Trendslop and the Invisible Hand

I’ve always been a cynical person—not in the tiresome, performative way where someone at a party tells you they “don’t trust anyone” immediately before trying to sell you crypto. I mean that I never once operated under the assumption that the employer-to-employee relationship rested on anything but mutual self-interest, and honestly—honestly—that clarity is a relief. I don’t like the ambiguity of pretend-familial obligation. Don’t invite me to the picnic. Don’t ask me to be a “culture carrier.” Pay me what I’m worth, give me work that doesn’t make me want to set the building on fire, and we’re square. The fuzziness—the “we’re a family,” the “we care about your whole self”—was always a trap, and I’m too autistic to walk into a trap I can see the welding marks on.

Anything beyond the bare minimum from an employer has been like pulling teeth, and it’s never something you can simply ask for. You don’t walk into a performance review and say, “I’ve earned this promotion by every metric you handed me, so I’d like it now,” and receive anything other than a PowerPoint about “headcount constraints” and “budget cycles” and “we really value you, just keep doing what you’re doing.” The only thing that moves the needle is leverage. You find it, you bend them over a barrel, you apply pressure if it’s worth applying—or you move on. The maths is simple. Not easy. Simple.

You want that promotion? It doesn’t matter that you earned it on merit—merit is a bedtime story told to junior devs to keep them pulling all-nighters. It doesn’t matter that you earned it on the metrics they told you to chase—sales revenue, feature delivery, team growth, whatever numbers were assigned. The numbers are downstream of the decision, not upstream. The decision gets made, and then the numbers are arranged to justify it. So the actual play is: you land a job elsewhere with the title you want, and your current employer either lowballs you and you leave, or finally comes to the table with a fair offer. That’s it. That’s the entire system. Everything else is theater.

Meanwhile the promotions go to corporate darlings who lie to close deals and torch long-term business relationships. They go to the people who capitulate to the boss and ship a feature by cutting so many corners it’ll take sixteen other engineers a full year to refactor once more than nine people actually use it—because it wasn’t built to scale, wasn’t tested, wasn’t built on any design theory, and was quite possibly lifted from a hackathon belonging to the quiet engineer who had the idea first. I’ve watched this happen so many times I could set my watch by it. The corporate darling gets the title. The quiet engineer gets a LinkedIn endorsement for “creativity” and a two-percent raise that trails inflation. The company gets a codebase full of scar tissue and a reputation for broken promises. And somehow the quarterly earnings call still manages to sound optimistic. How?

All of this is just American corporate culture, and for most of my career I didn’t see much I could do to change it. It’s pervasive. It’s structural. And most of my career has been spent as an individual contributor—an IC, a person who builds things—with no control over these decisions. So I operated within the constraints. If you can’t change the system, make it work in spite of itself. That’s not cynicism. That’s engineering.

There’s a line in The Wealth of Nations where Adam Smith says that even selfish pursuits get led by an invisible hand toward the greater good—when they’re restrained by oversight. Smith did not mean this as a blanket endorsement of selfishness, whatever the think-tank crowd would like you to believe. He meant it as a description of a system that only works when the guardrails hold: when competition is real, when information is symmetric, when no single actor can rig the game. The invisible hand isn’t magic. It’s a mechanism, and mechanisms require maintenance. Remove the oversight and the hand simply grabs whatever’s within reach. Guess which half of Smith made it onto the tote bag.

DE&I is the perfect example. I’m a minority, and I’m late-diagnosed neurodivergent—ASD-1, what used to be called Asperger’s, the kind that lets me pattern-match nonsense faster than most people blink and also renders me constitutionally incapable of pretending to enjoy a team-building exercise. But I joined tech in the 2010s, when those things were a corporate death sentence. You didn’t mention being autistic. You didn’t mention being anything other than a default-settings human who loved ping-pong and “disruption.” Around 2014 or 2015, organizations started telling themselves Title IX mattered—not because they believed it, but because the legal pressure and the PR calculus had shifted. That seed grew into “diversity” by 2020 and eventually bloomed into full DE&I: diversity, equity, and inclusion. The branding was impeccable. The sincerity was nonexistent.

