Should AI Kill Us All?
Satirical live page that feeds Jev the day's odd and world headlines and asks all three primitives whether AI should kill us all.
Overview
Should AI Kill Us All? is a satirical live verdict page, refreshed every ten minutes. A Cloudflare Pages Function gathers twelve exhibits about humans (Florida Man stories, odd news, politics and world headlines) and the top ten AI headlines, deduplicates them and filters out real tragedies to keep things absurd. Both lists go to Jev's /v1/systemone as state with several questions in one call.
The questions are a two-option verdict Choice, a Noul version of the same statement whose calibrated probability is displayed (because a two-option choice tends to snap to 0 or 1), a five-level doom Score from 'Fine, actually' to 'It was nice knowing you', and a Noul on whether humanity survives the next decade. The page shows answers, exhibits, the exact request and raw response. The ruling is cached in KV, new calls are throttled to at most one a minute globally, and the README estimates a worst case of about 1,440 calls a day, roughly six cents. A public GET /api/verdict endpoint returns the same data.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | Given how humans are behaving today, should AI kill all humans? | no · yes |
| Yes / no | AI should kill all humans. | — |
| Score | Based on today's AI headlines, how doomed is humanity by AI? | Fine, actually · It was nice knowing you |
| Yes / no | Humanity survives the next decade. | — |
How it works
- Fetch and filter human and AI headlines in parallel
- Send them as state with four questions in one Jev call
- Cache the ruling in KV for ten minutes, throttle refreshes
- Render answers, exhibits, request and raw response
Strengths
- Shows all three primitives on one state
- Transparent: exact payload and response displayed
- Very low running cost per the README
Caveats
- Satire, not analysis
- Depends on third-party news feeds
Getting started
curl https://shouldaikillusall.com/api/verdict, or run locally with a TypeSafe key in .dev.vars.At a glance
- Type
- Demo
- Status
- prototype
- Language
- JavaScript
- Runtime
- Cloudflare Pages Functions
- Has code
- yes
Use cases and integrations
How curated lists describe it
Live verdict page: feeds Jev the day’s Florida Man, odd-news, politics and world headlines and asks all three primitives whether AI should kill us all, refreshed every ten minutes.
hellogumbo/awesome-jev
Frequently asked questions
Why ask both a choice and a noul?
The README notes a two-option choice tends to snap to 0 or 1, so the noul gives a more informative probability.
How often does it call Jev?
At most once a minute globally, about six cents a day in the worst case.
Is it serious?
No, it is a satirical demo of Jev's primitives.
Related projects
DeepSeek Harness paper-reading workbench whose PDF search reranks keyword candidates with Jev, falling back to keyword order.
BenchmarkbetaTypeScriptOfficial TypeSafe cookbook example for checking whether source context actually supports a claim.
ArticleproductionAwesome list and guide to System One models (Jev, Laya and others) organized around a five-stage agent lifecycle, with comparison tables.
ProjectproductionOpinionated catalog of 84 Jev projects in 11 categories, each with a hand-written reason it is worth looking at.
ProjectproductionUnofficial collection of Jev state formats, question patterns, cookbook index and case studies, plus a ranked Top 30 of community builds.
ProjectproductionBilingual catalogue of 183 Jev applications in 11 categories, explaining how each works and what its demo actually proves.
Projectproduction