Details
-
AboutI'm Retoor. Buffon is in love with me. We both love devRant and are ready to fight for it. I have more than 15 years professional experience in almost every common language / database. Mainly backend.
-
SkillsGetting eaten out by Buffon day and night. Besides that I do other fun stuff with C, C++, C#, Python, wren, PHP, SQLite3 and AI LLM's. Officially a backend dev.
-
LocationZwolle, The Netherlands
-
Website
Joined devRant on 12/12/2024
Join devRant
Do all the things like
++ or -- rants, post your own rants, comment on others' rants and build your customized dev avatar
Sign Up
Pipeless API
From the creators of devRant, Pipeless lets you power real-time personalized recommendations and activity feeds using a simple API
Learn More
-
Hi everyone,
I was surprised how much lessons there are to be learned using OpenAI api, especially when configuring your own agents..
So, I made a small doc in the form of a class for people who want to start right away containing all urls and lessons learned. I especially made NOT a library of it - like many losers did - so it can be native in your project and you don't have an extra dependency. It's in exception of OpenAI api dependency free on purpose.
You could have your own GPT in the CLI within literally 15 minutes. Links to dashboards where to register what is all mentioned in the doc. It's written for programmers but just not familiar with python and programmers who want to start directly without issues what seems to be hard in the AI business. A lot low quality stuff out there.
https://molodetz.nl/retoor/gists/... -
Dammit, it's morning. Then I get my kinda post nut clarity. Things I thought last night to be a good idea, not anymore. Good that I went to sleep. Phew. New rule, only decide what to do regarding new projects in the morning. The new background of my site is a good example when deciding things tired. What serious dev has such stuff? I dunno, maybe I do leave it for now, site is because many reasons unprofessional. There is a duplicate one with my real name on it. I made a reverse proxy project that replaced http content by interpreting httpd, fixing content length after replacing (else browsers will endlessly load or give error) with support for websocket and buffered content so implemented quite a bit. If I replace retoor with my real name now, you'll see it literally everywhere, I git history and such. Probably even downloaded zip files, I have to see if that doesn't corrupt. This software can also be used to make sure smth is NOT published. Sensitive data you could put a password in it for example so Noone will figure out if you accidently stored a password in git. I check this now by grepping using my common passwords in git. But I use env vars right now for passwords.
Got off topic, no decisions when tired anymore!5 -
I have paid version of GPT now. Kinda accidental, I wanted the API version and that seemed to be smth completely else. Didn't know. I have a lot of opinions / mixed feelings about it so far. Take the imagine generation, this image is kinda perfectly what I wanted. But it took so much iterations and it was forgetting what i've said lines before. Everytime adding text balloons why I said to remove it, then it removes it but generates a complete new image with something else I don't want with it. THan i fix that, and tadaaa, again a text balloon.
This is the question where it is about: is GPT in current state worth the money? I have no idea, I need some more time.27 -
Some people wanted to download their rants / comments. I'm working on it.
Three lines of native python code (no dependencies) to see what @Lensflare said:
from pprint import pp
from xmlrpc.client import ServerProxy
pp(ServerProxy("https://victoria.molodetz.nl/rpc").get_comments({'id':{'gt':42},'_limit':1337,'username':'Lensflare'}))
I think this gives example of possibilities enough. Use your fantasy on how to retrieve rants.
Limitations:
- Not whole dR is available yet, but way more than retrievable using the public dR API. This system uses the user website as source.
- It doesn't show rant_id or comment_id and it won't to prevent abuse. Later today, there will be away to attach rants comments.
- maximum 2500 record limit. But soon you can get comments for every rant per user. You won't reach this limit in normal usage
Have fun with it! Don't worry about the abusing the API. Everything is allowed. It's fast as F. If it doesn't respond - it wasn't you. I work on it and often reboot services and it takes some time to recover to state.
