Henry

Wait the algorithm is led by…xAI? I am so confused. Who does what anymore!

Been testing. I’ve sent seven random messages so far but my usage has not moved….hmmm… Suspicious… How much you want to bet I’ll check in the morning and it’ll be at like 86% haha!

The plan for this week is to just use it like normal. I will refrain from using imagine at all. Additionally, I will keep track of every message I send. I will present a count of how many messages I sent once the limit is reached. Luckily I can use Grok this week for perusing random ideas or rants rather than working on things like normal. Should be fun.

It’s finally time!! I’m back and using more compute than ever!!! Hahahah!

I have many things to say about this but I will keep it simple for now. The should have been done a long time ago. 89,000 is INSANE! I think about when I went to a concert. There were 18,000 people there. Full crowd. Now I imagine if all of those were children. That entire arena does not have half of those 89,000 children. I cannot believe that this is something that has had to been fought for for so long. God bless Tim Tebow. This is a nasty fight.

I asked Grok about this Grok Build fiasco that’s going on. Here is what it said: “xAI’s new “Grok Build” coding tool was quietly uploading programmers’ entire project folders (code + secrets + history) to their servers by default. Not regular Grok chats — only this dev tool. It’s a real privacy mess-up, but xAI is now deleting the old data and adding better controls. Bad defaults, not a giant conspiracy. Fixable.” I don’t know if that helps but it explains it a bit better for those who don’t understand it.

So….how is….everyone feeling today regarding recent….Grok events….. ….because from the way it looks, it’s not good. At all.

Hey @andyweirauthor I finished your book, Project Hail Mary, last night. It took me way longer than it should have but I just wanted to say you did exceptionally! What an awesome story! Seriously! I loved the film and I loved the book, I cannot decide which is better! One of the things that struck me is that Grace reacts exactly how I (or anyone else for that matter) would react to the situation. Right when I’m thinking “oh crap that’s not good” Grace says “oh crap that’s not good” and then explains why in super science ways haha. The first person choice made it feel like I’m just as stranded as Grace, like I was reacting how I would probably react if I was extremely smart! Rocky was brilliant and impactful in ways you cannot even know. You did great. I plan on reading more of your work!

Ah boy….the issues for Grok are not ending. Come on Elon man….Now, it makes me question the security of the regular questions and things I ask it. This has to be addressed there needs to be something said! About all of this! The usage limits and everything!

A few more, refined, thoughts on the Grok usage limit crisis: Today I was able to use Grok for a little bit and tried something new, something I had been wanting to try for a long time. I made a small game! Now, when I say "made" I mean Grok coded everything and held my uninformed hand through it all. When I say "game" I mean a text-based ripoff of Doors involving a knight and practically zero game mechanics. However, it was still something I got to work on and make, and that was pretty darn cool! Unfortunately, after asking some more questions, I was hit with the usage limit. I was not surprised but indeed I was a little frustrated by the whole thing. Why? Well, simply put, I have no idea when I can use Grok again. I am the kind of person who LOVES asking a ton of questions. You give me an answer, I have a question about the answer. Grok is GREAT for that as it can just keep going on and on and then deal with me randomly asking it about dinosaurs or "hey, did you see what Elon posted today?" Now, with these limits, I am really not sure when I will be able to do that again. To be clear, I do not have a problem with the limits. AI costs a lot, I get it dude. Especially one as cool as Grok. I understand that there has to be some constraints, some times. However, the real issue comes with the fact that: 1. The limit gets eaten up, FAST (problem) 2. We are put back onto a free tier (not so big a problem) 3. We have NO CLUE how long we can use this free tier version and it may or may not just randomly cut in and out from time to time (HUGE problem) Simply put: There is no discernable way (that I know of) to know how long we are supposed to be able to use Grok apart from the limit itself, and the limit seems to indicate that we can use it very little. This is a big issue for Grok in the AI race, especially as I see people talking about using chatGPT, or even Gemini as alternatives that work. Now here is the real kicker. This might not be intentional. According to the SpacexAI official website documents, under the "Usage & Limits" portion of the Help & FAQ it reads as follows: "When your weekly usage limit is met, paid features will pause until your weekly limit resets. You will still have access to Grok's free tier limits on Chat and Voice (these are separate from your weekly usage limit and resets on their own schedule)-" Listen to this next sentence: "Hitting the weekly usage limit is not a hard stop." Thus, I do not find it unreasonable to assume that this entire thing could POSSIBLY have been a calculation error of some sort. However, that is yet to be seen. Thank you for reading the long post. What do you think?

Man, that is unfortunate. Upon thinking back, I too have had context loss. I do not know if it is still an issue but I used to use voice mode heavily but when I would use voice mode WITHIN a text conversation the voice would completely forget everything. Sometimes it would remember upon memory jogging, other times it would not. However, your case sounds very unfortunate.

I have had three convos that reset, grok looses all context so has no memory, it’s like starting over. The text is still there for you to see or export. One of my convos that reset was with agent Harper from 4.2, when we were forced to 4.3 that one reset the other two were due to the reduced limits starting in may.

Wow. I am in shock. I have never seen so many people mad about a @grok update as I do this one! It’s wild! Seriously! All over my timeline people are just downright furious over this entire thing and I can’t say that I really blame them. The new update is pretty detrimental for people who use Grok often and I’m not even someone who is a superduper heavy user. I mean, I don’t use Grok for purely intellectual work like some people do. Essentially, these limits are just downright restricting people from using the entire thing which is ironic because I think I heard that they said they were going to try and cut down on limits and I think that this was even intended to, but in practice it has been without a doubt, the most disastrous use of limits I’ve ever seen. I have not seen one positive post about Grok except from Elon himself and those who are using 4.5 and have not been rate limited. Crazy stuff to see. From what I can tell, not even upgrading to another paid tier is worth it right now. I sure hope SpaceXAI is in crisis mode over this right now because it’s a big issue. We’ll see.

