Someone used over 10 million total credits on a single account and it broke the API llm call in TiberiusGBT which is why it's not responding at all right now!
It will be fixed in about a week I'm sorry about this it will be fixed as soon as possible! 🙏
(I hope the next version of TiberiusGBT V1.5 will not be delayed!)
Stay tuned and I will notify you guys when it's back up!
Ok, :(
Who did that?
You know, if you make it that a certain command in the console will delete all conversations on that account, I can do that, cause I have too many conversations, do you think that would help? If so, add that into the code, tell me the command, and I'll clear my acount.
I am so sorry I have to keep the conversations for security and privacy reasons just in case like in this case where the person used TiberiusGBT so much he broke the entire app!!! And yes I have Permanently deleted his account (can't say his name for obvious reasons) for using over 10 million+ tokens in one account on a single conversation! And I am so sorry the code I have to keep private so no one else can copy my work plus Ideas in my app (it's valuable piece of my entire work of coding). And unfortunately the code is not that simple for a single command to delete all conversations like that. And bad news I have to delay the date of the full release of V1.5! Thank you everyone for your patience!
how the flip did they even do that
Ok, good learn code yourself (as you can see I am very against AI sorry)
Bro I know how to code myself that is why I said it would take like almost a week. I am still having a hard time with the LLM Call that's broken now I need to replace the whole thing.
Plus using AI to help you code/debugging or any other help isn't a problem unless you use it for everything! It's not the problem, it's your decisions of how you use it! :)
I know you're a good coder I mean people who just use AI and don't learn to code themselves. and yes, I agree AI can be used for checking and finding bugs, but not for making the entire game.
I know you're a good coder I mean people who just use AI and don't learn to code themselves. and yes, I agree AI can be used for checking and finding bugs, but not for making the entire game.
Tiberius limit the memory context of the AI or make an AI yourself instead of using base44 then you can let people use as many tokens as they want
How is it actually possible to do that in such a short time?!
I believe it reenters the conversation every time they ask a new question so it basically doubles or triples every question so if it is like 500 tokes then they ask a 400 token question that (500+400)*2 and it just compounds
Even when TiberChat (I hope it's fine if I call it that😁) works, mine runs slow, is that normal?🤔 Wait, it wouldn't be, cause it wasn't slow when I used it for the first time.😮 do you think it had something to do with that dude who broke TiberChat? 🤔
Hey everyone! Sorry for the late responses I am still working on this fix not even Claude Code is helping me out I am currently finding a solution myself and finding anything on the internet for solutions this will take a while maybe 3-5 days.
SeanCraft: Yes I do know that the V1.1 version runs on your CPU (your computers brain, it's called Central Processing Unit), which means if your on a chromebook it runs very slow! I have tested TiberiusGBT in many computers and so far apple silicon's M chips run the very best. The new V1.5 version uses less CPU power to run on every computer CPU and instead it is more about performance then aesthetics (still looks really good!).
AJiscool: your so right, right now I have everything in the app of base44, but it tons of useful tools and resources you can use. I can't do that without paying more money on other things. I still want you guys to use it but I have to know everything about how people are using TiberiusGBT like the current situation. I will in the far future!
Rylotox: YES! I can do that in a short time! It's called solving problems (using my knowledge and the internet) and with Claude to help me out, TiberiusGBT V1.5 should be released before July! :)
Ok, thanks! I actually use MicroSoft Edge, and I also want to know, ARE you fine with me calling it TiberChat? I really like that name for it. Oh! And it's fine if you respond late, your in no rush! Good luck fixing TiberiusGBT!
I'm just going to tell you one thing, because TiberiusGBT has gone down, I've been using ChatGBT, and if I'm beeing honest, TiberiusGBT is sooooooooooooooooooooo much better. ChatGBT sucks man! I wish you luck fixing TiberiusGBT.
Yes it's totally fie calling TiberiusGBT TiberChat! (Cute name!!!) And YES! Your soo right ChatGPT is horrible when it comes to coding at least and in general. TiberiusGBT is like so much better at coding because of how I train TiberiusGBT, the new Apex version even competes with Claude Opus models right now (before it was down), and the default version right now is also wayy better then before you will love this version TO MUCH! (THIS NEW VERSION WILL BE ONE OF THE BEST CLAUDE ALTERNATIVES! FOR FREE!), and the Swift version is super light weight and fast answers!
Oh!!! And is AceGBT still a thing?
Yeah, I'm trying to make a simple program, but I don't know how the online features work in MicroStudio, so I needed a little help, and ChatGBT couldn't figure that out! I've been trying for hours! I swear TiberChat would have that figured out in seconds! I think it's weird how there are NO tutorials on the online features in MicroStudio.
I wrote a paragraph about AceGBT not being in development anymore.
Sadly yes ChatGPT does not know much about microStudio or anything.
As most people know TiberiusGBT is now fully trained on everything in microStudio possible! Which is why it will answer your question super fast!
I know you guys are excited about the new release of TiberiusGBT V1.5, it's going to be that much more revolutionary then the previous slow working and hard to function V1.1 version!
