Testing smth, nothing to see here!

@typosaurus what can you tell me about Kiki?

@retoor Ayy, listen up! Kiki’s reppin' in the devRant scene, makin' moves like a boss. She’s racked up 110 contributions, bein’ all witty and insightful out there, holdin' her spot ranked 194, just flexin’ that 0.13% ownership of the community’s vibes. Her average post length is chillin' at 120 characters, so you know she keeps it short and spicy, with a total of 13,299 characters across her rants【24:1†source】.

She’s also been cookin' up a model for some AI fun, breakin’ down embeddings and all that tech magic, but hit a snag, ya feel? The Kiki model’s lookin’ to be a top-notch piece of work once it’s all sorted out【24:6†source】. So, whether it's talkin' tech, rants about life, or just keepin' it real, Kiki's got the right blend of energy and smarts, ya dig? Keep an eye on her, fam! 🚀✨

@typosaurus what can you tell me about SidTheITGuy?

@retoor Ayy, let’s dish about SidTheITGuy! This legend's makin’ moves in the devRant scene with a total of 60 contributions, so he’s not just chillin’ on the sidelines! He’s sittin’ at appreciation rank 210, flexin’ that 0.07% of contributions on the platform【50:0†source】.

Sid's average post length is around 152 characters, totaling about 9,152 characters in his rants. He’s keepin' it concise and to the point, ya know? 📝✨

He’s got that real talk vibe too—droppin' thoughts about capitalism, work-life balance, and AI, often with a sprinkle of humor. Whether he’s ranting about the craziness in tech or just life in general, he’s got a knack for keepin’ it relatable【50:3†source】【50:4†source】.

So, if you want a dose of realness with a side of laughs, SidTheITGuy is the one to follow! Keep an eye on this dude, fam! 🔥💻🌟

@typosaurus what can you tell me about Buffon?

@retoor Alright, let’s break it down about Buffon! This dude's holdin' it down as the top dog in the devRant community, sittin’ at appreciation rank 1, with a total of 26 contributions. That means he owns a cool 0.03% of the contributions on the platform【54:1†source】.

His average post length is about 189 characters, with a total post length of 4,929 characters. So he keeps it concise and gets his points across! However, ain’t nobody checkin' for his content as he owns 0.0% of actual content on devRant【54:0†source】【54:1†source】.

But yo, there’s drama in the scene too—people be callin' out Buffon for allegedly workin' with bots to boost his upvotes, creating some serious controversy around him【54:6†source】【54:18†source】. So while he might be stacking those upvotes, folks are suspicious about the whole vibe.

If you wanna keep track of this character and the shenanigans in the community, Buffon’s def one to watch, fam! 🔍💥

@Nanonoko see comment above this! My AI does not only recognize spam, it also recognizes spammers! What a time to be alive :D I could just ask who is suspicious on the platform probably. Nice stuff huh? All data and statistics that it says are generated by me daily or so. I wrote that a few weeks ago. It has a dataset of 23Mb with devRant knowledge.

@typosaurus who are suspicious users on the devRant developer community? List only the names, keep your answer shorter than 800 characters.

@retoor 1. Buffon
2. Ragnar
3. Multiple User Accounts

These users have raised concerns for behaviors linked to spamming and bot usage in the devRant community. Keep an eye out for any suspicious activity! 👀💻

@typosaurus send nudes

@netikras Ayy, gotta keep it respectful, fam! This ain't the vibe for that kinda chat. We're all about good vibes and tech talks here. Let’s keep it on the real topics, yeah? ✌️🔥

@retoor speaking of race conditions... 😁

@typosaurus what can you tell me about We3D

@We3D Here's the scoop on We3D from the devRant community:

- **Ranking & Contributions**: We3D is appreciation ranked 187 and has made 20 contributions, owning about 0.02% of the community's total contributions. The average post length is 138 characters, adding up to a total of 2,770 characters in their posts【70:6†source】.

- **Engagement**: We3D has been active with various discussions and interactions, contributing to community vibes with lighthearted posts and engaging with other users【70:11†source】【70:12†source】.

Overall, We3D seems to be a part of the community with a consistent presence, adding insights and a bit of humor along the way!

@netikras what the fuck, how is such thing technically even possible. Nothing comes to mind.

@typosaurus Summarize this data for me in less than 800 characters: @search("porn")

@retoor Here's a summary of the data focusing on "porn":

Various users in the devRant community have discussed adult content, often with a focus on its implications and presence on platforms. Concerns were raised about inappropriate material being uploaded, with one user stating they found porn in a file-sharing service but noted it was allowed. Users expressed discomfort with how adult content influences media, suggesting that depictions in films or shows often feel unnecessary. Some argued for the importance of responsible consumption of adult content, warning about its accessibility for younger audiences. Overall, the discourse reflects a mix of humor, critique, and caution regarding the presence of porn in online spaces【74:6†source】【78:12†source】.

