Do you actually care what a panel of “experts” thinks about your job security? Yes, but only if they’ve actually written a line of code or managed a production cluster in the last five years.
The career advice trap
There is a specific kind of vertigo that comes from reading “thought leadership” about AI. It’s the feeling of watching a cooking show where the chef spends forty minutes talking about the philosophy of salt but never actually turns on the stove. WIRED is inviting people to submit their questions for a livestream about how AI is altering work, and while the intent is fine, the format is fundamentally flawed. These panels almost always devolve into a vague consensus that we need to “adapt” and “upskill,” which is the professional equivalent of telling someone to “just be better” at their job. These sessions are designed for the concerned observer—the middle manager or the anxious journalist—rather than the person actually staring at a terminal.
For the people building the tools, the “now what” isn’t a question for a panel; it’s a question for the documentation. The real friction isn’t a lack of career guidance—it’s the fact that we’re fighting with context window limits and paying an arm and a leg for H100s while the “experts” are still debating whether the prompt is the new programming language. (Most of us know it isn’t). We are paying exorbitant token costs to maintain a state that should be trivial, and yet we are expected to focus on the “societal shift.” It is hard to worry about the macro-economic trajectory of the labor market when you are currently trying to figure out why a model that claims to be smarter is suddenly hallucinating JSON syntax in your production pipeline.
The obsession with the “future of work” as a conceptual puzzle is a distraction from the actual work. The shift isn’t happening in a boardroom or a curated Q&A session; it’s happening in the IDE. We’ve seen this cycle before with the cloud migration era, where consultants spent years telling companies how to “move to the cloud” while the engineers were the ones actually figuring out why the latency was killing the app. The consultants got the speaking slots and the fancy titles, while the engineers got the 3 a.m. PagerDuty alerts. Why ask a panel how to survive when you can just look at the GitHub commit history of the most successful projects in the space? The evidence is in the code, not in a livestream.
We are currently stuck in the “existential dread” phase of the hype cycle, where every article is a variation of “will a bot take my desk?” but that phase is ending. By Q3, this discourse will have shifted entirely away from job security and toward the technical nightmare of agentic orchestration and hallucination management. We’ll stop asking if the AI is coming for our jobs and start asking why the autonomous agent we spent three weeks building just deleted the production database because it misinterpreted a “cleanup” instruction. This kind of public forum is mostly theater; it provides a sense of participation in a change that is actually moving far too fast for a scheduled livestream to be relevant. By the time the panel answers a question submitted today, the underlying model version will have likely changed, and the “expert” advice will be a snapshot of a world that existed three weeks ago.
The only real way to figure out “now what” is to keep breaking things until they work.
These panels are just expensive ways to state the obvious.
