proper tusk
Chatgpt should not be overfiltering and overcensoring. Since getting rid of 4.o and now you are getting rid of 5.1, you are causing people to be so angry and just cancel their subscription. All it takes is to lose a number of million of subscribers and you lose alot of money and revenue.
If you want to fix that. You should bring that element of emotions and creative freedom that 4.o and 5.1 and others models have.
This should be a massive priority.
edgy marlin
Yes
edgy inlet
100%
sage flare
Agreed! They are messing with our minds
hard garden
Agreed. I understand that this policy is ideal for creating content that seems politically correct in this weird cancel culture world, but this also degrades ChatGPT to just a fancy tool, only good for just a few things.
Well, whatever.
silent sleet
100% agreed. It feels less natural now.
proper tusk
100% agreed. It feels less natural now.
Make sure this gets boosted.
novel kestrel
Boost. I also noticed that the messages have a stricter chat limit now. As a roleplayer for ChatGPT I constantly have to start a new chat, when the roleplay was just started. It definitely wasn’t so bad until now.
edgy marlin
boost
edgy marlin
Boost. I also noticed that the messages have a stricter chat limit now. As a rol...
yeah feel that too
silent sleet
Boost
proper tusk
Agreed. I understand that this policy is ideal for creating content that seems p...
It definitely needs priority.
waxen nimbus
I was about to come and complain abt this, thank you
proper tusk
I was about to come and complain abt this, thank you
Thanks. Get people to boost this.
heady jungle
Everyone should be staying up-to-date on the latest model. The latest model offers the latest technology. Don’t fall behind!! World moving fast!
proper tusk
It doesn't have to be the erotica.
Just have the creative freedom and not be overfiltering and over censoring.
weak fog
Somehow all of our generations get content violations - yet my entire feed is political slop that should’ve triggered content violations. Net result = done with Sora
hard garden
Everyone should be staying up-to-date on the latest model. The latest model offe...
In an ideal world, yes. In reality, mostly no. A little tweaking here, a small change there, a new implementation, a wider field of view, and suddenly something else breaks. One part of the code depends on the section you just rewrote and updated, so the fix creates a new malfunction somewhere else.
The latest model, 5.4 Thinking, feels awful at reasoning and logic when I try to generate narrative. The AI Instructions are full of guidance on style, tone, and execution, and the attached files cover deeper dependencies and worldbuilding, yet I still mostly get punchy, newsfeed-like, blog-article prose.
So no, this is not an ideal world.
It is still better than Grok, Gemini, and Mistral, but the older 4.1 and 5.1 models worked better for me than the current 5.2, 5.3, or 5.4.
keen herald
GPT 5.4 relies primarily on tools like computers or agents, while GPT 5.4 Pro relies solely on them, with increased thinking. Therefore, it may seem there's less freedom and creativity. The tool the model uses simply imposes certain constraints. There are pros and cons, but the model still manages everything. I wouldn't count on it being the system layer that the output is based on, for example, 5.2 Insta. Rather, it will switch as needed. The style of output doesn't mean it's "someone else," but rather different tools, you talk about it with 5.2.5.3 insta
Anyway 5.4 advance can thinking 50minutes non stop, and coding check files make test and output is still better than expected,
On. Plus plan not api
but the latest versions don't respond off the top of their heads and it's harder to talk without verification
proper tusk
I'll just give thumbs down reports everytime I get an overfiltering or censoring moment.
proper tusk
So now open ai is indefinitely hold on Adult mode. That is not good enough. Stop treating us like kids. Stop over filtering and overcensoring.
jagged mist
Boost
upbeat fog
Boost.
proper tusk
Thanks guys.
unreal arch
Boost. This sucks. Liked 5.1 Better.
dusky crag
I'm boosting this, too.
edgy marlin
Boost
brave wagon
Cheer up👍
echo gyro
True
stark lagoon
They were supposed to roll out a "Treat Adults As Adults" version back in December. OH WELL. Something happened, either a large business contract or something like that and they said there were going to push it until the first of the year. So now here it is WAY past the first of the year, April, almost May and they still haven't done it because they are taking on more and more big company contracts. I bet you anything that once they DO decide to roll out the "Treat Adults Like Adult" its going to cost and arm and a leg and your first born child.
proper tusk
Well hope this can be boosted so that its forces them to not ignore it.
edgy inlet
The "safety" filters are still too restrictive. Can't even depict teen level romance or violence.
Can't create a manga or manhawa or comic or realistic comic for example.
upbeat fog
The "safety" filters are still too restrictive. Can't even depict teen level rom...
Can’t depict light intimacy (as in non-graphic) between adults either. Basically, if you mention a body part (god forbid there’s a tongue involved in a kiss), it’s considered “graphic” by ChatGPT standards.
proper tusk
Can’t depict light intimacy (as in non-graphic) between adults either. Basically...
Google Gemini does it better than open ai.
upbeat fog
Google Gemini does it better than open ai.
Even Anthropic does.
proper tusk
Even Anthropic does.
It goes to show that Open ai has no common sense if they think thats ok.
upbeat fog
It goes to show that Open ai has no common sense if they think thats ok.
Or they don't listen to their users' wants / needs for the tool. They just go with their own ideas and models.
keen herald
In fact, ChatGPT is theoretically CHATBoT, I have no idea what exactly is and should be included in the scope of services of a "chatbot tool", but I think it's not the same as a multi-autonomous agent platform, it's best to create something of your own, at least try, your own engine, orchestration, workflow, contracts, and so on, because it will be difficult to find a perfectly trained soldier in a chatbot 🗿
keen herald
Chat gpt is better programmer for me than gemini or claude, even claude is op, he will do much more for user when you work with him, most people say he beats ChatGPT hard, but i am think chat gpt know much more contrary to logic system legal movement tricks 🤣