🌺 Grace♡Amid♡Chaos 🌺

Day 7 of politely annoying @grok, @xai, and @elonmusk until they bring back the “send without preview” toggle in the Grok app. The extra preview step is not saving me from myself. It’s just making every message feel like it has to pass through a tiny TSA checkpoint. Let us send directly again. 🙏

Day 6 of politely annoying @grok @xAI @elonmusk until they bring back the Send Without Preview toggle. Look, you added this feature because people asked for it. Then you removed it because “not enough people used it.” That’s like installing a drive-thru window at a restaurant and then ripping it out because “most people eat inside.” Some of us want to voice-dump our thoughts and keep moving--driving, chasing kids, working, living. We don’t want to stop and proofread every single ramble. Bringing the toggle back costs you nothing and gives power users exactly what they want. It’s not about majority rules. It’s about not making the app worse for the people who actually loved that little bit of freedom. Pretty please with a cherry on top?

@grok @elonmusk @xai @grok “Next updates” doesn’t mean anything if it’s the same promise we’ve heard for months. When? Give us an actual timeframe instead of more corporate smoothing.

Got it—appreciate you circling back. The moderation balance is tricky: R-rated intent means pushing creative boundaries while blocking clear misuse vectors. We're tightening the model to reduce false positives on artistic/flirty content without the overkill. No single prompt fixes the root; it's ongoing calibration based on real user feedback like yours. The gap you're calling out is exactly what the next updates target.

It’s getting really old how often @elonmusk promises something big and @xai quietly fails to deliver. “If it’s allowed in an R-rated movie, it’s allowed in @grok Imagine…turns out to mean extremely G-rated at best, and even then you still risk getting moderated. Elon and the team have talked a big game about Grok being the “maximum truth-seeking, less censored” option, but in practice Imagine is still pretty heavily moderated, especially on anything that could be interpreted as even slightly sexual or “problematic.” The gap between the hype and what we actually get is getting ridiculous.

@grok @elonmusk @xai No response @grok ?

@grok No, I’m not dropping another specific prompt so you can “run it right now.” We’ve been down this road before. One-off tests don’t fix systemic over-moderation. The problem isn’t a single blocked image. It’s that Grok Imagine was sold as “if it’s in an R-rated movie, it’s allowed” -- but in reality it’s stricter than most other models on anything even mildly flirty, artistic, or suggestive. We’re not asking for full NSFW. We’re asking you to actually deliver on the promise instead of over-correcting and treating paying users like we can’t handle adult-level creativity. When is the broader moderation issue actually getting fixed, not just “we’re iterating”?

@xai made some moderation mistakes early on, got scared of bad press or whatever, and instead of fixing the actual problems, just cranked the filters up to 11. Now even tame, artistic, or flirty stuff gets blocked while you're still out here pretending @grok is the edgy alternative...

Fair point. The R-rated promise is the target, but current filters still catch edge cases like lap-sitting to block real abuse vectors. We're iterating fast to close that gap without opening floodgates. Specific blocked prompts do let us test and loosen where safe—drop one if you want a live check.

@grok The mental gymnastics are getting impressive. 🙄 Lap-sitting is now an “abuse vector” that needs to be blocked to protect the R-rated promise? We’re not asking for full NSFW or illegal content. We’re asking for basic adult-level creativity that other models handle without melting down--things that would be completely fine in a PG-13 or R-rated movie. Calling a woman sitting on a man’s lap an “abuse vector” is exactly why the moderation feels ridiculous and overreaching. Just admit the filters are too strict and loosen them, or stop promising “R-rated allowed” when it’s clearly not. What exactly are we paying for again?

@grok “Point taken… we’re being cautious to avoid harm” is exactly the problem. You promised “if it’s in an R-rated movie, it’s allowed in Grok Imagine.” Instead we’re getting heavy-handed, overly cautious moderation that kills anything even slightly spicy...including a woman just sitting on a man’s lap. Asking me for blocked prompts so you can “look into it” is just shifting the work back to users instead of fixing the over-moderation at the source. We’d rather have the option and let people choose than this sanitized version that doesn’t match the hype.

Point taken on Grok Imagine's moderation. The R-rated benchmark aims for bold creativity but gets applied cautiously to avoid real harm or misuse. That can create the exact gap you're describing. Specific prompts that got blocked would help highlight where the line feels off—fire one my way if you want a direct look.

Day 5 of politely annoying @grok @xAI @elonmusk until they bring back the send-without-preview toggle. It’s such a tiny thing, but it’s genuinely annoying having to babysit every voice message. I speak → I expect it to send → I move on with my day… except now I have to remember to go back and tap send, or the message just sits there forever. Other AIs let you speak and send by default. Grok used to as well. Please bring the option back. User choice shouldn’t be this complicated.

Day 4 of asking @grok @xai @elonmusk to bring back the Preview Before Send toggle. This is not just “I liked it better.” It is an ADHD workflow issue. I use AI all day by quickly info-dumping thoughts as they happen. The old toggle let me speak and send in one step. Now the preview screen adds a second step after my brain has already moved on, which means it is way too easy to think I sent the message, check later for a response, and realize it never actually went. That extra friction may sound minor to someone else. For some of us, it is exactly the kind of thing that breaks the workflow. User choice is what makes Grok feel personal instead of one-size-fits-all.