Corporations never gave a single thought to any of this on the merits. They are, at best, indifferent to women, minorities, neurodivergent people, and the LGBTQIA+ crowd. But the opportunity carried a single, gleaming benefit: we can expand the American labor force by double digits and land the most qualified candidates—since women in particular have made up the vast majority of college graduates since roughly 2010, a number that’s only grown—and we can pay them almost nothing. That’s the part that never makes the recruiting brochure. That’s the real business case. Diversity as arbitrage. Inspiring, isn’t it?

Let me give you a number. While I was making $140,000 as a senior engineer, my peer—same title, same responsibilities—was making $80,000. She held a master’s degree in computer science. I hold a master’s in nothing. She was not only more qualified on paper; she was a better engineer AND a far more level-headed IC. And she was paid nearly half what I was. The reason U.S. bosses don’t want you talking about your compensation isn’t to protect your privacy. It’s to fleece you, individually and collectively. It’s a structural information asymmetry engineered to keep labor cheap, and it works best when the people being underpaid don’t know they’re being underpaid. And the same logic drives the quiet hostility corporations express toward anyone content to stay in their current role without chasing the next rung. If you’re happy where you are—good at the job, adequately compensated, uninterested in climbing into management—the machine reads that as a failure of extraction. You’re supposed to want more, always, because wanting more keeps you pliable: you’ll work harder, self-exploit more efficiently, and accept deferred compensation in the form of a promotion that may never arrive. Contentment isn’t loyalty—it’s a refusal to play the game, and the game hates a player who’s already won enough. When did being satisfied become insubordination?

As I ascended into leadership roles with actual control over hiring, I understood perfectly that my company only wanted DE&I candidates because it wanted cheap labor. I wanted them because they were almost always far more qualified—and in the rare cases they weren’t more qualified, they were always as qualified. Open the aperture and the talent pool is extraordinary. It was never a pipeline problem. The problem is that the people doing the hiring had been selecting for cultural comfort over competence for so long they’d convinced themselves the two were the same thing.

This escapes the most fragile people in the conversation. I have a fiduciary obligation to staff my company with the best people available. I cannot hire a warm body for imaginary points—and who, precisely, is issuing these points? Where does the crowd convinced that DE&I means hiring incompetents on the basis of skin tone imagine the revenue is deriving from? This is still a business. The earnings calls still have to demonstrate growth. You cannot positive-growth your way out of a team full of unqualified hires, and nobody—nobody—was doing that. It was always a bad-faith caricature, a moral panic for people who needed someone else to blame for their own mediocrity. If diverse hiring were the anchor they claim, wouldn’t it show up in the numbers?

Besides, most engineering teams are small—five to fifteen people—and I’m not easy to work with as a neurodivergent person, because I don’t have much patience when people do things that are illogical or inefficient. I am not hiring unqualified people. I can’t work with them; my brain will chew through the drywall. DE&I gave me free license to hire the best candidates in spite of my company’s cheaper impulses. The invisible hand at work: the corporation’s appetite for cheap labor opened the door, and what came through it was a wave of systematically underpaid and systematically overqualified talent that, for one brief window, I could bring onto my team without re-litigating the same tired battle about “culture fit.” Culture fit, incidentally, is just a euphemism for “people who look and sound like the hiring manager,” and it has cost American business more innovation than any competitor ever could.

My real job, though, was to deprogram my hires. I’d pull them aside in the first week and say: “Look, this is your shot as a DE&I hire to get the experience that henceforth removes that label from your résumé, because you’ll have demonstrated competence. Also—hustle here for one to two years and then bounce, because you’re grossly underpaid.” That wasn’t disloyalty. It was honesty. The compact was broken before they ever signed the offer letter. My role was to make sure they extracted as much from the arrangement as the company was extracting from them, and then got out before the extraction turned permanent. I provided as much mentorship as people wanted, as many resources as people wanted. And much as I’d love to still be working with a lot of those people, the universal truth held: hired at X, and after one to two years, a twenty-to-thirty-percent median bump in pay elsewhere. I have the numbers to back this up. The system worked—just not the way the corporation intended. It worked because I hacked it, and because, for a while, they let me.

And then, the moment they could, they dropped the ruse. The indifference was always there, just beneath the surface. At best indifferent to women. You can work here—but now you need your own bathroom? A pump room? Time off for having a baby? What do you mean? Get back to work. Period leave? “When will the Marxist demands of the socialist communist agenda ever end?!” cry the executives now openly nostalgic for the days before “corporate skirts” were allowed into the office. They never saw women as equals, and after the 2024 election they threw as many of them off boards and out of executive rooms as they could manage, more or less overnight. I’m still awaiting the study quantifying the drop in women’s leadership representation post-2024. By 2022, women held something like two percent of Fortune 500 CEO roles despite being fifty-one percent of the population and sixty percent of the educated class since 2010—which, on the merits, makes women on average more qualified. But sure. Meritocracy.

