ploum1d ago

I added the followingt to ublock origin:
https://github.com/alvi-se/ai-ublock-blacklist

It is helpful when I’m searching for technical stuff. Technical contents are plagued with SEO spam and now slop.

So I make my search using @kagihq , I open in new tabs the 5 or 6 results that look useful. Out of those, one or two are usually blocked, which means I don’t even need to parse its content.

The bad news is that the remaining 4-5 tabs are often not useful either.

Best are StackOverflow posts from the 201* era.

GitHub - alvi-se/ai-ublock-blacklist: Websites I personally found that are completely generated by AI. Pull requests welcome.

Websites I personally found that are completely generated by AI. Pull requests welcome. - alvi-se/ai-ublock-blacklist

GitHub
Show threadploum1d ago

In the end, I realize that reading official documentation and the source code of the tool you want to use is the best way to go. There’s always that 5 minutes dread when it seems helpless, when you think you could simply find a blog post from someone explaining exactly what you need.

But, this era is no more. You need to either blindly follow an hallucinating tool not under you control or build a real knowledge of the tools you use, through sweat.

Show threadnathan1d ago
@ploum Code never lies. Comments often do 😛
Show threadAstroBoat

@nathan @ploum The code contains many indirections and abstractions giving you a false sence of understanding.

The code is the how, comments are the why. Both are critically importent to grasp how a piece of software work and how it should be maintained.

Mar 14 at 4:15pmWeb
Show threadAstroBoat1d ago
@nathan @ploum For example, I'm pretty competent at coding in JS/TS, but I cannot make or maintain a Gnome extension without the documentation. You have to know many abstractions, APIs, entrypoints, libraries provided by Gnome, integration boundaties with Gnome Shell, were to store settings or how to handle caching. Code don't lie but it doesn't tell you much.