Zikeji

Roguelikes and roguelites tend to be my favorite. Ones where each run is new and you can toy with different builds and usually get pretty OP toward the end (or get cut down early because luck wasn't in your favor or you made a mistake).

A VPN is still a good choice, in fact if you setup your own VPN on a VPS that is an even safer choice because then you (sorta) control the certificate used for encryption. True, your hosting provider could still obtain that cert if they really wanted to, and they still have the data on your IP using it and for how long / how much, but it would make obtaining your data a targeted attack.

But there are cons to setting up your own, such as misconfiguration exposing you, or just the setup time in general.

A VPN introduces a new party who can harvest your data. It doesn't avoid IP tracking, it just shifts it from your ISP to another entity.

You have to trust that your VPN provider's claims of no logging/tracking are accurate, you can usually get fairly confident with research but it's never 100%.

Edit: to clarify, I'm not trying to dissuade VPN use. It's a still a great choice.

A not so funny story on a similar vein: roughly a week after my parent's friends of a couple decades agreed to be named as our guardians, the husband announced, on April 1st, he was getting a divorce.

It wasn't a terrible joke, he just chose the most inappropriate time to announce it.

He did get the raw end of that divorce in court though, the judge really didn't like him after he brought his mistress to court.

Thanks for the correction and clarification! I just assumed from the open-r1 post that they gave everything aside from the training data.

Open source isn't really applicable to LLM models IMO.

There is open weights (the model), and available training data, and other nuances.

They actually went a step further and provided a very thorough breakdown of the training process, which does mean others could similarly train models from scratch with their own training data. HuggingFace seems to be doing just that as well. https://huggingface.co/blog/open-r1

Edit: see the comment below by BakedCatboy for a more indepth explanation and correction of a misconception I've made

There's Filler Checker as well for those that don't have such a collection.

The thing that's crazy is that if I followed the 2 "best practices" of verifying the phone number + getting them to send an email to you from a legit domain, I would have been compromised.

Since when was "verifying the phone number" a best practice? Phone number spoofing is still a thing and trivial to do, which is why the best practice is to call back once you verify the phone number matches whatever the company lists (or, preferably, call their main number).

I work an odd schedule - two jobs, one WFH Sat-Tue from 8PM to 3AM, then a hybrid (2 days WFH) dayjob (Mo-Fr) from 10AM to 6PM. It's been this way on and off but so far I'm at over a year with this particular schedule, but I've had similar schedules in the past.

I would say I have a life, but my hobbies are more introverted anyway. Am I healthy? No, bit I wouldn't say that's entirely related to the schedule, I have other conditions.

It can be taxing at times but most of the time it's just life.

A combination of both. More bad mood going around, impacting you and causing you to perceive things more negatively.

There are quite a few text equivalents. text-generation-webui looks and feels like Automatic1111, and supports a few backends to run the LLMs. My personal favorite is open-webui for that look and feel, and then there is Silly Tavern for RP stuff.

For generation backends I prefer ollama due to how simple it is, but there are other options.

I'm not familiar with the drama itself, but dansup is the creator of both Pixelfed and Loops.video.