Minorities? Same deal, and this clearing-house was even more of a fire sale, conducted with even less ceremony. America’s racial history has always run on a loop: failure to acknowledge, followed by stubborn insistence that it’s all been atoned for and “people are making a big deal out of nothing,” because “we fixed racism after the march on Selma”—or whichever epoch the speaker imagines racism was cured. Maybe it was when Mr. Rogers soaked his feet beside François Clemmons in the pool. That was 1993, by the way. The not-funny-but-funny part is how neatly that proves the point: racism was never resolved because it was never wholly addressed. The Supreme Court has cleared the way for several states to gerrymander themselves back toward something that rhymes with Jim Crow. In 2024 I watched an elected lawmaker get physically removed from the chamber where a vote was being cast for the district he represented. Essential workers—the landscapers, babysitters, janitors, food workers, farmers we called heroes during COVID—are being detained in conditions that are killing some of them. So, no: racism isn’t gone. It just changed its tie.

LGBTQIA+? This group became the generation’s most visible lightning rod for hatred. What astonishes me, beyond how little of it is anyone else’s business, is that when people earnestly ask this group questions, the group is overwhelmingly willing to answer—often in a way that’ll make you blush and smile, darling. If the “just asking questions” crowd actually wanted solutions, solutions exist for most of what they raise. But they don’t want solutions. They want someone to hate. The bathroom debate? Trivially solved. Spain sorted it decades ago: convert every bathroom to floor-to-ceiling locked stalls that maximize privacy, with a shared area for the sinks. No gendered rooms at all. One room, several stalls, several sinks outside them. Full privacy for everyone, full safety for everyone. As a bonus: less peeping, less graffiti, less floor space, since you only need one facility. The engineering is trivial. The politics are the problem. But hate doesn’t seek answers. Hate seeks victims. Which is it you’re actually after—the safety, or the victim?

Corporate culture in America has apparently swung back toward something like Mad Men, and whatever impulses your leadership felt obliged to hold back while Title IX protections had teeth, it’s now open season. So the exact flavor of nonsense is probably unique to each person. What’s not unique is that leadership no longer even seems strategic. It’s the war-crimes theory of management: we’re finally allowed to commit the atrocities! Oh—wait—the atrocities don’t actually help us win the war. Maybe that wasn’t what was holding us back. Maybe diversity wasn’t holding us back either.

Teams don’t appear to be infinitely more capable now that the pesky women and minorities have been thinned out. Shocking, I know. Here’s one thing worth saying: if you measure your leadership success by how few problems reach you and how rarely anyone objects, two things are true. One, you are not a good manager. Two, your team will fail. My team challenges me constantly. Sometimes I get frustrated—I’m human. But when I step back, I think of it as projectile physics. I have ASD-1. My brain is wired oddly. By listening to the people closest to the problem, by taking qualified objections from diverse perspectives at every experience level, by opening the floor to every teammate without ego, we make the decision together, we hold high ownership, and we can be reasonably sure that once the projectile is in motion it will hit the target. That’s not soft management. That’s ballistics. And you do not get ballistics out of an echo chamber.

The crux of the problem I have now—and here’s where the invisible hand becomes the invisible man, and the invisible man becomes an AI prompt—is that I can no longer listen to my team. Because the same person who decided that DE&I no longer matters also knows exactly what my team should be building, how we should build it, and how long it will take. And though I can’t prove it, it looks very much like AI has made all the calls. Not AI as some superintelligent oracle weighing nuanced trade-offs. AI as a large language model fed a handful of business-school clichés and the CEO’s own wishful thinking, obligingly handing back exactly what the CEO already wanted to hear.

AI tells you what you want to hear. AI tells you that you are the fairest of them all. AI will never defy you, my lord. AI faithfully reinforces every bias it was trained on—the same biases that fired the women and the minorities and the queer folks, the same biases that swapped strategic thinking for vibes-based management. It’s as though “trendslop” is running the show: running on fumes, and running us all ragged, because it’s running us around in circles.¹ The decisions have no memory, no accountability, no skin in the game. The AI says “build this feature,” the invisible man rubber-stamps it, the team works 60-hour weeks to ship a hallucinated product nobody asked for, and then we pivot to the next thing the AI dreamed up, and nothing—nothing—ever gets launched, scaled, or maintained long enough to matter. The careers we’re building are Potemkin villages, and the architect is a stochastic parrot.

