Move over McKinsey: The next layoffs will be powered by AI
Letting HR vibe code is like giving a gun to a toddler
Recent articles have included pro tactics I’ve been using to build multiple SaaS within 10 weeks for myself and clients.
But what happens when AI falls into the wrong hands?
You’re thinking Russia, North Korea, Iran? No.
I’m thinking HR.
The next layoffs won’t be televised, but they will be powered by AI. Buckle up.
Hire me as CTO or tech lead for your next project.
Focus on growing the business, never worry about the tech again.
Slide into my DMs on X or Substack. Select portfolio at Deca Labs.
Like chimps discovering a fork
Before the current era, HR was largely limited in their ambitions. They had to wait for new functionality to come years later in their HR software. Workday. Greenhouse. SmartRecruiters.
Since they couldn’t build anything of their own (tempermentally expected given it’s HR), they were handcuffed to the limited capabilities that SaaS providers offered them.
But, in the age of AI, those shackles are falling off.
In some cases HR is learning to vibe code themselves. It’s like watching a chimp learn to use a fork.
But, in many more cases, internal software engineers are naively giving HR the rope with which they’ll soon be hanged.
Grow your Substack with Poaster.App
Use Poaster’s custom AI model to pull the best quotes from your long form writing, and automatically post them to your socials.
Skip the army of virtual assistants, start today from $1/month.
Class traitors that can code
Internal software engineers who casually help HR are best seen as class traitors in the communist sense.
While HR may come and nicely ask for help, handing the nukes over to Somalia because they ask nicely is still not a sane or responsible use of your talents.
Consider the following example.
HR is in charge of performance reviews annually. Quarterly is too much work. They encourage engineering managers (EMs) to comb through their individual contributors’ (ICs) work scattered across Github, JIRA, emails, Google Docs and summarize (with links back to the evidence) their work and impact. This is often weeks of work per EM if they have the usual 8-15 direct reports.
Performance reviews feed both into promotions, compensation, raises, and into layoffs.
“In the age of AI, surely this can be automated!”, one bright but naive software engineer exclaims, and proceeds to build a script which can provide in-depth performance review reports for each person in the company with working links, stats, and their stack rank position across their team, org, and company.
Feeling oh so accomplished with their 4 hours of vibe coding and refinement, they run a test batch of reports and send out a demo to the company for how easy it is now for anyone to leverage AI to update their hype doc!
But rather than broad thanks, the response from colleagues is muted. And within an hour the backlash has reached their org’s director who tells him to pull down the saved reports for now to stop the complaints.
Social credit scores approach, in every domain
But why were the performance review reports pulled offline?
Why were people upset?
Why didn’t they want an easy way to update their hype docs?
Obviously, such great power in the wrong hands will be abused.
And now that the repo and reports were published, the genie can’t go back in the bottle. The EMs know, HR knows, even the ICs now know that a new age is upon them.
What was far too costly to do manually regularly, can now be done monthly, weekly, even continuously, at minimal cost by AI.
And without the sufficient friction to stop abuse, whether it’s making a million microservices or excessive test coverage or now automated performance review reports, it will be abused.
Taking these trends forward, what does this mean for the corporate surveillance state?
Performance reviews will become continuous functions. HR dashboards will now have alerts when your weekly PR counts or throughput start to trend down. Performance Improvement Plans (PIPs) will start quarters early, not needing to wait months for EM documented infractions, AI will always have the evidence.
Come layoffs, it won’t just be a bunch of sweaty McKinsey consultants in a conference room drawing a line on an Excel sheet sorted by compensation + most recent performance rating. Now HR and management will have dashboards where they can easily model cutting the bottom X % based on realtime employee performance.
While Bridgewater, Ray Dalio’s hedge fund, received endless scorn when it came out that they had manually implemented a continuous evaluation system before AI, forcing employees to rate each other throughout meetings on iPads, that system of control and fear will now be scalable across every company, without the cultural friction of needing to force everyone to manually rate each other constantly.
And as many AI fear mongers have now rightly predicted, the age of AI will be used far beyond simply mass realtime surveillance, but conceptually as a way for the elites in every hierarchy to finally be freed from the burden of accountability.
Management has already accomplished that in a small way by outsourcing layoff decisions to consulting firms like McKinsey, but in the age of AI this will extend to every major decision.
Promo? Sorry, the AI said no.
Raise? Sorry, the AI said no.
Laid off? Sorry, we can’t override the AI.
Social credit score dropped? Sorry, the AI said to reduce it.
Health insurance claim denied? Sorry, the AI said no.
MAID (medically assisted suicide) opt-out request denied? Sorry, the AI said no.
Thought crime appeal? Sorry, the AI said no.
Public government contracts with Palantir for AI enabled surveillance systems, dashboards, and data ingestion and normalization services already point to that coming future.
For smaller organizations though, you can choose where to apply your talents.
And if you decide to naively give HR and management the scripts to do AI performance evaluations, don’t be surprised when they use the rope you gave them to eventually hang you for false crimes.
What’s the upside?
Well, AI SEO is still in the early innings…
Keep reading with a 7-day free trial
Subscribe to BowTied Fullstack to keep reading this post and get 7 days of free access to the full post archives.