A few weeks back I dedicated two issues of the newsletter to the AI Risk Weather Report – here's part 1, and part 2 – as a snapshot of risks and opportunities in the AI space. I kept that high-level in order to maintain a readable page count. (Believe it or not, that 5,000-word beast was the short version.)

That should be my cue to shift back to my usual coverage of AI-related news. Like, say, model providers setting their bots loose on the public internet for running tests. Or the ECB's note on the potential impact of a genAI market correction. (My take, since you asked: a sudden, sharp correction would cause widespread economic damage. Enough to give us a taste of Russia's "Wild Nineties.") Or datacenters turning into political poison. Or any other topic.

Instead, today I'm running an essay on a risk that stems from AI in the workplace. It's subtle. It's quiet. And has the potential for long-term impact.

* * *

A silent poison

AI-based risks in the workplace usually start with over-eager executives. Their company-wide usage mandates create the double effect of frustrating team members and torching tons of cash. Some execs have gone as far as to sideline trusted human colleagues by turning a chatbot into their advisor, confidante, or chief of staff. (Science may eventually pull that behavior under the "AI psychosis" tent. Until then we can only smile and nod.)

That forced-from-the-top genAI interaction pairs with job destruction. The immediate concern involves companies that cut existing headcount in favor of robots, even though said bots are rarely capable. Lurking in the background is the longer-term issue of assigning entry-level work to the machines. That may save a little money in the short run – again, assuming the bots are up to the task – but will eventually bleed a field of its talent. Once today's senior-level hires retire, who will replace them? At the current rate, not genAI bots. Not by a long shot.

That leads to the subtle issue I mentioned earlier: a quiet brain drain is forming as experienced tech practitioners, frustrated by genAI-related issues, leave the field early. It's the reverse of your typical mania- or bubble-driven talent displacement: instead of people leaving their existing roles to join the hot new field, these tech professionals are already in the hot field and are beating a path to the exit.

I'm frankly surprised that I didn't see it coming. But it keeps cropping up in conversations with peers. They're simply tired of genAI and what it's done to their workplace experience. Some have already been laid off and are trying to not return. The rest are laying the groundwork for their next steps, either waiting to be shown the door or preparing to leave before the axe swings.

This time, it's different

Granted, the tech sector has lost experienced minds before. Some practitioners, having amassed a wealth of domain knowledge within their industry vertical, hang up their spurs and migrate to other roles in that field. Like, say, a lateral shift from "software developer who works in the publishing industry" to "an editor with deep knowledge of software."

Then there are the people who launch businesses which leverage their tech skills under the hood, but aren't overtly tech-focused. It's as though they're tired of trying to help other companies be effective in their use of software or machine learning, so they're doing it for themselves. (You know who you are. I salute you. And for one of you in particular: I owe you a call.)

But the genAI-driven brain drain is less of a pull into other fields and more of a push out of the current tech role. Of the people I've encountered, one segment is frustrated by bots taking on more of their work. Especially since the bots don't do the job well, and they have to go back and clean up after their ersatz digital replacements. Another segment is tired of working under leadership that's trying too hard to force genAI into being. Even worse, they know that their next job hunt will raise the other spectre of tech-sector brain drain – ageism – so rather than starting that quest with two strikes against them, they're self-selecting out of the pool altogether.

(It's funny how the tech sector keeps pushing people out right as they've developed enough experience to spot patterns. That might be why tech keeps Columbusing its way through problems, "discovering" half-baked approaches that had been solved several times before. And solved better, at that. But I digress…)

Some plan to get as far away from tech work as possible and never touch it again. Others have the idea to take a so-called "dull" business that hasn't been overrun by tech, then use their technical skills to automate certain tasks and drive efficiencies. That's a shaky business model for a general-purpose, mass-market tool; but it's quite viable for building tools that only you will use.

Not the only ones

This is when you'll point out that the older portion of the tech sector is mostly genX and elder millennials – people who were already in their prime "what the hell am I doing with my life?" years. What if they were on their way out anyway, and genAI frustration is a convenient excuse? Perhaps. For some. But certainly not all.

You may also claim that this is a small-N problem, and that my argument rests on a selection bias. Maybe, maybe not. As I started noodling on the topic that would eventually become today's newsletter, I've brought up the idea in other conversations. Those people, too, are hearing murmurs of departure within their professional circles. For the same reasons. And with the same destinations i mind.

I suspect this extends well beyond my professional network, too. Since the core problem is "people in tech who are tired of overhyped tech bullshit," then we can look to the Meta and Google employees who have openly expressed frustration with their employers' genAI attempts. Plus there are the countless tech professionals who are less vocal, but still roll their eyes at the AI rollouts that are doomed to fail. Or those who intentionally burn LLM tokens so they can comply with execs who track token usage as part of adoption mandates. Some percentage of that group will also leave the space.

A one-way road

Then what? Since tech workers tend to earn well, those who have squirreled away their cash are positioned to take a leap and try something new. They're in an interesting position to apply their know-how – not just of "technology" but "how tech companies operate" – to new domains.

Years ago I posited that if Wall Street quants were to apply their talents to other fields, they'd give new-age data scientists a run for their money. That story has yet to play out – Wall Street's golden handcuffs are strong indeed – but we might see a version in which technology workers migrate into new fields.

Consider what I said about finding ways to tech-ify parts of a traditionally non-tech business and reap those efficiencies. If those ex-developers do well enough, they may package up their homegrown tools into products that other businesses in that vertical would find useful. (Let's face it: some areas have seen little technology advancement, since larger tech providers didn't think they amounted to enough of an audience.) While that would essentially be a return to technology work, one could argue that it's not quite a return to the technology field.

And that spells trouble for the companies that have driven them out. What if, in their post-genAI hangover, these former employers are unable to woo back the disaffected senior-level tech talent? You know, the people who have internalized a couple decades' worth of lessons and best practices? Writ small, those companies will suffer because they'll teleport themselves back to the early Dot-Com days when everyone had to figure everything out from scratch. Writ large, then, the entire technology field will be worse off – forced to re-learn a lot of lessons that the old pros figured out years ago. And anyone using tools built by those companies suffer.

In other news …

For more links to recent news, and with a slightly broader scope, I encourage you to check out my other newsletter. It's a weekly, curated drop of what I've been reading.