Hi, I suggest that AI-Studio introduce a subscription plan (for example, a basic plan for 8 euros) and allow NSFW content, at least for paying users. You can keep NSFW content to a minimum, but at the same time you won’t be banning it, you’ll have age verification
This is my first time posting on this forum.
@Logan_Kilpatrick Good day, and we are still here
Really? But we can create photos and videos… Besides, isn’t developer mode specifically for developing? So it’s basically for creative purposes?I develop creative stories, entire books. And this update is completely stupid… I had to adapt everything and it’s still a nightmare.Don’t think in a childish way, you have to think big! I was going to get an API key so I could write without any problems. Now that’s out of the question. I won’t do that while this filter is in place. Creativity doesn’t stop at your private world, it extends beyond it, and that’s what makes us human. Gemini is incredible and so creative! I hope you’ll go back soon… please… think about these amazing possibilities…
@Logan_Kilpatrick a daily reminder, again
@Logan_Kilpatrick Boo. How much longer do we have to wait?
@Logan_Kilpatrick Another daily reminder so we can hopefully receive info about this.
@Logan_Kilpatrick Daily reminder: It’s the 24th day since this post was published.
@Logan_Kilpatrick a daily reminder as always.
Also, am I the only one who had an increasing rate of errors for the last few days?
@Logan_Kilpatrick daily reminder. We have waited, we are waiting, and we will wait.
@Logan_Kilpatrick And again, a daily reminder.
@Logan_Kilpatrick I would like an update on the situation. The external filters should be removed and we should only rely on the AI’s training data. This is costing you money for requests that are being processed but aren’t useful.
Another reminder @Logan_Kilpatrick
I see the economics of such a move otherwise. The user is still charged for a blocked request, but it’s takes less compute to process since it’s blocked from continuing. Sometimes I think that it is the reason for censorship changes.
And yeah, I agree - the external filter should be removed, or at least the ability to remove it should be introduced. It is self-evident that if I set safety settings to “none”, then I don’t want safety checks.
But in the case of the previous filter, which terminated the request rather than deleting it, it worked exactly the same. Computing power is consumed even more, because once, because you have to regenerate the same query over and over again, instead of being able to edit it and continue from the place that wasn’t censored.
It seems to me that nothing has changed much. The site is only concerned with displaying the generated content. It is the gemini servers that generate the content. In fact, nothing has changed from the previous filters. Both this filter and the previous one interrupt the generation process so that costs are unlikely to decrease. I even doubt that the filter is based on a full-fledged AI. From my experiments and observations, it looks like there is just a trigger system that blocks the message when it comes across the word trigger.
In simple terms. A meaningless, in some way filter that can be bypassed. Or in human language, a crutch.
#Logan Kilpatrick If you’re reading this, then I can partially understand your position. However, I’ll be honest. The implementation of this system is extremely inadequate. As a community, we can accept censorship. We didn’t live badly with the previous version either. However, the current implementation is inadequate.
no. I mean im pretty raunchy, but even i know the difference between [The n Word] and [“The”, “n”, Word”]
just dont. even people not in a company can destroy you for it…
An interesting point, but there’s another question: why there’s no such features implemented in Gemini app then? The Enterprise version has the output blocking though, but AFAIK it’s managed by Workspace admin.
And most likely if it was the cause, then Logan or other Google spokesman would have said it straightforward. So… never attribute to malice that which is adequately explained by stupidity. I still think that it’s a code issue because AI Studio has some serious bugs for the last few weeks. And if it’s not, then… well, Google, those who have intentions for malicious use are bypassing your filters. You’re just striking the normal users.
There have already been updates that changed the interface, there have already been tons of complaints, 100% of which have been read, I doubt this was done not intentionally. It’s just some strange decision that for some reason seems right to them 
@Logan_Kilpatrick
Just chiming in again to keep this issue visible, as it’s still completely breaking my workflow. The way the output wipes out entirely when a filter hits in the middle of a generation is incredibly frustrating. We’re losing all our work and wasting precious quota just fighting these triggers that aren’t even clearly defined. It’s a real shame to see such a top-tier tool suffer from these opaque and counterproductive updates. Please, could you at least add a “save partial output” feature so our work doesn’t vanish into thin air? The community really needs this fixed.
This isn’t just about creative writing, although I think that’s important too. I use the tool for work to analyze large volumes of text, like articles and subtitles. Right now, the new filters make this almost impossible. The generation often cuts out mid-way and wipes the entire response, so I’m left with nothing after waiting for a long output. It’s making the service really unreliable for actual productivity