Nostr only post: The AI job creation no one is talking about
"In partial darkness, a rope becomes indistinguishable from a snake."
- Vedic aphorism
LinkedIn published data in January showing AI has already created 1.3 million new roles in two years
(source:
https://www.weforum.org/stories/2026/01/ai-has-already-added-1-3-million-new-jobs-according-to-linkedin-data/ ).
So the "AI will create jobs" argument is settled.
It will.
What's not settled is what kinds of jobs. Most people are imagining more engineers, more data scientists, more prompt writers. Bigger versions of the same boxes.
The more interesting jobs are the ones that don't have names yet. The ones that exist because AI is reaching out and grabbing everything it can, in partial darkness, and someone needs to check whether what it's picking up is rope or snakes.
I'll explain what I mean.
I currently use AI to do the work of about 15 different functions. EA, project manager, editor, research analyst, product architect, social media strategy, content, coaching, IP management, accountability, among others. I've lost count of the exact number, which is sort of the point.
The result? I've created more human roles, not fewer (the first: a business growth coordinator).
I didn't expect that. I assumed it would be an efficiency tool, that I'd do the same things, just faster. Instead, because the overhead disappeared, I could see opportunities for people that I couldn't see before. I've already created a new role and increased someone's hours. The new role exists because AI freed up enough bandwidth for me to notice it was needed.
Jensen Huang said you're only making people redundant if you lack imagination. source:
Jensen Huang says CEOs ‘out of imagination’ for culling workers because of AI. Why he’s doubting his biggest customers
Of course, an AI chip manufacturing company would say that, but I believe he's 100% right because it matched what I was experiencing. The companies cutting headcount with AI are telling you something about themselves They don't know what to do with their people when the busy work goes away.
And people can tell when that's the plan. More on that later.
I read this week about websites that can detect whether a human or an AI is reading them. And they've already developed three distinct ways to stop competitors using AI to scrape insights from their content. Some strip down the content. Others deliberately feed misleading information to poison AI training data.
Cloudflare launched something called AI Labyrinth (source:
The Cloudflare Blog
Trapping misbehaving bots in an AI Labyrinth
How Cloudflare uses generative AI to slow down, confuse, and waste the resources of AI Crawlers and other bots that don’t respect “no crawl” ...
) - when it detects an AI crawler, instead of blocking it, it redirects the crawler into a maze of AI-generated pages that look like real content. The crawler keeps reading, keeps indexing, and none of it is real.
The AI reaches for what looks like information. It's not always information. Sometimes, your AI thinks it's found a bunch of rope to build stuff with, but your competitor is feeding AI a container-load of snakes designed to look like rope that, without some careful snake-wrangling, will poison your entire data-intelligence and cause you to make bad business decisions.
Then there's the AI-on-AI problem. As companies deploy more AI agents, it's inevitable that some of those agents will be rogue - designed by a competitor or a bad actor to exploit the blind spots of your AI. Your agent trusts the data because it looks clean.
The rogue agent was built specifically to look clean to your agent. No AI can reliably spot an adversary that was engineered to sit in its blind spot. You need a human for that - someone whose job is to watch the agents and ask the questions that neither agent is asking.
Meanwhile, employees are doing the same thing at their desks. 31% admit to actively sabotaging their company's AI strategy (source:
31% of employees are ‘sabotaging’ your gen AI strategy | CIO
). And the most effective sabotage is the quietest kind.
Companies across China started ordering workers to document all their knowledge as AI "skill files" - the plan being to replace those workers with AI agents trained on their own expertise (source:
X (formerly Twitter)
Milk Road AI (@MilkRoadAI) on X
This is WILD.
A secret workplace war just broke out in China and it has gone fully viral on GitHub.
Companies started ordering their workers to d...
).
The workers figured it out fast. Someone built a GitHub tool called colleague.skill that scrapes a coworker's chat logs, emails, and work docs, then clones them into an AI agent.
The logic " digitize your colleague before they digitize you", give the clone to the company, and watch your coworker get laid off while you survive. Ouch.
Then someone else released anti-distill.skill - a tool that takes the skill file your company forces you to write and strips out every piece of real knowledge before you hand it in.
The output looks professional, detailed, impressively complete. But every critical insight has been quietly removed. It even has three intensity levels depending on how closely your bosses are watching. It went viral on GitHub within hours.
Same dynamic, different scale. When people believe AI is being used against them, they corrupt the inputs.
So now you need someone whose job is to check whether the AI is operating in partial darkness. Someone who triages: is this reliable, or is something being hidden? That role doesn't have a name. It didn't exist a year ago. It's already overdue.
No one knew what a "social media content manager" was in 1998. Not only did the job not exist, the entire industry that recruited that category of job didn't exist. Two decades later it employs millions of people.
I think we're at that point again - roles forming that don't have names yet. All of them requiring the one thing AI cannot provide, the ability to see what AI is not being shown.
The question everyone keeps asking is "which jobs will AI take?"
I'd start with a different question. What could your people do if they stopped doing the things AI does better?
AI can grab a lot of ropes and snakes very fast. Among the myriad of new categories of jobs that will get created, we're entering an era where we're going to need a whole lot more snake-breeders ... and snake-charmers
