Superhuman의 AI 초안 60%는 수정 없이 사용 — 그래도 아직 못 믿는다
이메일 클라이언트 Superhuman이 OpenAI와 Anthropic의 프론티어 모델을 활용한 새로운 자동 초안 작성 기능을 출시했습니다. 발송된 초안의 60%가 수정이 전혀 필요 없다고 주장하지만, 정확도는 여전히 들쭉날쭉합니다.
email client superhuman released a new version of its auto-draft feature, utilizing a mix of frontier models from openai and anthropic to generate replies that mimic user tone. beta testing data indicates that 40% of drafts were sent within 24 hours, and 60% of those were sent without any edits. the system learns from user behavior over time to correct mistakes, such as refusing unreasonable meeting times, but still requires oversight. this shift moves beyond older gpt-3.5 implementations to offer higher intelligence, signaling a trend where ai agents handle increasing amounts of routine communication.