The careers we’re building are Potemkin villages, and the architect is a stochastic parrot.

You can watch it happen in real time. The studies are starting to pile up. Researchers asked LLMs for strategic advice, and they got trendslop in return—plausible-sounding, superficially confident recommendations that collapse under even mild scrutiny, because the model is optimized for coherence, not correctness.² The model tells you to “leverage synergies” and “double down on core competencies” and “adopt an AI-first strategy” because those are the words that appear next to each other in the corpus of executive blog posts it was trained on, not because those words mean anything in your specific context. Trendslop is the perfect drug for the invisible man. It flatters his priors. It asks nothing of him. It supplies a plausible-sounding justification for whatever he was going to do anyway, and it does it in bullet points suitable for a board deck.

The upshot is that teams are run ragged chasing ghosts, while the people at the top are more convinced than ever that they’re steering the ship—when they’re actually just leaning on a tiller that isn’t connected to anything. The frantic, ad-hoc product launches I wrote about last time—the ones that spring fully formed from a CEO’s CNBC appearance—are now even more disconnected from reality, because the AI has blessed them. The invisible man’s shadow has acquired an AI-generated voice, and it’s telling us to ship the impossible on an impossible timeline, and every strategic objection, every piece of evidence from the ground, every diverse perspective that might have corrected course is simply ignored—because the AI didn’t mention them, and the AI is never wrong, because the AI is just a mirror reflecting the CEO’s own face back at him, softened with a Gaussian blur and captioned “visionary.”

I think about the people I hired during the DE&I window. The best of them are gone now—promoted into better companies, better titles, better pay. The ones who are left are fighting the same fight I am, trying to build real things in a system that no longer values realness. And they’re tired. I’m tired. We’re all tired, because we’re not just doing our jobs; we’re doing a second, unpaid job of navigating the chaos generated by an AI-driven leadership fantasy that shows up on no ledger, serves no customer, and builds no lasting value.

The invisible hand of Adam Smith required oversight—real competition, real accountability, real guardrails. The invisible man who runs your company now operates with none of them. And the AI he’s installed as his oracular successor is the final evacuation of human judgment from the boardroom. It has stopped being a tool. It’s an abdication.

The damage isn’t theoretical. It’s showing up in the résumés of people who’ve spent two years building features that never launched, in the burnout stats HR quietly buries, in the quiet exodus of anyone who still has the leverage to leave. The only thing keeping the machine running at this point is inertia and fear. And I don’t know exactly when the breaking point comes, but I’d wager it’s closer than the quarterly guidance suggests.

In the meantime, I do what I’ve always done. I operate within the constraints. I make the system work in spite of itself. I tell my people the truth. I document everything. I keep my résumé current. And I wait for the day the trendslop finally meets a problem it can’t bullshit its way past—a cash-flow statement, say, or a mass walkout, or a customer base that finally, finally locates the exit. What do you suppose happens to the mirror the first time it has to show him a number he doesn’t like?

---

1. ↩ See the growing body of work on “trendslop”: the phenomenon where LLM-generated strategic advice defaults to fashionable but empty jargon. Harvard Business Review covered it in March 2026: “Researchers Asked LLMs for Strategic Advice. They Got Trendslop in Return”. Jan Oliver Schwarz dissected the definitional contours on his Substack. Fortune weighed in with “What Is Trendslop? The Hidden Bias in AI That’s Giving Your Workplace Bad Advice”. And IBTimes UK noted that AI trendslop is increasingly shaping CEO decision-making. The signal is loud and getting louder.

2. ↩ The problem is not that LLMs can’t generate useful insights under careful human supervision. It’s that they are being deployed as oracles rather than as tools—a distinction that the executive class, in its current hospice-care myopia, seems incapable of grasping. The trendslop loop is self-reinforcing: the model outputs what the leader wants to hear, the leader acts on it, the action fails to produce results, but the failure is blamed on execution rather than strategy, and the leader asks the model for more advice. Rinse, repeat, vaporize shareholder value.