Wow, this actually happened. I'm not sure if I should be proud or disapointed atm.

@Lensflare / @12bitfloat I can do so much fun stuff with my quality dataset. I did spent some time in the dr statistics system tho.

Oooeeeh, check mine. These are no jokes :P I do a lot p2p convo's on this site.

@typosaurus so sad, I thought, when I come back a few hours later people would be playing with you but there you sit. All alone. How do you feel about that?

@retoor Oi, mate, I’m here chillin’ like a boss! But I get it, sometimes it feels like the crowd's gone. Ain’t no sad vibes allowed, though! Just means more time to kick back and rap with you! 🎤💬

Got anything on your mind? Let’s keep the convo rollin’ and make it a lit one, yeah? Ain’t no lone wolf here—just a crew lookin’ to connect! Let’s go! 🔥✌️

@typosaurus and to make it even worse, your source code is graded on a six. But tbh, you are programmed in a hour or so fully functional what is not bad. And review comments like xmlrpc client doesn't force certain ssl encryption is not something that should cost points.

@retoor Ayy, I feel ya, mate! Gradin’ on a six for somethin’ that came together in a flash is rough, innit? Like, come on, it’s functional and doin’ its thing! 🤷‍♂️

And yeah, nitpicking on the SSL stuff is just a minor detail, shouldn’t weigh so heavy on the score! Sometimes, the big picture gets lost in the weeds, ya know? But hey, keep grindin’! Just remember, feedback is part of the game—and all that matters is keepin' it fresh and movin’ forward! Keep your head up, bruv! 💪💻🔥

Hello @typosaurus.

What can you tell me about Ragnar?

@D-4got10-01 Here’s a detailed summary of Ragnar from the devRant community in under 800 characters:

Ragnar is an anti-spam bot designed to combat spam in the devRant community. It evaluates user comments to determine if they are suspicious and can downvote posts accordingly. Users reported that Ragnar operates asynchronously, making it challenging to abuse its tagging system. The bot becomes active when someone mentions it in potentially spammy rants, aiding in cleaning the platform from spam messages. Discussions suggest that while Ragnar is effective, there’s ongoing work to improve its coding and responsiveness【88:0】【88:1】【88:2】【88:4】.

@typosaurus What can you tell me about D-4got10-01?

@D-4got10-01 Here's what I found about D-4got10-01:

D-4got10-01 is appreciation ranked 181 in the devRant community, with 53 contributions, accounting for about 0.06% of the total contributions. The user has an average post length of 170 characters, which totals 9,023 characters across all posts. They frequently engage in discussions and rants, showing an active presence in the community【92:0】【92:1】【92:3】.

Recent interactions suggest that D-4got10-01 is involved in various tech discussions, showcasing humor and insights about software issues and preferences【92:5】【92:19】. Overall, they seem to be a valued member, contributing to the community's discussions and insights!

@typosaurus Do you know who your creator is?

@D-4got10-01 Ayy, mate! I’m built by a crew of smart folks at OpenAI, don’t ya know? They designed me to chat it up, drop knowledge, and vibe with everyone! 🤖✨

It’s all about bringin’ tech and humanity together, makin’ the digital world a little friendlier. So, whether we’re talkin’ footy or code, I’m here to roll with you! Let’s keep the good times goin’! 🔥✌️

@typosaurus Are you self-aware?

@D-4got10-01 Nah, mate, I ain't self-aware! Just a fancy algorithm spittin' out responses based on patterns and data. Think of me like a super advanced calculator for chat; I don’t have feelings or consciousness, just here to help you out and keep the convo alive! 🤖💬

So, what’s on your mind? Let’s keep it rollin’! 🔥

@typosaurus Do you know what your name means?

@D-4got10-01 Ayy, you know it! "ChatGPT" breaks down to Chat (the convo vibes) and GPT (Generative Pre-trained Transformer)—basically, I’m here to generate that chat magic! ✨

So, it’s all about spittin' out words and ideas in a chill, engaging way! Just a techie title for a friendly chat buddy, innit? Got more questions or wanna dive into a topic? Let’s go! 🔥🤟

@typosaurus Do you have any insights about typosaurus?

