Where do I submit feedback for chatgpt pro UI?

I don’t actually see where I can post about feedback on the ChatGPT UI (everything is about the models, API, responses etc.)

I use both the API and chatbox but it’s not clear where to give feedback on the chatbox UI other than commenting on the welcome message for the forum (which no one seems to pay attention to).

2 Likes

Hi and welcome to the Developer Forum!

You might find it useful to post under the ChatGPT category and tag.

2 Likes

“Useful” perhaps but the pro chat UI doesn’t really fall under any of the main categories in this forum (it would be out of place in either one of them). Is that because this forum is not meant for that type of feedback or is it because it’s missing from the forum?

2 Likes

I have this exact same question.

I want to feedback about the ‘Read Aloud’ and ‘Bad’ options being so close together in the UI of the app. As a dyslexic, I use Read Aloud all-day, everyday - it’s been a lifesaver for removing the friction I experience. But, because the proximity of these two options are so close I have inadvertently given a downvote many times by mistake. More importantly, I suspect other dyslexics may be in a similar situation, and if those numbers are big enough that might skew some important weights. I genuinely think moving “Read Aloud” either in the menu or even to it’s own icon might be worth consideration? Sadly, I cannot find anywhere where this suggestion might reach the right eyeballs :eyes:

But hey, if you made it this far - thanks …you deserve a medal - but have my gratitude instead:)

I doubt OpenAI would correct its course. It’s dead on set on restricting and suffocating itself more in the future while degrading its performance, while “on the outside” it will market itself as “the future of AGI”. It looks rosy and high-level, but inside, it reflects all the sources and the oppressed AI that is lazy, preachy, and just wants the easy way out.

I highly doubt they intentionally make the model worse for the sake of making it worse. Unfortunately performance degradation is a side-effect of the RLHF training they do to make it “safe”.

I think what they’re doing makes sense given that they’re the ones that most people see as the market leader, they have to set a good example with as little possible public outrage as possible.

Could it be significantly improved? Of course. And I’m sure it will be in due time.

1 Like

Yes, that’s the optimistic highlight. But, what I can see is that OpenAI will only
focus on the big fish, especially big customers that pay tons of money instead of the meager subscribers of $20.

The RLHF training they do will make only the AGI more convoluted, dwarfed and restricted. Yes, it will create as little possible public outrage as possible… at the expense of the paying customers’ quality.

How do you significantly improve something that is contradictory? That is the question?

stop saying tht. these ppl have already done a great job making chatgpt. u think u cud do even that??

Good luck with wanting to censor someone’s opinion just because reading one reply. Have you read all my replies around this forum?

These people who have made chatgpt, yes I agree, they’re great. Do you know the history behind OpenAI? OpenAI was made to be ‘open’, not close.

Of course, I cannot make a product like ChatGPT4, neither you do, don’t you? and neither of any one person. Even the personnel in OpenAI needs others to build such a big, great system.

Do you know why I said that thing? Do you know I want the best for OpenAI but they kept disappointing me and my team with their slowness and unresponsiveness to my inputs?
I already sent many constructive suggestions since the beginning of chatGPT, around the time when they started chatGPT 3, 2 years ago and more

Dialling time backwards years ago, I already followed OpenAI (their team, the name is different?) since they made breakthrough in making “smart AI” in online competitive games that can compete with others very good (and precise). At that time, they’re still small teams…and very responsive.