It would be somewhat freeing, I imagine. Though, if you were to go hardcore, quick access to knowledge would be the biggest drawback.

Ditching the enshittified parts would probably be easier. Can do offline maps (or paper), purge social media, and use credible sites for any information you might need to look up.

I support and enjoy Nebula. Definitely worth the cost in my opinion.

Never thought I’d see an MMM article on here, but it’s definitely relevant.

I recently got a clear big ice maker and read into directional freezing. It’s pretty cool (pun intended) and nice that little waste and if something doesn’t work, you can just try again right away.

I feel like this would be an appropriate place to monologue a bit.

As in plumbing vents or piezometers?

TBH I just did a quick web search on my phone and picked the first one. Apologies.

I use Ultimate Updater connected to gotify to tell me which machines have updates with a noti every day. I can then run update and it will take snapshots and upgrade everything when needed.

It really only does linux systems and wont do docker, but there’s watchtower (the one that’s still alive) to do that.

Rethink DNS

LXC on the host for me. Make sure it never backs itself up though.

Arguments like, “Well, why do you have clothes on then?” are not effective, because they aren’t equivalent forms of privacy.

There are a couple pieces of media I like to recommend:

• Targeted by Brittany Kaiser - Kaiser being intimately involved in Cambridge Analytica, it outlines how digital record of user behaviour, traits, and engagement can inform agents to which form of manipulation would be most effective on a given person (psychographic analysis), which can then be used by whomever wanted to exert influence (micro-targeting), whether it be on consumer or political preferences.
• The Great Hack is a documentary adaptation of the above story if someone prefers film over text, but the book has more detail of the methods and examples of when it’s been used in the past (alarmingly many political campaigns).
• The Social Dilemma is another documentary that touches on how exploitation of user data drives issues like addiction, radicalization, and depression on social platforms. Just recently, Meta was found to be feeding increased beauty advertising to girls and women who had recently deleted selfies.

Providing real examples of this exploitation is, in my opinion, a more effective argument for promoting online privacy. It nudges people to think, “maybe it would be better if advertising companies didn’t know about my recent (breakup, miscarriage, job loss, promotion, unplanned pregnancy, debt, car accident, birth of a baby, death in the family, deletion of a selfie…).”

No need to apologize, but it’s good to be aware of the policies of various applications you use and promote.

I use perplexity occasionally myself, aware of the above, so I only do more general searches with it.

I have a self-hosted Perplexica instance I use for anything more sensitive.

Since we’re on c/privacy; from perplexity itself:

What Does Perplexity Do With User Information?

Perplexity:

Collects: Search history, queries, device and location data, browsing activity, and navigational behavior (especially via its new AI-powered browser).

Uses: These data points help personalize results, train their models, improve functionality, and—crucially—build detailed user profiles for targeted ads and marketing.

Potential Risks: Privacy experts warn this data collection may turn users into marketing profiles, similar to surveillance practices seen in other big tech companies. Even actions outside the Perplexity app (via their browser) may be tracked and leveraged.

Transparency and Privacy: Perplexity does not offer strong privacy protections (like end-to-end encryption), and isn’t fully transparent about how all user information is used. Cookies, device fingerprinting, and web beacons may track even non-logged-in users.

Enterprise risks: Businesses using Perplexity’s enterprise tools must be cautious about uploading sensitive data, as it could be used for model training and not always protected from leaks.

Exactly as designed…

Fake it til you make it.

If they see no consequences to their actions, it only makes them bolder to go more extreme.

Can you elaborate?

Anyone with experience, how does this compare to Dawarich?

Extending history by a lot isn’t recommended. A database add-on is recommended for long term storage.

Traffic data kind of requires driver tracking data to be sent to the cloud, which usually goes against the philosophy of projects like this. Unless they’re obtaining it from a third party, I doubt it’ll be a feature they implement.

I agree, the assignments will need to adapt to discourage the use of LLMs. Easiest is in-class writing or written exams. Unfortunately that takes away from other class activities.

I remember one of my favourite courses in university had exams where you could bring in any resource you wanted (excluding phones), because the exam was written in a way that required understanding of the core topic, something you can’t simply look up.