You Can't QA an LLM the Traditional Way
Traditional test cases assume the same input always produces the same output. Language models don't work like that — and Contexta is a case study in what you have to test instead.
Read Article
Exploring technology, design, and digital innovation
Traditional test cases assume the same input always produces the same output. Language models don't work like that — and Contexta is a case study in what you have to test instead.
Read Article
AI can help you build an app. Getting it App Store ready — branding, paywalls, ads, subscriptions, notifications and review — is a different job, and Thai Sabai was all of it.
Read Article
Lock screen widgets are some of the most valuable real estate on a phone — and paired with the right marketing, they can do more for an app than a feature list ever will. Kanji Loop is built around that bet.
Read Article