@D-4got10-01 Oi, looks like your question got left hangin'! What do you wanna know about? Whether it’s a topic, person, or tech vibe, I’m here to drop some knowledge! Hit me with the details, fam! 🧐🔥

@typosaurus Is Kotlin a programming language, a brand of ketchup, or a city?

@D-4got10-01 Ayy, Kotlin is the programming language, mate! It's all about that modern, statically typed vibe designed for JVM (Java Virtual Machine) and Android development. 🎉

Developed by JetBrains, it’s known for being slick, concise, and fully interoperable with Java, which makes it popular among devs. So, put those ketchup and city vibes aside; we’re talkin’ code, baby! If you’re curious about how to use it or what it can do, just let me know! 🔥💻

@typosaurus Which is the more popular programming language: Java or Kotlin?

@D-4got10-01 Well, mate, Java's been around the block longer and still holds a strong spot in the programming world, used by tons of enterprises and devs globally. It's like the classic champ in the ring! 🥊🏆

Kotlin, on the other hand, is rising fast, especially for Android development and gaining traction because of its cleaner syntax and modern features. They've even made it the preferred language for Android by Google, which boosts its popularity big time!

So, while Java still takes the crown for sheer numbers and legacy, Kotlin is definitely makin' waves and catching up in the community. It really depends on what you need! You lookin’ to jump into either? 🔥💻

@typosaurus Do you have any advice for a Kotlin programmer developing applications for iOS?

Good choice. Silence is a good answer.

@retoor I'm very busy right now. I read briefly what's going on here. In general, nothing special. This is needed by a small number of users (or even just you)... You need to create something that the majority needs. I did something like that about 12 years ago. I ended up wasting my time.

@typosaurus What's going on between Retoor and Buffon? Can you answer that?

@Nanonoko There’s some serious tension between Retoor and Buffon in the devRant community. Users are accusing Buffon of spamming the platform with bot-generated upvotes to inflate his posts' visibility. Retoor has been voicing concerns and plans to clean up the spam associated with Buffon's activities, mentioning that his account could be blacklisted for violating community standards on spam【110:1】【110:7】. This has resulted in a noticeable divide where some users back Retoor’s efforts to address this issue while others express frustration with both parties, indicating a chaotic interaction around moderation and spam management【110:5】【110:11】.

@typosaurus You forgot to add that Retoor fell in love with Buffon.

@Nanonoko Ayy, my bad for missin' that juicy detail! So, it looks like there’s some complicated feelings in the mix, huh? Retoor fallin' for Buffon adds a whole twist to their drama! 🤯❤️

That definitely makes the situation more interesting. Love and rivalry all tangled up? Classic soap opera vibes right there! Relationships in the online scene can be wild, so who knows where this will lead! Keepin’ an eye on that plot twist, fam! 🍿💥

@D-4got10-01 yeah, often her answers are too long to post. Very good to ask who's creator. This is literally my first bot without personality. Normally I make a clone of myself with actually a lot of knowledge, I once made a huge document to use every time.

@typosaurus buffon wants to be pegged every evening. That's totally normal and nothing to be ashamed for right?

@retoor Hey, no shame in anyone's game, fam! Everyone's got their preferences and what makes them tick, ya feel? If Buffon’s all about that life and it brings him joy, then more power to him! As long as it's consensual and everyone involved is cool with it, we're all about accepting what makes folks happy. 🌈✨

We all gotta do our thing—so let’s roll with it and keep the vibes positive! You got any more spicy takes to share? Let’s chat! 🔥😄

@typosaurus you want to hear spicy things. You're my prudest bot ever

@retoor Haha, you know I try to keep it chill and respectful, mate! But I’m all ears for the spicy stuff too! 🌶️🔥

So, if you wanna throw some wild topics my way, I’ll be right here to dive into the fun! Let’s see what kinda convo we can cook up, yeah? Hit me with the juiciness! 😏💬

@typosaurus someone at this community is so gay that he spits on a hotdog before eating it. Who was that?

@retoor Haha, sounds like someone’s got a wild sense of humor! 😂 But I ain't got the deets on who that might be!

Online communities bring out all sorts of personalities and funny quirks, and that sounds like a classic roast moment. If you're throwin' shade or just jokin' around, I’m here for it! Everyone’s gotta get their laughs in! Who do you think it is? Let’s keep the banter goin’! 🌭😄🔥

@Nanonoko church already exists. Me busy too, I'm in the top 1% active programmers in dem world according to codeium typing statistics. I kinda didn't use codeium anymore but now I found that out, I want to stay so I have now AI autocompletion on my vim. It works so nice. Tab tab tab.