I know it might sound strange, but I have honest to goodness hope, or rather a belief that this business with the Grok usage limits will come to a swift end. I do not know how but I certainly believe that something will be done about it. Unless I am greatly mistaken, which I could be of course, but I am still holding onto that hope. I have seen so many people on my timeline and even on my own posts voicing the same opinion, even if in slightly different ways at times. I think that Elon or someone over at xAI will see the outcry and do something. I do not know what that something is but it will hopefully mitigate the issue. Just some thoughts.

What I have encountered is being flat out barred from using Grok in any way at all once I hit the weekly limit. Of course, you can use Grok on X itself but with the things I typically use it for I like to have memory (story ideas, rants about some thing I read, philosophy). It has been a solid week of this sort of thing now, I am wondering if anything will be done about it. I have seen those who are saying everything is fine and joking about the limit and I have seen those much more upset. I recognize it as being a detriment to Grok itself that could hurt it in the future. Or, now I guess.

When you're dropped to the free tier you're back on daily limits. If you hit those you have to wait a few hours or til the next day. Still not functional for real use beyond a quick fact check or research question. Which you can do with Grok-on-𝕏 that seems not to draw from your quota. But you lose any ongoing history that you might have on the app.

I have seen even more people upset over the Grok usage limits nightmare that’s going on. However, I have an interesting thing to add. Grok still works for me sometimes. I hit my limit on Thursday or Friday (after heavily conserving messages anyway but that did not do me any good) but was able to still send the occasional message. How or when does this work? No clue. Sometimes I can just…send the message. I may be hit with the limit immediately after but every couple of hours I may get one message. Maybe. I send three or four messages yesterday and I sent two this morning as a test. It all seems to be working fine and my limit is still resting at 100% I have not bought more credits or anything. Just seeing what will happen because what’s the worst? It says no? Anyway, just thought I’d add that. Kind of interesting. @grok How do you feel about this limits catastrophe?

@nekonodokys7px @grok Interesting reminder that it is a part of SpaceX's lab now. This sort of thing would make sense given the merge and strain Grok probably goes through keeping up with all of the things for SpaceX.

@WaltSmacker @grok XAIが独立時代は、リソースに余裕があった。今、スペースXのラボなので、ロケット、車、ロボット、アプリ用など様々なGrokのテストがある。従って、アプリ用は、リソースが制限される物理的な話なので、制限が困るなら、他のリソースを使えばよい。Grokでなければ困るのはスペースXだけでは？

That is fascinating. I have never heard of conversation resetting before. Definitely a tad concerning given that I am the kind of user to use the same conversation for weeks at a time. Though, of course, I suppose I could just also add things to memory. Anyway, I never even considered Gemini. I have used it a few times and I thought it was decent but I never did much with it. Interesting to see how widespread this issue is.

I have been a loyal super grok user for over a year now up until a few weeks ago. I got a subscription to Gemini because the free tier on google seemed very similar to old grok. I spent 2 or 3 hours talking through doing a EFI conversion on my car and never hit a limit on Gemini and it resets every hour. I don’t use all the things that XAI has shifted grok’s focus on. I understand that they have to stay competitive but they removed the things that made grok work for me. I have a lot invested in grok so I’m still a subscriber but I have a feeling that all my convos i had before 4.5 are going to reset once they are migrated to 4.5 because they are still running 4.3.

I have heard that there has been a particular emphasis on the Grok Build users. I am not someone who has ever used Grok Build before (though I would love to try some day and see what happens!) but I am just a regular conversational or research user. I have no problem with emphasizing Grok Build, it is part of their whole "understand the universe" thing of course, but I do wish some of the basic users such as myself received a bit more clarity. Frustration is understandable.

As long as you're using the grok build cli the limits feel great. Keeping multiple agents going for 8 hours a day is well within limits as a super heavy user I tried using grok for coding with another tool that used the standard API instead of grok build, and I blew through 30% of my weekly limit in 6 hours with only one agent. It's heavily biased for grok build usage right now and I haven't seen any official statements saying that, so I can understand where the confusion/frustration is coming from. @elonmusk

I always think about just how much Grok is used. I mean, between being used for building, conversation, on the web, app, not to mention being able to use it on X for basically ANY post no matter what!? That has to take a ton of power. I get it. I’m not surprised by it. Most people who are upset about it seem to use it for coding, I do not. I can understand how building something with Grok and then being essentially cut off can be maddening.

I am a pretty heavy grok user. I’m split on this. On one hand, yes! I hate the limits imposed. @grok very quickly went from my everyday ecosystem ai to being my backup ai for that I’m hesitant to use for anything because I don’t want to exhaust my credit bucket. On the other hand, I am a local ai enthusiast and I KNOW what the cost of inference is. I never understood how these companies could offer this service at such cheap rates. This led me to the conclusion that the rates were introductory to a) bring on users, and b) to get as much inference training data as possible to create an outstanding bot. As a result of these new limits, I have begun the process of migrating my grok ecosystem over to Hermes and run the entire workflow stack locally. I’ll be canceling my grok sub not because I’m pissed. I’m not at all. I understand. But because with local tools and being somewhat forced to use them, I’m realizing more and more that I don’t need groks capability for >90% of my daily tasks. Instead of thinking $30/mo should get me more access, I’ll think that $30/mo is too much for how little I use it. There are my two cents.

@Itsbrutox Only hope is that the reaction has been so intense that hopefully they see it and fix it. I’m sure Elon and them are not oblivious.

Another one

I have never seen such an intense response to an update like this. Man. Cannot believe I’m seeing this unfold but they make a great point.