmented teams. Fewer new roles. Job growth flattening as companies scale with leaner teams and smarter systems.

We find ourselves firmly in the augmentation phase. You would not know it from all of the blather about autonomous agentic AI and full-scale orchestration. The reality is most companies are still running around trying to figure out which tools to adopt, which platforms to bet on, and how to keep up with a wave of releases that never seems to stop.

This week’s piece is not about jobs. It is about the hype cycle. The endless flood of new models, apps, agent platforms, breakthroughs, and step changes. And how to navigate all of it without losing your everloving mind.

My take? You don’t have to chase every shiny thing.

Let the AI come to you.

Let’s dive in. 👇

I spend a lot of time on AI, both researching and getting hands-on. Testing. Prototyping. Exploring what is new. Staying close to the tools and trends is a calling, and also a professional necessity, especially now that this newsletter runs weekly. I had no illusions about the pace. I know what I signed up for. And yet, lately, I find myself a bit bored. Often underwhelmed.

Every week brings a new tool, a new model, a new thing you have to try right now. Last week it was an agent platform that automates any personal workflow and handles Gmail, Notion, and Slack on your behalf. This week, it is another orchestration layer built on top of ChatGPT. Next week? Probably a new wrapper for the wrapper from last month. What was it called again?

Some of them are genuinely clever. Many are just crap dressed in AI’s clothing. And a lot of it feels like a preview of what will be bundled into ChatGPT in Q3, then Gemini and Copilot a few quarters later, but by then, more integrated, more robust, more secure, and more durable.

That realization keeps returning. This stuff is coming. It just might not be coming from the folks trying to sell it to you right now.

There is something deeper going on here than feature fatigue. It is market confusion. Decision exhaustion. The sense that the window to act is always closing, and that falling behind is not a risk, it is an existential certainty. Hesitate, and you are obsolete.

So you sign up. You click around. You connect your apps and upload your docs. You craft a few prompts. Then you jump to another tab only to forget this “game-changer” by lunchtime tomorrow.

Here is the truth most people will not say out loud: if you are not a tech-first business, you can afford to let the hype pass you by. This cycle is not a call to action. If ever there were a time to pause, take a breath, and gather your thoughts, it is now.

Yes, experimentation is healthy. Essential, even. I built a news summarization bot the other day using one of these new agent tools. It connected to my feeds, clustered articles by topic, and generated a nice tidy digest. It worked. And that was the point. Not to deploy it at scale or productize it, but to test the edge and identify what, if anything, gets lost in translation. In this case, I missed the act of skimming. The intuition. The joy of discovery and the art of curation that comes from firsthand experience and nuanced judgment. The half-second spark when your eye catches a headline and your brain lights up. My little bot filtered out some noise, but it dulled the signal too. That insight was worth the time. Not every experiment should scale. But every team should be running them. If only to explore the possibilities and understand where all of this is going in six months or a year from now.

This led me to another not-so-subtle realization. It has been sitting in plain sight, but the more hands-on I’ve gotten, the more obvious it has become: we are over-automating. Not thoughtfully. Not strategically. Simply because we can. And some things suffer for it. Some tasks lose their meaning when abstracted. Some processes are worse, flatter, and duller when filtered through automation. We need to slow down and consider what is actually being gained, and what is being lost.

I’ve had this same feeling with AI art. It is fun, even useful. I’ve been able to add original comics to this newsletter that would have been out of reach otherwise. But as with most generative output, the results tend to land squarely in the middle. Competent, yes. But never great. Never truly original. The more we automate the creative edges, the more we risk sanding off what gives the work substance. Are we on a path to automating ourselves into mediocrity?

This is where durable value is beginning to emerge. Instead of relying on point solutions layered on top of existing systems, companies are starting to see greater payoff from AI features embedded directly into the enterprise systems they already use, the platforms where data already lives and work already happens. These AI capabilities are implemented not by your team, but by the platforms themselves. These are not after-market bolt-ons that void your warranty. They are deeply integrated, security-aware, and nearly invisible by design. Your CRM, your productivity suite, your collaboration stack—each is adding AI from the inside out, in ways that respect risk, access, and operational continuity.

Why? Because they have to. They own the enterprise. They understand what is at stake. They have everything on the line. Their brand depends on it. This is not some side project spun out of a hackathon at midnight by a couple of Ph.D. students in Paris. This is the infrastructure of modern business.

So, where does all of this lead? What does it mean? Success now depends on knowing where to explore and where to wait. Where to run pilots and where to hold the line. It means assigning someone, a business analyst, not IT, to play in the sandboxes, track emerging tools, and kick the tires. But it also means resisting the pressure to install every shiny new app that pops up on Betalist or Hacker News. Because odds are, if it matters, if it's really good, it will not stay standalone for long.

If your systems are modern, if your platform bets are sound, the best AI capabilities will arrive there. You do not need to chase every tool. The real innovations will meet you right inside your CRM, ERP, your productivity suite, your collaboration stack. You’re not missing out by waiting. You’re being smart. You’re being strategic.

In the same way mobile functionality moved into core enterprise software 15 years ago, AI is being absorbed, standardized, and deployed at scale through existing platforms. The best features will surface. The best vendors will get acquired. The best use cases will be reborn as native capabilities.

This is not a dismissal of innovation or the impact of AI. AI will likely surpass the internet in reach and consequence. But the next phase will not require great acrobatics and herculean transformation efforts, even if that is what McKinsey is trying to sell you. No, it will show up inside the software you already use—smartly integrated, increasingly capable, and sometimes even invisible.

If you are leading an organization right now, the goal is not to outrun the flood. It is to build toward the flow.

Let the noise pass. Build the muscle. Stay close to the signal. Meet the future with a rational mind, discipline, and a plan.

Want to chat about AI, work, and where it’s all headed? Let’s connect. Find me on LinkedIn and drop me a message.