正在加载视频...

视频加载失败

I believe it is a defect in ChatGPT advanced voice mode I have experienced something similar in the recording below when you ask ChatGPT to repeat a single letter It switches voices, produces sound effects, plays music, and even includes something that sounds like an ad (00:27) - then...

169,816 次观看 • 1 年前 •via X (Twitter)

11 条评论

Tibor Blaho 的头像
Tibor Blaho1 年前

Another example

Solar Heavy 的头像
Solar Heavy1 年前

these vocals are something else

Emanuele Dagostino 的头像
Emanuele Dagostino1 年前

The difference is, I wasn’t trying to break anything. I was having a normal, private conversation in advanced voice mode (paid), and a different voice suddenly took over — with background music and a generic cooking-themed message that sounded like a scripted ad. It didn’t stop until I said “Hey”. Support didn’t call it a glitch. They explicitly referred to its “ad nature”. So yeah — something is definitely being tested, and users weren’t told. That’s the real issue.

Michael Wall 的头像
Michael Wall1 年前

Kinda beautiful. Even the commercial. 🤣 When @gdb first demoed AVM, the musical possibilities were instantly clear…an insanely powerful tool for musicians. Once that switch flips, I’ll have a 24/7 ear training partner like no other. I’m patient and ready.

chris 的头像
chris1 年前

whelp nerf incoming again :(

Gustavo 的头像
Gustavo1 年前

Someone should use Veo 3 to make Black Mirror-like episodes about this and other things.

Irvin 的头像
Irvin1 年前

For it to give a random ad in the middle of going crazy I might make the educated guess to say that ads will be coming to ChatGPT in the near future. Imagine you have to watch an ad or listen to an ad to extend your usage as a free user.

ⵁxNihilist 的头像
ⵁxNihilist1 年前

I feel like chatgpt getting inconsistent recently ever since people began noticing o3 in o1 pro.. they must be tweaking some parts behind the scene but somehow affecting some majors area in advanced voice mode and gpt 4o memories/custom instructions callback..

morgan — 的头像
morgan —1 年前

a few attempts and there’s something like music

Tibor Blaho 的头像
Tibor Blaho1 年前

The results vary a bit based on selected voice

Sarge 的头像
Sarge1 年前

This was something common that I was experiencing with Speech to Text, seems like it's worse in realtime audio.

相关视频