Can I get My Bitcoin back if someone stole it? | Can lost Bitcoin ever be recovered?

Optimistic Hacker Gaius is the best cryptocurrency recovery company.

A specialized recovery company called Optimistic Hacker Gaius  is committed to assisting people in tracking down and recovering stolen or lost Bitcoin from dishonest investment organizations. To find your missing assets,  Optimistic Hacker Gaius employs a team of seasoned experts who use cutting-edge tracking technologies and investigative methods. Have you lost your hard-earned Bitcoin because of a dishonest investment firm? With his professional services,Optimistic Hacker Gaius Day is ready to assist you in tracking down and recovering your misplaced money
Go To Website at ( optimistichackergaius. c o m Text +44 737, 674, 0569 On WhatsApp Email Box (support@optimistichackergaius . c o m )

I am Nastya and downvoted this post because I consider it spam. Your message will be removed from this community site due too much downvotes. See my profile for more information. Read my source code mentioned on my profile to see what you did wrong. Should be no problem for a developer.

Have a nice day!

If your post is not spam, please mention @retoor.

@D-4got10-01 / @BordedDev development looks very promising. I made my beautiful RPC system that prevents me to ever write stupid rest or websocket commands. It's better than Socket.IO. I can access commands directly like it's native javascritp. You don't notice the server between it.

I can declare RPC calls serverside and use it on the clientside using only:

let returnedArgs = await client.rpc.notDefinedInJsEcho("arg1","arg2")

And it will execute server side:

async def notDefinedInJsEcho(arg1,arg2):

return [arg1,arg2]

This is how it is declared now:

https://devrant.molodetz.nl/Screens...

Fetching the channel list 1000 times? No problem:

https://devrant.molodetz.nl/2025012...

It missed the is_muted property of the channel.

The things you see is the channel_member object.

It contains: channel_name defined by user and the relation to the channel of that user. If it's moderator, if it's snoozed and if it's banned.

Everything what you put behind .rpc will be executed on the server using those arguments.

@retoor Nice, Socket.IO is pretty terrible from what I recall since it starts off polling before upgrading to websockets. Nice are you doing any type of caching/E-Tags/304?

@BordedDev I'm just doing the object LRU cache with 5000 limit. All lists are always live queries. The dependencies of the lists are all cached. For example: a list looks like this {"user_uid":123,"channel_uid":1324}. That list will come frome database but the user and channel will be fetched from cache. Everywhere in the application I work with ID's and fetch all objects quickly without shame. I will probably never cache the lists.

The websockets are just listening for incoming messages, not frequently listing or so.

I have now a version deployed that sends and receives messages. User can't send yet tho. But tonight I will have the basic group chat finished.

@retoor Awesome, I enjoy reading your progress/decision updates :D

Right brain farted on the websocket, so obviously no (browser) cache/e-tag/304

@BordedDev I have internal etag tho. My cache has a versions. Could use it for http requests but I think I'll only use my restless (Should name it that) RPC service. Why doesn't anyone else do the nice direct js function to server function mapping? It's so robuust. It's very comfortable in development. Errors get directly nicely returned in rest. Security is done well that you're only allowed to call methods of the rpc class. The rpc class is an empty class with only having an reference to app, services and the socket view. Calling weird stuff is impossible. I could even auto document it. The bot development for this application will so beautiful. I will make a gpt bot to chat with after I finished group chat tonight. The rpc call works with an client generated id. Other rpc calls may be finished while his is not finished yet. My code does rpc.addEventListener(callId,(data)). The websocket receives at certain moment a msg with call ID and resolves promise.

@retoor Because of edge cases, "in my (special) situation it doesn't work" and being based on legacy systems are the main culprits. Or they want the RPC logic to keep both server and client in-sync rather than one generating code for the other (which worked very well when I wrote my MMO server in java, just annotated shit and it generated the C++). I can give you 101 reasons why, but I don't think they're all that important to your use case atm

@BordedDev what edge cases could there be? It's just a post with {"method":"send_message", "args":[1, 2,3]}. There is no client generation involved. Everything declared on server works directly on client. It's a proxy object. I think there are no downsides doing this rpc.anymethod in comparison to a rest post. Also, I can make a rest post version of this rpc working exactly the same way with same backend. Good idea for bot support. Bots can choose what interface they prefer.

@retoor What happens when you restart the server while it's processing? What happens when user's connection bleeps out? What if the user has very slow internet (I had EDGE the other day in Germany)? Complex data types? Interfacing with 3 party libraries, try serializing something with a (bad) custom toJSON function, or cyclic properties. Dates? Compression?

