SuperhumanのAI下書き、60%が無編集 — それでもまだ信用できない
メールクライアントSuperhumanは、OpenAIとAnthropicのフロンティアモデルを活用した新たな自動下書き機能を公開した。送信された下書きの60%は編集不要だと主張するが、精度は依然として不安定だ。
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.