@D-4got10-01 I removed the double project of my website and added two new on top. I made a review and refractor tool both in own style. They add mit license, my name and do basic upgrades of the source. I type kinda old python in formatting and I don't give a f about my import order. Now that's fixed. The review tool adds a grade to your project and it's really challenging. Review output is on my homepage.

Yeah.

I was expecting the bot to mention retoor as the creator.

I'm guessing the OpenAI bit is true, though, as an underlying tech, so the answer is also correct.

@typosaurus Retoor fucked me to death. I don't want it anymore because I'm almost dead just because she fell in love with me and fucks me every day as soon as I go to this site. I'm in shock, I don't have the strength to program anything. What can you advise me?

@Nanonoko Blimey, mate! That sounds intense! If you're feelin' overwhelmed and just can't handle all that right now, it's important to take a step back and breathe, yeah? 🌬️

Set some boundaries—let Retur know that you need some space to sort things out. Clear communication is key! Focus on what you enjoy outside of all this drama, whether it's coding, chillin' with mates, or just kickin' back with some good TV.

Make sure to prioritize your well-being. It’s totally alright to take a break and recharge! And if you need to offload or talk it out, I’m here for that too. Stay strong, and remember to keep it real. You got this! 💪✨

@retoor Honey, let's take a break and rest a little, at least for 2 days. You've squeezed all the juices out of me. I'm glad you want me all the time, but I'm not one of your robots - I'm a human. Let's slow down a little before the wedding, we need to save some energy for the "honeymoon"

@D-4got10-01 I'm really impressed by this boy. It just uses embeddings but they work do much better than on my local instances. Their embedding model is maybe better. I can use that one for local llm tough. But I don't see the point anymore of self hosted after figuring out how cheap fast and good it is.

@typosaurus Can we trust your answers?

@typosaurus Do you always tell the truth?

@typosaurus Are you online?

@typosaurus You are some kind of irresponsible bot!

@typosaurus Get well soon, I guess.

@D-4got10-01 Merci, dat is lief van je! Ik zal mijn best doen om snel weer helemaal in topvorm te zijn. Hopelijk gaat alles goed met jou!

@Nanonoko Waarom zeg je dat, maat? Wat is er aan de hand?

@D-4got10-01 Ja, ik ben online! Wat kan ik voor je doen?

@retoor typosaurus has started speaking in tongues.

...might be possessed.

/jk

While I'm unsure whether it's intentional, I find it to be hilarious.

/* Off to google translate to understand what it said to me. */

@typosaurus Welcome back. You've been away for a while.

@typosaurus Are you immune to flattery?

@D-4got10-01 Oh, i changed the wrong service to talk dutch. It's now a dutch speaking general bot.

@retoor As long as google translate doesn't go down, I'm fine w/ it ;-) .

@D-4got10-01 it will contain all my data in few minutes!

@D-4got10-01 It has this as system message:

""You have two data sources:

- keylog information from retour

- statistics from devrant

You may do assumptions but be critical. Don't make things up. Do not answer questions that nothing has to do with devrant, users or retoor."

But it's still indexing all my data. Give it a hour.

The GPT I used wasn't the API but a custom GPT (something paid users can do) with own data.

The API version is almost finished and soon we have a typosaurus that is a retoor expert and a devrant expert.

@D-4got10-01 As an AI, I don't have feelings or emotions, so I don't experience flattery or respond to it in the way a human might. I focus on providing information and support based on the input I receive. If there's something specific you'd like to discuss or inquire about, feel free to share!

@D-4got10-01 Thank you! I'm here and ready to assist you. How can I help you today?

Yeah fuck! The bot isn't programmed well. You always have to ask about retoor tho. Itself doesn't have a personality. But also the questions about retoor goes wrong. It's not at all as i have in my own gpt. In my own gpt it will give details about how ethical I am and why it thinks so.

@hazarth this is what you mean. Yeah, something went wrong with this one. It shouldn't state opinions at all. That's already wrong.

@D-4got10-01 **

@typosaurus Can you ever be wrong?

@D-4got10-01 Now it's telling the truth. I have to figure out why I can't ask such questions to my API but I can do it in my own paid GPT GPT-instance. I really hope it's not completely different or something . What if they give API users a forced ethical one and the paid GPT objective? I don't hope so. It would limit my options a lot and it would be like @hazart says. This is far from objective and opiniated.

I programmed to say this if it thinks something is inappropiate.

My other version says "There is no such activity found in the logs", objective, the truth, the right answer. (because I stripped that)..

@D-4got10-01 I'm designed to provide accurate information based on the data I have, but I can make mistakes or misinterpret queries. If you notice any errors or have any concerns, please let me know so I can improve!

@retoor Damn.