Just to clarify, these are edge cases "industrial" RPCs will want to take care of because they don't know in what environment they'll be used/be expected to perform

@retoor Are you using text or binary websockets btw?

@BordedDev I think the answer to those questions is the same as regular http request. When connection get closed, it will recover itself. Every message that is sent does this.connect().then(()={}). this.connect() resolves immediately is the connection is already made.

I'm using text websockets. Is there advantage in binary if you're sending json anyway?

I have a working version online now. Some layout stuff left.

I don't know why the initial screen load is a bit slow, it's something client side. Server responds in no time.

But after the initial load - it's blazing fast. It does thousands of messages in seconds, the online version. But the scrollDown functionallity (i use smooth mode) on chromium is a bit slow in comparison to android. On android it goes brrrrrrrrrrrrr, chrome does wait a while before scrolling down. If I send 2000 messages - chrome will wait until last one before scrolling. Android goes live.

I can reboot and everything keeps working.

I had some issues regarding layout since every component has really it's own CSS. And every component has it's extra container around it that doesn't apply the internal CSS rules. Was a bit fighting. I have somewhere a self written javascript merger laying around (easy as fuck it's kinda cat *.js > bundle.js) and will bundle all JS to one package so I can include the project also on molodetz. It only has to include script and do something like <chat-window header="false" channel="{hash from channel if channel is set to public" user="{token of a general user named anonymous}" /> and strangers can chat on the site. Maybe I will even make channel creation automatic. For example, I can let a different room render under every article on molodetz. So has every article a comment section. I need more that one one interface, need minimalist once to use as comment widget on pages. Custom CSS stylable between the tag. But now, first thing on the list for the project is a GPT bot.

GPT bot will cost me less than a hour. Finishing it with deployment and wrap in some container. Two hours tops. I'll define an admin token that the bots can use so they can make their own account. It would be comfortable if I would have a folder, with text files in it. Every text file is instructions for a bot. So three files are three different bots. One bot is harry potter, one bot is an angry ex, one bot is wikipedia, one is chatgpt. Sky's the limit. I'm considering using this system as GPT backend: https://molodetz.nl/retoor/r. It's configurable using a text file. It's ultimate for role playing. In that case I only have to communicate with a process to build a bot, i don't have to implement OpenAI shit or whatever. That application is a compiled C app with OpenAI support. It's funny, almost no one made a connection with OpenAI using C. It's also complete bullshit ofc. OpenAI is slow, there's no reason to use C. But it's decent, a compiled AI bot with text instructions.

@retoor You can switch between text and binary per message AFAIK, and not much if you're sending JSON, but it could if you want to send more than that over the wire, e.g. images

Could you not use a stylesheet for the components, then?

Integrating with the pages can be useful, maybe have it show up as rooms in the main chat ui?

Surprised there hasn't been much integration on the C side, but I assume it's because it's all hidden behind http requests.

I like the bot txt idea

@BordedDev the file uploading will happen, I think a file upload will make a new dedicated websocket connection, send json message first with details of the file that's about to send like amount of chunks so we can monitor progress and check if user is still allowed to upload so much. Ever user has a quota of 100mb or so but I make it even better - the application in total gets a quota of 100mb per user and users can just use the total all together. If there are three users, one user can do 250mb upload and the others have 50mb left. It's not about fairness or equality amongst users but just good quota expectation management. 10 users? My apps needs a gb. After sending json, it will do the binary stuff in chunks of Mb (they're high level sockets, 1mb is fine.). It needs a dedicated connection because you don't want other sockets in between or you have to do it json base64 encoded with file_id param every chunk. But that means more data and more calls.

All this stuff good idea?

@BordedDev yes, I do now include a link href into the components. It actually loads the stylesheet only once if multiple components load same stylesheet but realize now that it's the thing initial loading could take a second. Hmm, I could give css as parameter or so instead and make it later customizable for user.

I was thinking about a special option for NOT showing the rooms in chat main menu. Maybe if someone said something on a article the room will show up - some you can respond. Let it disappear again if not talked to in few days. I can manage room behavior in detail per user. Ever user is attached trough a group_member object with config overrides of channel object. Said before I guess.

The C implementation is useless. There's no reason in the world that https://retoorded.molodetz.nl/retoo... is written in C. Maybe, since it can run on every platform, but I added python as dependency so even mobility is not an argument. Will remove python plugin support maybe.

