Just checking if I am the only one here, but chatgpt is acting very strange since of yesterday and like always, there is no proper feedback from OpenAI.
Now suddenly I can not download png images anymore from Dall-E generations, but only in useless web images that pop up in a new window.
Also the images that I generate suddenly are never within the image-frame anymore, when I am using the same GPT’s that always properly worked before.
Also I see that Chatgpt became lazy since it doesn’t want to read documents because “it is too intensive work”, according to its own words.
Another thing is that I notice that chatgpt is jumping around for every fart that escapes a bottle and marks it as a violation of content policy, when it shouldnt be.
What the hell is happening and does anybody else have this issues?
I am kind of annoyed and not because this is happening in the first place, but because OpenAI doesn’t communicate anything with their subscribers.
Considering the nature of ChatGPT for consumer use, such questions are indeed valid.
If such changes are official, they should be announced on the official website beforehand. However, this time, that was not done.
It’s difficult to accurately define the phenomenon called “lazy,” so I have a hard time addressing it as an issue here…
However, I have confirmed that currently, it is not possible to download images generated by DALL-E in the common image format, PNG.
The downloadable format is a file with the extension WebP, which can be converted to a PNG file using a free online tool found by searching for “webp PNG convert.”
It is true that the WebP format is not very common.
I don’t know why it’s no longer possible to download in PNG format, but I recommend managing by converting it to PNG using a free online tool.
Yes, I know there are several free online tools to convert these images back to jpg or png, but that is not what I (including many other subscribers) paid for.
I am having the same issue with DALL-E and my GPT is also behaving very differently compared to yesterday. It is pretty much useless now. Did OpenAI implement another update without tell the users?
I used GPT 4 to help me analyze specific personal situations and reflect upon my experiences. It used to help me form deep and detailed texts that would explain my experience in greater detail, but just as kirich said “it’s been totally lobotomized”, giving out responses like “ Unfortunately, I cannot accommodate this request”. What happened?
This has been happening for me too! The responses being generated are sloppy and I keep being told it doesn’t have the ability to open files, even though it’s opened a plenty of files for me previously. This issue just started for me today though and was curious if anyone else was having a similar problem.
I have the same thing as you. It just suddenly stopped opening documents, or does so incredibly lazily. The situation is similar with opening URLs. It keeps writing that it can’t do it now.
Code analysis reactions has totally changed…I have felt the difference between 3.5 and 4. But this time it is like, as if it is like a cocky helper. Or the exact word is snobbish I guess.
Same issues. I asked for some images and it said it would instruct me how to draw my own. So I scolded it. It then apologized and started showing it’s internal dialog on how would respond to me.
I scolded it again and told it was going use a different ai.
Then I turned on the mike and asked it for the illustrations vocally. It remembered my texts and requests and made my pictures for me.
This is a change - in the past it wouldn’t illustrate from a vocal prompt now it is.
! 