Photo by Kris Møklebust.
This weekend, we found ourselves up to our elbows in olives. And if we’re being honest about it, olives contained to just elbows and forearms would have been an improvement over what actually happened. In preparation for a dinner party, we bought two giant jars of green olives from the nearby Greek grocer. With pits. Because the Greek grocer isn’t selling you pitted olives. What are you, a baby? And no, they don’t sell olive pitters. Just olives.
Not to worry, there’s a kitchen gadget store a few blocks over and they will definitely sell us an olive pitter. Eighteen dollars and a few hours later, we’re finally ready to start the first step of this dish. Pour the giant jars of olives into a bowl and start pitting them. The first pit gets stuck. We take turns trying and the pit won’t budge. With enough force, the pit finally dislodges and goes flying across the kitchen counter. The pitted olive looks less like an olive and more like tapenade. And the counter and floor now have a sheen of brine. One down, several hundred to go.
About half an hour into pit extraction, one of us quotes an adage that’s been following us around for weeks. It’s a variation on “no bad weather, just bad gear.” And the first time we heard it, it was on a website advertising soap for washing graffiti off of buildings. You should brace yourself because once you’ve heard it, there’s no going back. Ready?
With the right tool, the job is easy. So if the job is hard, you’ve got the wrong tool.
Very few of you will have direct experience with getting sharpie off of windows. But let us be the first to tell you that soap doesn’t do shit. Windex doesn’t do shit. Elbow grease doesn’t do shit. It’s not about how much soap you apply or how hard you scrub. You can be at it for hours and still, nothing doing. But if you find yourself a razor blade, the window is clear in less than five minutes. Because with the right tool, the job is easy.
Olive you get back to work
The reason the Greek grocer doesn’t sell an olive pitter is that most Greek recipes leave the pits in. And when they don’t, a large chef’s knife does the trick. Without getting brine on the floor. And without macerating the meat of the olive.
In very very related news, remember the MIT study last year that said 95% of all AI pilots in the workforce were failing and then the CEOs were like, you’re just not using enough elbow grease. Probably you just need to use more tokens. And then they conducted mass layoffs to root out the non-believers and fire the middle managers and declared “that oughta fix it, let’s go productivity.” But then, oops, months later, the windows are still covered in sharpie.
And in other, very very related news, remember when the CEOs of major AI companies estimated that half of jobs would disappear. And the youth were like, hey that sounds really bad and the CEOs were like: shrug, learn to code, lolz, maybe don’t, sucks to be you, I guess. And then recently, the youth who are now graduating are booing when AI comes up in their commencement addresses. And the AI CEOs are like, wait, that hurts our feeeeeelings. And the grads are like, are we doing feelings now?
And in other, other, very very related news, hey has anyone noticed those same AI CEOs are doing rather a lot of media spots right now. And that the latest universal talking point is about urgently revising down the estimates about the percentage of the human workforce that they expect to displace. Because, you know, in their optimism about the demise of the modern workforce, they may have overstated those numbers. And they seem to have just figured out their role in the plot of Les Mis.
A tide fit to be turned
That the Greek grocer doesn’t sell an olive pitter should have been our first clue that we had the wrong tool for the job. That said pitter exploded the olive was our next clue. And the third and most damning clue was that the pitter was sold as an olive/cherry pitter. It’s not that there’s no use case, it’s that for the use case we had, it was the wrong tool. Because it made the job more difficult and messier and slower.
More difficult. Messier. Slower. These are not selling points. Ruh roh.
There are places where AI adoption makes sense. If you need help spotting them, use our graffiti removal company’s definition. You will know you’ve got the right tool when the job is easy. Not when the job is made more complex. Not when the job takes more cycles to complete. Not when hundreds of agents participate in the job. The right tool makes the job easy.
If, otoh, use of that tool makes the job at hand more difficult, or slower, or 10x more expensive, take a moment to consider that you have the wrong tool. No matter how big the funding announcements are for those tools. No matter how much breathless press coverage the tools get about their no-humans-just-tools vision of the future. At some point you’ve gotta reckon with the fact that you’ve got accidental tapenade. That reckoning is coming.
