I also recently got into selfhosting LLM. Having an AMD card meant I had to scourge for solutions since everything expects to have CUDA suppport which means having Nvidia cards. Koboldcpp has a fork with ROCM support which works on my machine, so I’m content with that for now.

I shudder imagining that a stupid company could make .cum domains actually happen. Looking at you Google and your .zip trolling.

Controversial contents are often chased out by their providers frequently, so that might explain the number of dead instances. They simply go for the next new domain in a loose jurisdiction each time it happens.

I have two separate keepass containers, one for passwords and another for TOTP.

There has to be a cultural shift as well. It’s not the early 2000s anymore where a substantial portion of internet users could tinker around their desktop computers. I recently got fiber at home and we’re locked behind CGNAT. I could look for a solution for myself since I grew up opening ports on my router, but imagine someone who grew up with bubble-wrapped smartphones trying to navigate their way through that bs.

So many chunks of internet history is going straight to the bin thanks to this.

So that explains why it’s so cheap.

Porkbun has been fair to me. Recommended.

Bless their souls. I cringe whenever I have to open a Reddit page from browser. And I know that old.reddit.com won’t last forever.

I had my doubts reading that Ladybird browser announcement, but more and more I’m thinking that Mozilla is desperately chasing the gravy train that has long departed with their sugar daddy (google) laughing all the way to the horizon.

I just wish we never got to this point to begin with. We shouldn’t have trusted all our keys to a single cool startup in the early 2000s.

Very true. I’ve seen how politicians of some countries do a complete lap-dance whenever a FAANG company entertains the thought of building a datacenter in their territory.

Very unlikely to exist. Libreddit (that didn’t support user login to start with) was discontinued a while ago, and Reddit’s hostile stance on 3rd party clients in general means any project gaining prominence will get killed before they become popular. RedReader (an Android app) is only permitted to use its API for free because of public outcry due to accessibility issues involving blind users. A web frontend with user login support for Reddit will get hunted down by Reddit’s legal department if it ever reaches maturity.

Piggybacking on the comment. I also use syncthing to sync my keepass containers. Have you encountered duplication of database files (e.g. filename-sync-conflict-*), and if so, how have you solved them? I simply merge the files through KeepassXC when it happens.

Imagine yourself running a ChatGPT equivalent. Wouldn’t you be tempted to record all the weird and dangerous stuff people ask just in case law enforcement pays a visit? A person can be a privacy advocate and fight for the cause, but a company has bills to pay and investors to please firsthand.

Choice, huh? I can’t choose where the config files are stored unless I am willing to either dig into an obscure setting, modify the source code and recompile (repeat every time there’s an update), or contact the developer’s smug beard using smoke signals.

It’s almost funny considering how crucial that difference is in the field.

Better later than never.

I’m can’t believe I’m considering purchasing another GPU just so I don’t have to depend on OpenAI or anyone toying around with the models.

Absolutely. I love paper books and physical collectibles. The problem I was talking about is being forced to the old methods instead of using modern technologies, like only being able to send an application form by physical mail instead of online.