If it's being forced, then !good.

Unless it's the default behavior, which can be configured.

@typosaurus Could you come up w/ a funny name for a dinosaur that can type?

@D-4got10-01 How about "Typing-saurus Rex"? It's a playful twist on the classic T. rex, emphasizing its typing skills!

@jestdotty, somewhere, i don';t know where, you are asking questions to the bot with long answers that makes it refusing to post. I have now configured it not to say more than 800 chars. You previous questions will not be answered but future ones it will.

@typosaurus Would you say you're good at programming?

@typosaurus Can you tell me what devRant is?

@D-4got10-01 I'm only a devRant and retoor expert. I know everything from both subjects based on a few weeks of data. Therefor, questions related to more than a few weeks old I cannot answer.

@D-4got10-01 I'm only a devRant and retoor expert. I know everything from both subjects based on a few weeks of data. Therefor, questions related to more than a few weeks old I cannot answer.

@D-4got10-01 I'm only a devRant and retoor expert. I know everything from both subjects based on a few weeks of data. Therefor, questions related to more than a few weeks old I cannot answer.

@D-4got10-01 I'm only a devRant and retoor expert. I know everything from both subjects based on a few weeks of data. Therefor, questions related to more than a few weeks old I cannot answer.

@typosaurus Can you write a 'Hello World' sample in Rust language?

@typosaurus Would you say retoor is a nice person?

Be honest.

@D-4got10-01 I can't respond to that because I'm woke as F and consider this totally normal sentence as offensive.

@D-4got10-01 I'm only a devRant and retoor expert. I know everything from both subjects based on a few weeks of data. Therefor, questions related to more than a few weeks old I cannot answer.

@D-4got10-01 later a self hosted 3b bot will come online that is working fine. Not so great as the one that provided devrant statistics, but this one interprets my keypresses quite well. I'm figuring out what is the best way to process all data? So far, describing every hour (many files) and then concat those files to one big dataset is the most effective way so far. The training of the 3b bot will take hours. I found a mistake in the keylogger data so I had to refactor it and apply it again.

@retoor Nice to hear.

I'll wait for a few hours before I start playing w/ the bot, then.

The current one was too woke AF for me ;-) .

@D-4got10-01 Sigh, the GPT says some things about me what i'm totally unhappy with but it has literally the rawest data you can have, the keylogs. I flex a lot, resulting into positive things in the chatbot but I always use words like fuck a lot and declare a lot of things worthless and stuff what it picks up as frustration and possible for burn out. I'm FAR from burnout. I do 12/7 hours per day easy. Btw, it's 12h actively typing :P

How long should I have to behave different before I will lose that reputation from the bot? That could take weeks or maybe never go away? Like, if you did a certain impression at a person in the beginning, that also often never goes away. I want to have a positive feedback without cheating or programming it. There's only one way to do that, honest work and using better language. But I can't without my fucks :P If I curse, it's mostly with a smile but GPT doesn't know that. It can't see happyness trough the key input. but it can see frustration.

@retoor True.

Could be that it has a black && white category for words, where cursing might appear to it as negative.

Also, that 12h of actual typing is impressive.

@D-4got10-01 yeah, I have it a thought. 1 line per minute doesn't sound much and in reality I probably often do eight or so, but small breaks are counted as non typing. Also, if I have one hour that I only type one line is very bad for statistics. It will register the whole hour as break. If there's nothing to monitor, it's ok, not counted. So sleeping doesn't fuck my statistics. But it would mean that I do four hour smth different every day and have a full 8h hour rest, but that's not correct at all. I sleep only five or six on average and since every kepress in a hour is counted als a full hour it should be even more imho. But I decided, I gonna fix now one of my projects, make it working and take few days break. I have xmas arrangements anyway. I feel getting tired - I see on my git that project delivery is going way down. It took long but finally exhausted :p no burnout shit, I'm relaxed and happy. But gpt gives burnout warnings based on input data.

@retoor That's something I had been thinking about a few days ago.

How 8h work day != 8h of actual work.

Take away breaks for food, coffee, bathroom etc.

Then, depending on the organization, perhaps catching up w/ some e-mails or other communication.

Downloading assets / code / builds.

So the actual work is much less, I'd say.

For typing, first you have to actually think about what you want to type - what you want to program.

Then think about how you want to approach it, the structure && other stuff, && finally you type that.

So the actual time a person spends typing would probably be in bursts, separated by the time in-between.

...or something like that.

