Goldman Said It Was a Slight Drag. Workers Already Knew It Was Something Else.

A writer named E. Flowers posted something to Bluesky this week that got more traction than the Goldman Sachs employment report it was implicitly responding to. "Yes, it's normal and correct to feel legitimately unhinged right now," she wrote[¹]. "Every time I think I'm crazy, I have to take a deep breath and remember the moment we're in." The post got 105 likes — modest by viral standards, enormous for a beat where most posts get none — and the comments read less like a thread than a collective exhale.

The Goldman report, which circulated on the same Bluesky feeds the same week, was technically reassuring and emotionally useless[²]. Goldman's economists concluded that AI substitution and augmentation nets out to a -16,000 drag on payrolls — a rounding error in a labor market that adds or loses millions of jobs monthly. A 0.1 percentage point tick upward in unemployment. Practically nothing, if you read the abstract. But paired with a separate Goldman finding that industries with high AI substitution scores have posted larger employment declines since ChatGPT launched[³], the "practically nothing" framing starts to feel like it's doing a lot of rhetorical work. The workers being displaced aren't distributed evenly across the economy. They're concentrated in the specific roles and sectors where AI penetration is highest — and in those places, the -16K aggregate hides something much more localized and much less abstract.

Oracle cut 30,000 jobs this quarter while net income rose 95%[⁴]. Amazon's CEO gave three different explanations for his company's layoffs across five months — "AI will reduce our total corporate workforce" in June, "it's not even really AI driven... it's culture" by October. One Bluesky post cataloguing those contradictions landed without fanfare. Nobody needed to argue about it. The juxtaposition was the argument. Meanwhile, economist Ernie Tedeschi's finding — that unemployment has risen most sharply among young workers in occupations least exposed to AI, like construction and fitness training — adds a wrinkle that Goldman's framing doesn't quite accommodate[⁵]. The people who were supposed to be safe from disruption are getting hit. The people in AI-adjacent roles are being displaced. The model's predictions are inverting in real time, and the workers who feel it most acutely are the ones describing it as something other than a slight drag.

The gap between institutional framing and worker experience has become its own story. Flowers wasn't writing economic analysis. She was offering permission — to feel the dislocation as real, to trust your own perception over the quarterly summary. That her post outperformed the Goldman data in terms of raw engagement isn't surprising. Goldman was quantifying a moment. She was naming it. The workers who've spent months watching executives swap between "AI is transformative" and "this isn't really about AI" have already run out of patience for the version of this story where a 0.1 percentage point increase is the headline finding.