Technology that works for the public: a builder’s case.

For more than a century we have known how to build things the public can count on — roads, water systems, libraries. We built them to a standard, staffed them with skilled people, and governed them in the open. The software that now runs public life deserves the same treatment.

Most technology aimed at the public is judged by whether it ships. Public works are judged by whether they hold — for everyone, for years, under scrutiny. That difference in standard is the whole argument.

Treat software like public works

When a benefits portal fails, it doesn’t fail quietly. It fails a person who needed it, often the person with the least room to absorb the failure. A public work is accountable to exactly those people. Public-interest technology should be too: durable, maintained, and answerable to the communities it touches.

If the public can’t understand it, inspect it, or push back on it, it isn’t a public good — it’s just software pointed at people.

The builder’s case, in three commitments

• Build to a standard. Collect the least data possible, aggregate by default, and make every result reproducible — the way a financial close is reproducible.
• Prove it. Ship the receipts: open methods, audit trails, and numbers anyone can reconstruct. Protections you can verify beat assurances you take on faith.
• Sustain it. A build is never finished at launch. It needs a workforce to maintain it and rules to keep it honest long after the contract ends.

None of this is exotic. It is ordinary engineering discipline, applied with the public in mind. The case for it is simply that the public is worth building for.