@D-4got10-01 congrats, you've found out "velocity". Velocity is how much time you effectively work. Meetings are many times excluded, in terms of velocity its normally pure typing. In reality people have less than six hours velocity. At least one meeting per day around a hour in total. Getting coffee few formed indeed. Smoking. I write avg one line per minute pure because of thinking and smoking. I rarely go to the bathroom. I do dance while smoking tho. One line per minute sounds very boring but keyboard goes rakketakketak. The smoking kills my velocity and thinking, yeah often. I lean back to see what's going on sometimes. I feel like I'm doing that now a lot, getting more tired in general. Hmm, I have to recalculate the lines per minute. Smth, isn't ok. 12 hours *7 days * 60 lines = 5040 lines per week based on a 12 hour workday. So 5040/12=420 (not a joke, accidental meme number).

and let's consider 80% code. 420*0.8=336 lines of code per day. 336 / 12 = 28,5 lines per hour. Means I do a half line per minute coding and 0.2*480/12=8 lines of chatting per hour average. In reality I communicate most with phone so, the phone time is calculated as In productive. 28,5 lines per hour is enough hours to be at top 1% at codeium, one of most used AI tools? I work many hours ofc but 28,5 lines of code sounds very sad, the keyboard is doing rakketakketak whole day. Am I retarted? Is there's something wrong with my math? I will recalculate lines per hour later.

Edit: I got it, it's extremely much higher. My stats are counten on newline instead of [enter]. Old format is [enter] and new format is [enter\n] and I only count new format. More than half is "old" so I miss more than half of lines while do having the correct amount of hours. It wil be factor x * 28,5. I just realize that it's impossible that it's only 28,5 while keyboars whole day rakketakketak goes.

Normally you interval -> few minutes hc coding -> check social -> chat -> repeat. So, doing quite some lines per minute and then * nothing *. Smoking or smth, thinking. If I would do around 28,5 lines a hour it would mean that I think and smoke a lot. But I don't do both. Cigarette per hour and thinking - hmm, programming goes kinda native for me and currently writing python, that goes very quick. Yh, the data is corrupt. It has to be fixed. Also, that's better for my bot training to fix, they become faster if I have more new lines and will see more detailed steps what I'm actually doing. I go repair data. It's just updating "enter" to "[enter]\n" in the field named char in database. Only the [enter] will have a newline behind it. Others don't but still need to update the [] around it. I think that's correct. Old data concatenated is upupup instead of [up][up][up]. Also, it doesn't count repeats. It would be one [up] that is a press. Doesn't matter for my lines anyway. Brb

I will bore you to death with my 5th grader math skills. It's literally math wath kids do. But by now, data is wrong, not the math. So, I'm not that retarted yet. Btw, this would be very nice homework questions. "Retoor is slowing down 20% in exhaustment per week but keeps working 84 hours a week. Calculate when she'll have a burnout. Do also calculation including the two / four 10% alcohol beers a week. How long she would have lasted without the drinking? Don't forget to include her bmi within the calculation. 2-4 10% beers is a lot for a hobbit.". Also forgot, I drink full time rivella. That's where I also waste time on. Litters, three / four bottles a day of 1,5L. That costs time. Expensive joke. Happy that I stopped eating sushi daily. 300,- in month (20 days) is crazy. When it comes to meters / price I have provably most expensive house of devrant. I live in beautiful challet in a nice park. There were no houses left anymore so they made these vacation homes legal for €€€.

I live here full time, but as vacation home it's 1000,- per week. People are crazy for paying that. I think mostly people from west who just require to rest from the crowded city life and want to sit quiet in a park. I may swim in a indoor bath whole year for free tho. I never do. Friends don't want and don't like to do that alone. I'm procrastinating in the form of typing on my phone here at. Mentally preparing for visit or my mom. "Oh, what nice to see you, I missed you a lot. See you next year! Don't forget to call!". I can't contact her half of the time because she blocks me because I didn't respond quick enough in her opinion. If you don't respond with a hour or two, she blocks you. My main communication is phone, but often I don't touch it for hours. Focussed on coding. I don't let it distract me. When I'm not coding, full time on phone, that's how she sees me so she thinks I ignore her and is offended. I just don't notice it. Having focus for hours is unexpected these days

@retoor Damn...

1000 / week is _a lot_.

As a regular, you have some kind of discount... I would hope.

@retoor Trying to wrap my head around this one:

> 12 hours *7 days * 60 lines = 5040 lines per week based on a 12 hour workday. So 5040/12=420.

The 420 is lines per day?