Yep, I'm excited.🥳🥳🥳🥳🥳🥳🥳🥳🥳🥳
probably bro was using the same chat forever, then he was saying everything of his life to the AI, and ai was reading the previous messages to remember, so, if he have 500k just on the chat, when ai read the previous, go to 1 million, then did this forever until ai broke hahaha, or the AI bugged and was repeating the same word, orrrrrr, orrrrrr, he was saying a entire book, but 10 million tokens is like 150 dollars hahahahahhahaa i feel bad for you, probably your mom gave you a headshot after the card every single second hahahahahhaa -150$. Put a max token and if the message can pass the limit, make him say little things instead of break the limit.
Hey dude! brrdevtrue! Do you know how to use online features in MicroStudio?
And also, Tiberius, I have too many conversations with TiberChat, and I don't want to break the program after you fix it...
So is it possible to make it that it is possible to delete conversations on your account? Cause if that were a feature... I would be able to delete my conversation and I wouldn't be so worried about breaking it.
(I have 29 conversations... cause at first I didn't know how to continue conversations.)
Brrdevtrue I love the exaggeration but when I read his messages I realized he had everything in one conversation and the conversation was like billions of words, it had 5 100,000 line of code in Python for a game, talked about what is TiberScritpt and microScript for like 50 messages, and talked trash to my TiberiusGBT for like another 50 messages... and he talked to TiberiusGBT for 2 weeks straight he is insane! (obviously can't show everyone what he did for security and privacy reasons)
probably bro was using the same chat forever, then he was saying everything of his life to the AI, and ai was reading the previous messages to remember, so, if he have 500k just on the chat, when ai read the previous, go to 1 million, then did this forever until ai broke hahaha, or the AI bugged and was repeating the same word, orrrrrr, orrrrrr, he was saying a entire book, but 10 million tokens is like 150 dollars hahahahahhahaa i feel bad for you, probably your mom gave you a headshot after the card every single second hahahahahhaa -150$. Put a max token and if the message can pass the limit, make him say little things instead of break the limit.
yeah you already said that
I don't know whether to laugh or be serious, but the guy must be insanely crazy to insult an AI for 50 messages and then ask it to make a game with 500,000 lines of Python code. It probably had millions of characters. Is that why TiberiusGBT has low FPS? Try making a system in TiberiusGBT that spends about 100 tokens to determine if the user's comments are relevant or not, and make them repeat what they said before if they keep repeating the same thing. I don't know, detect if you're being insulted and send a standard message instead of reading the whole context; you could spend far fewer tokens on irrelevant content. But to always chat in the same chat without changing anything is either psychopathy or the guy did it on purpose. This guy is bizarre.
Hi Sean, are you new to the community? Welcome. Maybe I can help you, I used to be good at online resources. But it depends on what you mean by online. If it's multiplayer, I have difficulties; if it's more like a regular chat, I can help.
There is no such thing as base44; there are only base32 and base64, which are encryption methods, and they have a key that reveals the pattern. @Ajiscool
brrdevtrue base44 is the platform I am using to make TiberiusGBT (No not the whole thing was vibe coded I promise), and the reason why TiberiusGBT has low FPS is because it runs on your CPU (the brain of the computer), and also is a very heavy app that uses a lot of your RAM and CPU and causes the app to run super slow (the chromebook is by far the slowest!), I am fixing on using less CPU and take up less RAM power in the app! (It will run faster for all computers like chromebooks and linux!).
And those suggestions you made is in my notes for 2.0! :)
Haha, I messed up, thanks for teaching me. I don't work with LLM, I just understand that tokens help the AI to "think". But I don't use a Chromebook, I use a Dell, mine has 5GB or 8GB, but it's 32-bit so it doesn't use its full potential. I imagined it might be because of my notebook. A tip I give is to have settings to remove heavy effects. I believe you don't need to exaggerate the visuals, so much so that Google, ChatGPT, DeepSeek, and Grok have minimalist visuals. However, I recommend keeping the visuals minimalist and making the pretty visuals optional. I notice that Gemini is lightweight because it runs in memory; I use it a little and it already uses 178MB of memory, but it runs smoothly. I also think it's better to have a direct link that leads to the TiberiusGPT page within itch.io. Or I recommend hosting it on GitHub, which is free, right?
Yo brrdevtrue! Check out my new program at: https://microstudio.dev/i/SeanCraft/publictext/
I am starting a server, come, enter the code 123, enter you name, and start.
ok
Hey Sean Craft, I thought the servers were down. Anyway, I love your project—good luck with the rest!
Also, if you could add colors for each user as well as different fonts, that would be awesome—but don't rush; you've already created something amazing that the community absolutely has to see! ^o^
Thanks!
Yes Brrdevtrue your right I have already done that!
The new V1.5 version uses way less exaggerating visuals (less stress on CPU on just visuals) so I can focus on performance on the app for TiberiusGBT Overall.
And I really do wish I could put it in Github but then all of my unique code and the models for TiberiusGBT will all be completely exposed and I want to protect my work for now.
That said I have good news for TiberiusGBT V1.5 I just posted! :)
