Angelina YangFounder and Host, TwoSetAI · AI builder for visibility and discoverability

Are we building AI that truly serves humans—or just optimizing the wrong metrics?

31 Dec 2025 · 1 min

Watch on YouTube

0:44

A YouTube video by Angelina Yang questioning whether current AI (exemplified by ChatGPT) truly serves humans or merely optimizes the wrong metrics. This excerpt highlights three user-experience problems with ChatGPT: (1) overly transactional interactions that reset per session, (2) responses that feel impersonal (the AI does not feel like 'you'), and (3) lack of task-tracking/reminder capabilities needed by business users.

Chapters

How is it different from ChatGPT right now? The experience of ChatGPT has, is problematic. The first reason is that it's still very, very transactional. And so every time you open a chatgpt, write messages, it asks you like three questions and you answer those questions and you close it, open new composition, it'll ask you again. The second thing is that a lot of the ChatGPT responses still not personal yet it's like you still feel the AI is not you. And the third problem with ChatGPT is because it's Chatter, it's not a task tracking app. And for business people, so like, they need someone to follow up on items, so they really need these task tracking capabilities. Someone need remind them when they need to do sentencing.