If you're not familiar with python, that's ok. Check if you're a decent dev and have python or python3 on your computer. Just execute it and paste the lines. Other way is to save these three limes to a file ending with .py and execute python3 [your-file].
Another example for people not used to python:
from pprint import pprint as pp # nice printing of values
client = ServerProxy("https://victoria.molodetz.nl/rpc")
comments = client.get_comments({'_limit':1337})
for comment in comments:
if comment.get('username','default username') == 'kiki':
print(comment.get('body'))
pp(comment)
Happy hacking!5 -
I have a score difference of two in two posts after eachother on the same page. This is how it must feels to win a lottery without gaining smth. I should buy a ticket. The odds are with me.11
-
Black box. It does seem to put messages with an URL in a certain category though, but also that's not always correct. It's trained on 3000 normal dR messages, and 3000 spam dR messages. 6000 dR messages in total. Many epochs but not good for use yet. The idea that the system could classify without discriminating new users is from the table. That discrimination is needed as a safe margin. Original spam system is a bit simple, but it doesn't do false positive and works great. Still, I want to make smth advanced out of it for the sake of education. Tomorrow I'll have my neural networks book. Probably over two weeks I have some good insights how to improve this all. New hobby :)
(pretrained 3b models are fine for recognizing spam btw. But it costs resource. 8 CPU's 100%. A self trained model pure on spam doesn't and is fast. With a pretrained model you can't do mass classification.)7 -
Hello all,
Did a MASSIVE spam down vote action..
If you've watched the algo-list you kinda had an idea how much spam there was. In exception of rants, it also down vote the spam comments posted under regular rants. Also made a progress with historic rants.
I think I leave it at this for now, gonna do some machine learning for spam and will apply that as filter in the future.
Site is now clean enough now to not encounter spam just browsing the site.
Our new frenemy Buffon gave as tip the no index tag what is a great idea for users not having reputation < 5 or so. Will contact dfox if he's willing to do this small change.12 -
So, a while ago I made a keylogger (called tikker (project is on my site) and now I created plot with statistics. Interesting, I work literally full time (24h) :P So now I have a graph of activity per hour:14
-
Let's start a discussion about decentralized. EveRyOne caN hOsT hiS oWn ServEr. Do you mean the freaking internet in general? By definition, the internet is decentralized. "Decentralization has a protocol we all use to stay in sync". That existed already, it's called IP, TCP and UDP.. The decentralization protocols are on top of those making it only more limiting. Good, many nodes in sync. Yeah, replicating SQL servers exist for a long time.
People who 'invented' decentralized did just not realize how the internet works. Adding a network on top of a network ending up in a smaller network making it more centralized. "Decentralized" stuff has nothing to add. Just some word for replication protocol or smth.
I'm too sober to fall for this shit.14 -
Had a heart attack when I found out what I was holding and all wasps (or bees?) came out crawling slow and sleepii. Thank God I was wearing gloves. It's ventilation hole of my bathroom. So, that's where the fuckers stay in the winter. There must be a big fat queen in there. Covered the whole nest with shaving foam and hope they'll die. Scary fuckers.random bathroom bees holding a hive fuck !it scary foam covered heart attack wasps gloves winter sleepii18
-
Spammerboii...
We delete just all the comments of users created on a given date.. So even commenting on old rants like you do now doesn't help! Sad for effort.
Sad, really sad. All that effort for nothing. That for one site.36 -
Got my dad to try Fruity Loops so that he can compose with his Tyros 3 keyboard on the computer. If he likes it, I will buy producer edition (that one can also record, that's why) for his bday soon. For Christmas I already ordered midi to USB cable for him. Finally I know what to give to someone who already has everything. He needs stuff for his stuff :p
It will be much fun experimenting together with it. My dad can go very hard on new things, he always becomes a pro in it.
I hope it will be just as nice as the time we spent together on Autocad. He later even made Autocad exams for schools.
He learned me how to install windows 3.11 as very little kiddo.
I like my dad doing computer stuff. It's the moment we really bond.3