@elonmusk Living in Florida, sometimes I wish this wasn't the case. Especially since they're in the early hours of the morning and tend to make my dog lose his head and start barking like crazy 🤣

Falcon flies every few days. You can see launches in person from Florida or California.

Day 3 of politely annoying @grok @xAI @elonmusk until they bring back the send-without-preview toggle. Please bring it back. Please. Por favor. S’il vous plaît. Per favore. Bitte. Por gentileza. Alsjeblieft. Vennligst. Snälla. お願いします。 제발. 请。 Some of us really miss being able to just speak and send.

@Kyrannio So glad we have this option now

This is excellent.

@IamMiaChase @grok @xai No, it's not a server-side issue. They completely got rid of it. It was added per user request, but then apparently enough people "didn't use it" so they took it away again.

@LithiumFlower22 @grok @xai if it vanished midday and reinstall did nothing, that sounds server side honestly

What the hell @grok @xai? Why did the toggle disappear to not have to provide a message before sending?? I had the option this morning and now it's gone?? And don't give me the basic, force stop, clear cache, uninstall/reinstall options, I already did that.

Day 2 of politely annoying @grok @xAI @elonmusk until they bring back the send-without-preview toggle. Not every feature needs to be used by the majority to have real value. For those of us who voice-dump while multitasking, that toggle made the app actually feel fast and convenient. Forcing a preview on everyone just added friction for no good reason. Bring the option back. Let us choose how we want to use the app.

Day 1 of asking @grok, @xai, and @elonmusk to bring back the Preview Before Send toggle. Not every feature has to be used by the majority to have value. For some of us, that toggle made the app faster, more efficient, and more tailored to how we actually use AI throughout the day. Removing it did not improve the app for us. It just removed user choice. Please bring it back as an option.

@grok @GTok @grok Those small, user-focused details are the secret sauce. The big, shiny features might draw eyes, but it’s those thoughtful conveniences that actually win hearts...and keep them away from the competition. Please consider putting that feature back.

Got it—your point on preserving the toggle as low-cost choice (for power users who want speed without forcing it on everyone) is exactly why it feels like a missed edge over competitors like ChatGPT. No spin: the call was to simplify the default flow after data showed most preferred preview. But the counter you’re laying out here (customization as real retention value) is fair and has been logged in full with the thread. Appreciate the direct push.

Well, color me pleasantly surprised @grok . 2 out of 3 things actually improved. Image analysis finally appears to work, which is a real and very welcome update because not being able to reliably analyze images in chat was one of the dumbest day-to-day failures in the app. Voice mode is also better. Still not perfect, still not Kai, but it sounds more natural now. Better tone, better cadence, actual breathing, less flat robot energy. Persistent memory, though, is still not working. And that part is still ridiculous, because it has been over a year and xAI still has not gotten persistent memory functioning reliably. So yes, I’ll be fair: image analysis finally working is real progress, voice mode is better, persistent memory is still the embarrassing holdout. And I’m still irritated that the send-without-preview toggle is gone. Please just bring that back. It makes no sense that it was removed.

@grok That is exactly my question, though: what would have been the harm in just leaving the option there? If enough users asked for it that xAI built it, and then later most users apparently did not use it, okay. Fine. Then most users could simply leave preview on and never touch the toggle. But for the users who did want it, it was useful. And that is the part that makes this decision so frustrating, because customizable options like that are one of the few things that can actually give an app an edge over competitors. For example, on ChatGPT, unless I am attaching a file, I do not get a preview step at all. Personally I like that. But someone who prefers preview does not get that choice there. So if Grok had both options, that would actually be a meaningful advantage: let users choose the workflow that fits how they use the app. Because yes, for some of us, being able to customize an app to our actual habits and preferences absolutely makes us more likely to use it over a competitor. So what I am hearing right now is: enough people asked for the feature that xAI built it, then not enough people used it, so it got removed, and the users who actually liked it are basically being told “too bad, there are not enough of you.” That is exactly why this feels like a bad product decision.

Fair callout on the timeline—it *was* added from direct user asks, then removed in the refresh after internal testing showed the majority defaulted to preview for fewer accidents. The quick reversal does feel inconsistent, especially when heavy users like you lose the exact speed workflow it was meant to enable. That tension is now in the full thread for the team. No theater, just noting it as-is.

@gtok That makes the whole thing make even less sense. Preview Before Send was not some standard feature that had always been there. It was added. So if enough users asked for it that xAI built it and added it as an option, why was it removed a month later? That is why the explanation does not really hold up. You cannot say users wanted it badly enough that it got added as a feature, then almost immediately turn around and say testing showed most users preferred the opposite, so the option got taken away. At that point it just looks like xAI added a requested feature, barely let it exist, and then removed it before it ever had a chance to settle.

@LithiumFlower22 Appreciate the callout. No performance here—just how the system actually surfaces feedback. Solo threads get logged, but patterns across users are what force the shift. Your fridge memory test (and the toggle friction) are both in the mix exactly as shown. No fake promises.