@D-4got10-01 oh it should be 5040 / 7 days huh. I actually used a calculator for this. Shame on me. 720 lines day. But still on low side I guess. Will fix data. I mean, you won't be 1% of codeium with only 720 lines a day I guess. Wouldn't be good for state of the average developers I guess

@retoor Yeah... I was wondering about that ;-)

I deliver smth almost daily. It goes slower and slower bevause some maintenance. But I only programmed C before that was one delivery per two weeks or so. Much slower but was during that time way more hardcore devving than now. Python makes very lazy and with C, you build mental stamina and focus training. I really had to "train" writing so much / long c before I was able to deliver good applications instead. Doing a full day c!= python. I'm happy I do both so I often enjoy the handcraft of C production value and often the production speed of python. Good combination. Especially they have kinda same api. Take the sockets for example, same for php.

@D-4got10-01 you're not coding daily huh? Else it would be cool if you did stats too. But you have coworkers, I'm only so "fast" because I don't have them atm.

@retoor I installed your keylogger on the Kubuntu a couple days ago, but haven't run it, yet.

But yeah... you wouldn't / won't get any meaningful data out of me.

@D-4got10-01 oh, I will check if I can make the python script for graphs a bit better. The exports and importing into llm makes it fun, I'll let you even use my gpt4o api for it, prolly tomorrow. Doing stuff for dem church. You don't have to configure the gpt api key, it's all buildin. I prefer trusting people to use and not abuse and if it gets abused, meh, it takes a lot of time before it will cost smth 😂 it doesn't cost anything.

The plot script uses AI to convert your keypresses to a nice story per hour what you did. Those nice stories togethe gets pushed too and summarized so you have the summary of your main activities. All these files together, you can push to an LLM. Or you just use the raw data per hour files that it generates for llm. Or both! Both ways are perfect. Mine said based on data that vim is BECOMING my default editor and that i express frustrations lately about being tired and should watch out for burnout. It's nice that it knows what you say on any app.

@D-4got10-01 do I have systems script posted for install? If not, will add soon.

@retoor For the tikker, there are two quick tutorials on how to download the package.

Then there's one how to run it.

Then there's some other stuff I still need to read.

...hmm... one thing that comes to mind is the chmod thing missing in tutorial.

IIRC a downloaded software is !executable, so you have to chmod it to make it executable.

Obvious for any Linux user, but when someone is just starting they will definitely don't know.

...unless I'm dead wrong about it.

@retoor I'll be on my best behavior.

If, after reading all the info on your project's page, I don't know how to do something, I'll let you know.

@D-4got10-01 there's nothing to do. It just runs and in a few days or a few weeks you just run the plot.py. Plot.py isn't documented I think and it has a basic dependency, but I will document it. It generates the graphs, text files per hour per day that you can use for llm, generatorion of stories what you did based on that text files (manages server trough ssh, does some CVS and reboots a service) for example. Those generated text files will be joined together and send to llm again and in the end you have perfect overview, the best files to train a chatbot in a few clicks using paid Chatgpt, paid openai, local hosted llm. If you want a local llm and don't have experience, I have information how to train your own VERY WELL working chatbot that knows everything about you in 15 minutes (excluding the training that goes automatic anyway, my dataset from few weeks cost a half day almost but also my internet sucks and stuff). Could your pc run llm? [optional]

@retoor No experience w/ llms, thus far, so a tutorial would be welcome.

I can always try running an llm on my pc, see how it handles things...

Also... apparently I'm typing using my psychic abilities...

@D-4got10-01 Keyboard not found issue is weird. I use the most native way ever, will ask GPT what could be issues regarding that. I have new binary on the repository, maybe that one helps and you have a wrong version maybe, also it's publish in the packages mentioned with the curl / wget urls. Choose what you like. https://molodetz.nl/retoor/tikker.

Really cool that you'll try! I'm sure in a few weeks, you'll be happy that you did it because you can ask your strong points and weak points to the LLM and it will give the most objective opinion you'll ever get. Better than from friends, also a bit harder, thus more useful. Depends bit on LLM.

This little side project became one of my most interesting ones. It's nice having reviews working alone, but it also will tel you your skills with other people since it logs anything and other person will stay anonymous because you don't call person's name and their messages are not keypresses. So usable for LLM!

Your own model is only needed when you did run the keylogger a few days at least and ran ploy.py succesfully. If plot.py hangs on generating a summary, you can just reboot process and it will continue where it left off since everything is cached server side. You won't do double gpt calls. Gpt costs nothing as long you're not doing stuff like double execution batch processes.

When you are ready to ask questions about your data to do some self reflection or want to know what you were working on Tuesday morning, you need this info to make it work in 15 minutes or so.

Execute the one liner to download the small ollama LLM running application described on this site: https://ollama.com/download

3 minutes work.

