Choice, Score and Noul: three question types
How to build a Choice, Score or Noul question, what the answers mean, and why a score is not a precisely measured quantity.
Open guideExplanations, how-tos and practical proposals for developers and online shop operators.
How to build a Choice, Score or Noul question, what the answers mean, and why a score is not a precisely measured quantity.
Open guideTypeSafe, OpenRouter and Vercel: different interfaces to the same model, described without promises of unverified credits.
Open guideWhen to consider Laya, a traditional classifier or a generative model instead of Jev, and what a fair comparison has to include.
Open guideThe direct TypeSafe API, the route via OpenRouter and a secure adapter design that keeps the API key on the server, not in the browser.
Open guideTechnical limitations, privacy, contract terms for customer data and telemetry, and the questions to settle before a live Jev pilot.
Open guideA dated price list for Jev, the direct API’s token limits and an input-token cost calculator, without misleading comparisons.
Open guideA plain-language explanation of Jev by TypeSafe AI and of the difference between a model’s decision and the execution of an action.
Open guideHow to read Jev’s probabilities and confidence, set thresholds for automatic decisions and design human review into the workflow.
Open guideProposed pilots for online shop catalogs, product search and customer inquiries, with no false promises about automation.
Open guideA decision adapter in agents and developer tools as a supporting signal, not a replacement for permissions and approval steps.
Open guideAbout JEV.sk: independence from TypeSafe AI, how we work with sources, and the limits of the demonstration tools on this site.
Open guideState, criteria, calibration, escalation, fallback, p50 and p95, and other decision-model terms explained in simple language.
Open guideA practical plan for a Slovak-language test set for Jev, with separate data splits and a fair evaluation of the results.
Open guideHow to tell vendor data, a provider’s measurements and future JEV.sk tests apart, and what the published numbers on Jev prove.
Open guideTypeSafe AI’s founders and seed investment, and why we keep company announcements separate from evidence of the model’s quality.
Open guideNew preprints on Jev and what they focus on, with a clear line between reading an abstract and a verified replication.
Open guideNo article matches your search. Try a shorter term.