Nathael ⚔️

I hate that when I read some blog post from someone I don't know, I sometimes have a hunch that a part is AI generated and it makes me question everything else. Also hate the fact that it happened multiple times on blog post recommended by someone else

This is also a great blog post / interactive guide growingswe.com/blog/quadtrees quadtrees!

this series of 3 blog post is super cool jakelazaroff.com/words/an-inter… I love interactive blog posts, and I love learning new tech stuff, now I gotta find an excuse to use one of those...

Toook a bit longer than I thought because I made a few security changes to try and make sure this doesn't happen again, I think the culprit might have been my old ubuntu config since it wasn't updateable easily, but could have also been a few other things so I made a small pass

I wanted to finally finish the blog post on my board game logging app today but I got hit with a ticket from my hosting provider "your dropley may have been used in a DDOS attack, please delete it" so I've spent the evening recreating the droplet and re-upping the project!

Used a workaround for now, but I doubt it will hold very well, will use the API later on if I like my tool! I need to load up a few other destinations tho!

the X api is so confusing, and their errors aren't great, I get a 503 because I think on the free tier I can't post tweets ? (I say I think because the app I created is supposedly on a "free" tier but I have no clue what it can actually does without paying for credits)

took me 2 clicks to get the tokens to post on 🦋, worked first try, so at least this was easy!

I might load up a few $$ on the api for a few credits anyway cuz I don't plan on using a lot, but it's not a great DX

Great vid, hoping it helps people realize how widespread this already is. And ffs on the onlyfan portion, I can't believe someone would publicly say they work on that tech and scam people and just say "it is what it is" how can they watch themselves in the mirror man

sometimes I wonder what the heck I'm doing and I feel very very stupid, but then I read some tweets of ppl I consider highly successful and it's the most insane thing I've ever read and I feel a bit better. Still feel stupid and doesn't make me successful tho.

I'm cooking a new interface to explore your own games, not ready yet but I think I'm onto something fun!

got a rpg game to plan for saturday, got some very good ideas of the whole session, but didn't write down any prep yet, gonna be busy next two eveninnnnnng

@AmpCode Maybe someone can tell me, is the 14% from the baseline too high ? I have only a 40 lines agent md The initial reads seems to take a lot of context, this may be because my react files are too big and the agent spend tons of tokens to read them entirely ?

@AmpCode The change often works and all is well in those cases but if I want to iterate a tiny bit on it, I'm reaching a full context super fast and then have to do a handoff or start again to keep refactoring or updating a specific feature

sometimes i'm surprised at the rate the context of my @AmpCode agent fills up, I point at ~3 files, and explain a refactor I'd like to see, then the agent fills up the context to almost 40% before starting anything! By the end of the first agent response, I'm already at 50%!!

Ok fixed some things by hand in the frontend, but gotta call it for tonight and handle bigger refactor tomorrow with the agent

It went horribly wrong, I checked the first feature I asked the agent to build, and it wasn't bad, but it took a few iterations to complete it and now I'm out of credit and the first feature is still not ready 😄

I'm doing an insane thing which is launching an agent on my codebase to add a feature or rework something but then I don't have to check it and I get more credits so I launch another one and another one and another one. The tests before my next prod will be painful