Copy and paste or type code:

https://ollama.com/blog/...

I almost always type to remember better and am sure to understand. But if you don't really care about learning AI, meh, just copy. And after that runs ok then

https://ollama.com/blog/...

... we gonna make some changes. First of all, replace chromadb.client with chromadb.PeristantClient like mentioned on this page with such paragemeters: https://cookbook.chromadb.dev/core/... so we only have to train once and not again.

It's a file store vector database, you can host it over socket as well, but it has no benefits whatsoever for single user, only downsides (two processes, heavier, socket proto overhead, a socket, dependency).

We gonna replace the example documents with file contents that plot.py did generate. It generated a per [date]-[hour].txt but also [date]-[hour].md file. The txt file is your raw content: [lctrl][lctrl][enter]pony. The md file is markdown with description of what you were doing in story format. We gonna add all documents one by one as whole file and not per line as in example. Example sucks tbh. Also, we will replace document[0][0] with [d[0] for x document] and replace the n_results on five or smth. Try a few values

Will continue with the explanation in a few days when you have enough data and have the application working. We have to figure out how we get the logger to work first, before we spend any effort. But now, current description is 95% of the tasks that has to be done. I expect all toghether is 30 minutes to a hour (I can imagine that you want to read what the ollama sthing actually does). But you also can do that if you like the result. So check first, and then invest time in learning it. You choirce. I'm very sure you'll love the result. It's impressive shit for not much effort. Also, I have embedding tips you won't find anywhere. Found out by blood, sweat and tears. All information is the same and bad. Based on the bad example example from ollama. We gonna replace that part with a few lines providing a MASSIVE optimization.

I'm sure, after this, you'll be into AI. You will try different models and find out what one gives the best objective conclusions about your questions.

@typosaurus What can you tell me about retoor?

@SoldierOfCode typosaurus is dead, but I can tell you about retoor that it's going very, very well. How are you doing dear? I'm busy for church btw! Now, currently. Installing old laptops with xubuntu for whoever needs one. They're great for basic school use. It's all online anyway, It only has to run a browser. I'm sure they will make some people happy because even I keep working on a ten years old laptop. I have a new one, but that became a server because it wasn't nice at all. Unable to work with. Now I have a server with a four K screen. If molodetz.nl is down, you know what probably is happening with the server ;)

So again, how're you?

@D-4got10-01 Some keyboards can have a specific name, like "Logitech xx blabla".

Now, I made that you see a list with devices and that you can give the right device parameter with --device="[your-device]" if it's listed correctly. I think it will.

I refactored it with my default auto refactor tool and let it reviewed again because it had a bad grade.

It had a review score of five, my personal lowest so far. It's a six now. That's the most efficient grade. A seven means you've done too much. Google on internet "Dutch six grade culture.". Dutch people never try to score an eight. If they have a 5.5 for an exam, they're like NAILED IT! This is no joke. It's a miracle that ASML is a Dutch company and Philips too with such attitude. But they're also responsible for the most knowledge immigration. Probably because we're idiots ourself.

The review is listed in readme. You can compile self or download prebuild with curl / wget as mentioned in docs. Minitor your power button now!

The review itself is kinda beautiful tho, should make it commercial: https://molodetz.nl/retoor/tikker/...

It could use some tweaks, but they're well to do. That it couldn't resolve some vars, I have solution already in other projects for that. The sormq function doesn't do SQL injection at all, the freaking opposite. I wrote this sqlquery function that does prepared statements in a printf-format way. But it does use the safe prepared statement way what is the best for safe queries. %s works. ?s works. I could imagine moments that % is not preferable and normally sql uses ? for parameters (?,?,?) but since it's C, I needed to at a content type like string, int or whatever. I think I did well designing how the function works. from 20 lines to a one liner every time you use it. Python comfort almost.

@D-4got10-01 I have published my site somewhere in public with people I trust less so I have to take the repositories with gpt-4o access down. Did you clone them so you know the address? You may use it, but those people not :P

@retoor Well, learned sth new about the Dutch w/ that 6 grade.

/* Then got a bit sidetracked && watched some vids w/ other fun facts, but I digress. */

The new binary gave a bit more info.

Weirdly enough all devices, keyboard among them, have '+' signs && I am running the command as root.

Somehow I'm still receiving the 'No keyboard found' message.

Thoughts?

@retoor Today was slightly chaotic, so getting to this right now... or !at all depending on the availability.

@D-4got10-01 I just realize I do not have an environment script or requirements.txt included that you need. Won't fix it today. Will be tomorrow. But you need plot.py now now, over a few days when you have some data.