Authorized After the Fact

Two women in this batch of stories, working in fields that have nothing to do with each other, keep circling the same problem: authorization. One spent her first year seriously studying AI writing governance frameworks around a single question: who authorizes an AI agent to act? A few stories later, another is asking almost exactly the same thing about the autonomous systems her company tests: not whether they're capable, but whether anyone has said yes yet, and who gets to say it. Neither is writing about her own career. Read back to back, though, it's hard not to notice they're describing the same problem twice: once for the machines, once, quietly, for themselves.

Elsewhere the question loses its technical vocabulary and gets blunter. One woman walks into cybersecurity with a degree in French and Economics and is told her certifications don't count, until, eventually, they do. "Who was born knowing?" she asks, a fair question for a field that kept raising the bar for newcomers right after it had cleared for the people setting it. Another, still debugging her own Terraform files, decides she doesn't need to reach some finish line before she's allowed to mentor someone earlier on the same road. A third builds an app for women whose doctors decided their pain wasn't documented enough to investigate.

Not everything here is about that. Plenty of this issue is just people doing exacting, specific work: patching cloud security debt, mapping vendor risk, arguing carefully about what the AI bias research actually shows. But often enough, working from opposite ends of the industry, the stories arrive at the same shape of answer: capability shows up first. The sign-off, a certification, an endorsement, a green light from someone with the standing to give it, arrives later, if it arrives at all.

Nobody in these pages waited for it before starting. Some of them are still waiting, and building regardless.