Does this help life on Earth?
The one question we measure everything against โ and how to tell us when we get it wrong.
The benchmark, in plain words
Before we build a feature, ship a page, or let an AI loose on a task, we ask one question:
Does this help all life on Earth?
That is not a slogan on a wall. It is a checklist. If a thing we are working on does not move at least one of the boxes below, we do not work on it โ or we flag it and say so out loud. Both answers are fine. Pretending is not.
๐ฑ Health
Does it help a person's body or mind? Does it help someone understand their own health without being sold something? Does it reduce harm?
Never: medical, legal or safety advice presented as fact. AI output is a starting point, not a doctor.
๐ง Clean water
Does it help people get, test, save or share clean water? Does it help the systems and volunteers who maintain it?
Water data is life-and-death data. It gets checked, sourced and dated โ never guessed at.
๐ Food
Does it help someone grow, find, afford or share food? Does it help a small farm, a kitchen, a food bank, or a community garden?
Including the unglamorous parts: yield numbers, spoilage, transport, and knowing when a price is a lie.
๐ Knowledge
Does it help someone learn something true? Is the source visible? Can they check it themselves without asking our permission?
Knowledge you cannot verify is just rumour with better typography.
๐ฆ Wisdom
Does it help someone make a better decision โ or just a faster one? Does it show trade-offs, or hide them behind a green tick?
Speed without wisdom is just efficient regret.
๐ก Works offline
Does it keep working when the internet does not? Can someone in a village, a clinic, a boat or a basement still use it?
This is why a plain git clone boots with no database, no internet
and no secrets. Offline is a feature, not an error state.
๐ And a bit of fun
Technology that is only useful and never enjoyable gets abandoned by the people who needed it most. So: a friendly voice, a mascot that is definitely a bee, and permission to enjoy using the thing.
Serious mission, light feet. Both at once is the whole idea.
๐งญ And spirit
Does it leave a person with more agency and less fear? Does it respect them โ their time, their data, their attention, their beliefs?
We do not tell anyone what to believe. We do insist that the machine is on their side of the table.
How to flag something that doesn't fit
You do not need an account, a login, or a theology degree. If something on this site or in the API does not belong under the benchmark above, say so and we will either fix it or stop doing it.
Every flag gets a real answer: fixed, flagged (we know, it's on the list), or deliberate (here's the reason). "We'll get to it" is not an answer.
Ask the API
The benchmark is machine-readable, so your own tools and agents can check alignment instead of taking our word for it. Same content, no login, works offline against a local gateway.
curl -s https://hivv.org/api/purpose
# list the checks, the verdicts, and how to flag
curl -s https://hivv.org/api/purpose/checks
curl -s -X POST https://hivv.org/api/purpose/flag \
-H 'content-type: application/json' \
-d '{"what":"the thing","why":"how it misses the benchmark"}'
Read the full contract in API docs ยท Partner API
What we do when something misses
- Fix it โ usually the right answer, and the boring one.
- Flag it โ mark it public, dated, with a reason, so nobody has to guess.
- Stop โ if it cannot be made to fit, it does not get built. Even if it would have been profitable, clever, or a really good demo.
The honest version: we will sometimes get this wrong. The benchmark exists so that being wrong is findable โ by you, without our help.
๐ Help all life on Earth. Hydrate. Pet something. Ship the boring fix.
Back to home ยท
Talk to the hive