The answer isn't technical anymore, it's political.
The Summary
- The conversation about AI and work has shifted from "what can the technology do" to "what should it be allowed to do", a question that has no engineering answer.
- Some jobs are changing so rapidly that interview processes can't even keep up, creating a lag between what employers ask for and what the role actually requires.
- The fight isn't about capability anymore. It's about who decides how AI gets deployed, and whose work gets automated first.
The Signal
Sarah O'Connor at the Financial Times frames the central tension: we've moved past the question of whether AI *can* do knowledge work. Of course it can. The real question is whether it *should*, and who gets to decide. That's not a question for engineers. It's a question for workers, managers, and the people who will live with the consequences.
Meanwhile, specific roles are mutating faster than hiring processes can track. Companies post job descriptions for skills that were relevant six months ago. Candidates prep for interviews testing knowledge that's already been automated away. The gap between what a job *was* and what it *is* keeps widening.
"The real question is not what the technology can do but what it ought to do."
This mismatch isn't just inefficiency. It's a signal that the entire structure of work is being renegotiated in real time. Job descriptions used to be stable enough to last a hiring cycle. Now they're snapshots of a moving target. The interview process, already slow and bureaucratic, can't keep pace with roles that change week to week as new AI tools ship.
What we're seeing is two parallel shifts:
- Technical: AI can now do parts of almost every knowledge job
- Structural: The institutions that define work (HR, job postings, career ladders) are too slow to adapt
The people fighting for the future of work aren't arguing about whether LLMs are smart enough. They're arguing about power. Who decides which tasks get automated? Who benefits when productivity doubles but headcount doesn't? Who gets retrained, and who gets managed out?
The Implication
If you're hiring, your job descriptions are already out of date. If you're job hunting, the skills listed in the posting probably aren't the skills that matter most. The real skill is knowing how to work *with* AI tools that didn't exist when the role was defined.
Watch the people, not the tech demos. The companies that figure out how to redesign work around AI, not just plug AI into old workflows, will pull ahead. The ones still hiring for pre-AI skillsets will spend the next year confused about why nothing works.