What Thatch and Lath Were Never Told

What the Program Actually Funded

The resilience program funded three new systems at once — Thatch, Lath, and Brace — each building toward the same eventual load. Its rule was simple and, on its face, reasonable: go live within the deadline window, and continued funding followed automatically. Miss the window, and funding reverted to case-by-case review.

Nothing in the program's public terms mentioned structural robustness at all. It measured one thing precisely — time to first live traffic — and left everything else to whatever judgment each system's own team brought to the build.

What Thatch and Lath Chose, and Why

Thatch went live first, in the minimum viable configuration the deadline rewarded — fast, cheap, structurally untested beyond what the window allowed. Lath took a middle path, live a few days later with somewhat more hardening, still inside the bonus window by a narrow margin.

Brace alone let the window close. She spent the extra weeks on load paths and failure isolation the other two never had time to build, and watched her own funding revert to case-by-case review — the one visible, immediate cost of the only choice that would matter months later.

What the Surge Actually Tested

The Surge, when it came, tested exactly what each build had actually invested in. Thatch failed within minutes. Lath held longer, then failed under the same sustained load Thatch had never survived to meet. Brace's structure absorbed it without needing to shed a single dependent service.

The program's first incident report read the outcome as three individual engineering judgments, two of them poor. It did not ask what had shaped those three judgments in the first place, because nothing in the report's own template had a field for the question.

What the Incentive Actually Rewarded

A second pass, prompted by someone who bothered to reread the program's own funding formula, found that Thatch's and Lath's choices weren't poor engineering at all — they were exactly what the formula, as written, had told every team to optimize for. Brace's choice was the actual outlier, financially penalized in real time by the same rule.

Nobody at the program had lied to Thatch or Lath about what was being measured. Nobody had told them either, in terms either could have used to weigh a robustness they'd have needed to give up an active bonus to build — until a stress event with no announced schedule made the omission visible.

What Thatch and Lath Were Never Told

The funding formula was rewritten to decouple robustness investment from the go-live deadline entirely — a system could take the speed bonus and separately apply for a hardening extension on its own schedule, neither penalizing the other. The incident report was reissued, correcting the record: Thatch's and Lath's builds had been rational responses to what the program itself had chosen to reward.

Brace's structure hadn't proven her more careful than the other two. It had proven that the program had never told any of them, in a form that could actually be weighed, what speed was going to cost — until something arrived that made the bill impossible to keep hidden inside a formula nobody outside the program had ever had reason to read.

A choice made under a formula nobody has to justify isn't poor judgment. It's the formula's judgment, wearing someone else's name.