Bulk-buying mission accomplished banners
If we’re right, the next 12 months will be a flurry of CEOs declaring victory. Not, like, a victory-victory. But a “no need to continue” victory. A sort of “that’s enough of that” victory. CEOs and boards who went all-in on the AI froth — the mandatory AI experimentation policies, the suspension of regular business and performance metrics in order to maximally encourage fucking around with AI, unlimited token budgets but punishing employees who don’t use enough — they are gonna have a hard time calling that whole fever dream a mistake. Claude is quick to grovel and self-castigate when caught screwing up, but businesses still run on people and people, especially people in power, are pretty bad at owning their fuck ups.
But a victory? A win you can declare arbitrarily? And then move on from? That feels great. And you can predict the phrasing of the mission accomplished email just as well as we can.
We’ve learned a great deal from our experimentation … Remember this neat example? … I’m proud and impressed with the work we have all done to modernize, to challenge our assumptions and each other, and find new ways to save money and time … We know where to focus now, and where the greatest promise is … And now that we have fully integrated AI into how we do business here, we’re ready to move into the next phase…
The next phase won’t involve rejecting AI use. That would be too close to admitting failure, and anyway isn’t necessary for them to move on. There are times when it is genuinely a helpful tool. What the next phase will do is remove the mandates, the forced adoption, and the unlimited budget. Obviously continue to find savings and efficiency, they’ll say. Use AI where it makes sense, they’ll say. We’re just done hemorrhaging money and ignoring our core business, is a thing they probably won’t say. At least not out loud.
But that will be the core of the next phase. CEOs are already having the hard conversations with their CTOs and CIOs about AI costs and compliance issues. But they’re only just starting to have the conversations with their product teams about falling quality, their marketing teams about poorer engagement, and their customer service teams about the spike in churn. To the extent that they haven’t entirely gutted those teams. What you can see them realizing, in real time, is that their house is on fire.
CEOs are a particularly vulnerable population for AI hype, and make all kinds of silly mistakes, but even still most of them will not knowingly drive their own companies into a wall. And so the other part of this new phase will be a return to business focus. We’ve seen this before. Whether it’s the CEO re-asserting, or announcing that some new executive has their axe, the message will be clear: we have to focus on our core metrics and get ourselves back upright.
(But also, Mission Accomplished!)
Okay so now what
We know it doesn’t exactly hit as an optimistic take. Oh, thank goodness, the CEOs are back to focusing on profitability. That always works out well for humanity in general, and labour in particular. We try our best to give you all our best sense what will actually happen with work, not what we wish might. But having said that, this is actually hopeful.
It’s hopeful because AI has been, for many people, an unregulated workplace hazard. The evidence has been clear for years now that some people are deeply harmed by this tech, and that many more experience subtle forms of cognitive and emotional impairment. It’s not yet clear what safe handling practices look like for this one but, while we all figure it out, it’s a good thing for it stop being compulsory.
It’s hopeful because “what are we here to do as an organization?” is a standard that makes any fucking sense. One of the cruelest things you can do to an employee who cares is to strip their work of meaning. Taking you off of your day job, a job that you’ve likely spent years learning how to do, to make you, ahem, a prompt engineer, is such a fucking waste. Things making sense again is better.
And, listen. It’s hopeful because in business, in the entire history of any kind of paid employment enterprise, for thousands of years and across the world, do you know what has consistently had a positive return on investment? Building great teams. You don’t need to grasp at edge cases or invoke an org-wide mandate or try to find a 3% win somewhere to justify the exercise. It turns out that bringing together teams of people, investing in them, creating conditions where they can thrive and goals worth pursuing, pays off massively and repeatably.
And when you bring those people together, by all means, give them tools. Humans are long-time tool users. It’s part of our whole origin myth. Long before AI. Long before graffiti removal chemical compounds. Long before olive pitters. Not because of mandates. Not because of circular logic around what counts as productivity. But because when you have the right tool for the job, the job is easy. And because we all genuinely enjoy getting shit done.
— Melissa & Johnathan