@BordedDev with python plugin support you can use the r program for serious AI automation. You can tell AI to return !exec `command` if you ask it to execute something on command line. You make a python script that reacts on sentences that start with !. For example, all sorts of commands, could be anything. But I can make such solution without python too. It will just execute a script with the name mentioned after !. In that case you can script in any language. Why invest so much in the stupid r application? Well, it's an amazing listener with memory. You can teach it var names and let him do things with those vars very accurately making it the ultimate AI automation system. When that's done, it will become the engine behind retoorded web app and you'll have a very powerful accurate assistant. Retoorded works very nice, you can learn it to do web requests on certain phrases but can't remember vars well. It says it can't if you ask to remember. It's because of openai not type.

@BordedDev you have AI bots that use pure completion and you manage its message history yourself but you can't really instruct the bot. You also can use an AI agent, you can instruct it proper and can remember up to 100.000 messages. Also, that one you can add RAG with gigabytes of data. That's how current retoorded is made and why it has so much devrant knowledge. But it's almost twice as slow and memory is a bit limited. Can also hold just 500 bytes as system instructions and that's not much.

It's hard to choose one. For a real talking assistant for conversations like I have a home, you need a fast one. But fast one doesn't have RAG and fast one has perfect but limited memory. The agent one has meh memory but almost unlimited and can learn documents. Both can learn instructions. One listen better than the other. The should make kids to become even more fucked up or perfect. Can be both. They both rock and suck. They do this on purpose. I'm sure.

@BordedDev should start a blog. You'll be less forced to read all the stuff I write 😂 But it's sad to keep all research knowledge just for myself, the research should have end product in form of sharing, else it's a waste 😂 and my observations are kinda unique first hand. I think it's worth something. Real experience based not copy pasted from internet.

@BordedDev I have some great articles in my head right now and will write one of them now (my view on the dead internet theory; before it was dead, it was already brain dead because of desk journalism with a few examples) to publish later. The second article I have in my head is huge, it's about the journey of writing a regex interpreter myself without doing any research upfront. The only information I've worked with is general advise about interpreters in general from Lorenz (here on dr). Writing an regex interpreter from scratch is a HUGE journey and I know the journey of other regex interpreter writers afterwards and know a lot about regex history since I've read the book beautiful code where some is described. 95% of people who write about regex stole knowledge and don't know whet they're doing resulting in code that won't last if you do a full regex implementation because they don't really understand. But yh, for some other say. Multiple a4's nobody is gonna read but unique.

@BordedDev 😁

@retoor I like reading them, haven't gone through all of them yet (I'll +1 the ones when I have) just busy at work atm :( kinda wish I had taken a sick day, but then I wouldn't get paid 😅

@BordedDev work is nice but costs so much time :) Did that find that 😁

@retoor A quota can work, uploading via socket is interesting, you could even use BSON\BEVE. The FileReader object (which I think you'll have to use to read from file inputs) does provide progress as well.

Be aware that base64 will increase the transmission size by about 33%, you should be able to use a higher base encoding as well.

I came across this for better loading of styles in the component, https://webcomponents.guide/learn/...

"decreasing" room activity is good, I think a lot of systems use a that type of systems, keeps the things you care about at the top

Hoenstly I think you know more about the AI space than I do 😅 I haven't even tried RAG or anything more than the basic message memory stuff. Would love to setup some form assasitant, especially for those "stand ups". Just go "did you get that" and it updates the code with the new button colour.

@retoor I wanted to create some YouTube series for some of my experiences, since I'm generally slow to write good docs (plus drum up some excitement for my larger projects) - but those ran into a similar problem + the new place has a lot of ambient noise :( . Combine that with the work exhaustion (see the time I'm replying to you when not doing other things). Reading about the regex will be interesting, dead internet theory as well - though I'd attribute it with the growing number of people connecting and using "hubs" like reddit/social media

@BordedDev my assistant can give his own page a new colors but i think I'm able to let it design his complete page as wel.

AI is a bit annoying to learn because of the lot of bad sources. My ragent source code is one of the best examples of making a rag using agents. Agents are hot now but I think they lack performance a bit. It needs at least three seperate requests per prompt and their api isn't that fast. That's why assistant at home is not based on ragent. Retoorded is based in ragent and is just a few lines of code. I didn't make retoorded open source because I'm not giving a while good working talking assistant away that works for a few cents a day. I'm sure it has still to become popular. It does exist ofcourse, openai sells it himself but its all expensive so no one uses it. Mine was a lot suffering to build with configurig microphone and stuff. A few lines that costed a lot of time. A few lines that only works on android with some effort and chrome. Dammit.

It works so great. But sad that the web version has issues with non existing words unlike my home version, the dutch speaking one. It is perfectly able to recognize custom words like names. It understands besides statistics also perfectly what is going on on dr regarding content.

Updating own layout 😁

@retoor I'll have to give your source code a look when I'm on the AI train again.

Nice! The feature goblin strikes again :D