People usually have different tolerance level for pop-ups and distractions based on their neurobiology.

I have heard autistic people say pop-ups hurt and overstimulate them, ADHD people say they can’t ignore pop-ups, while neurotypocal people say they don’t even notice the pop-ups.

So both experiences can be true at once. You have no issues with pop-ups, while for someone else physically can’t tolerate a single distraction without stressing them out.

There might be, but very few. Most (like gpt-oss, or Deepseek R1) are open weight, because no one wants to make their model training data public.

Open sourcing a model would likely give away they are using pirated content or scrap internet websites without permission. That’s something Anthropic was fined for last year, they were caught using pirated books for model training.

Moreover, training data is one of most lucrative resources in AI race next to GPUs, and AI companies are going all in on finding it, desperately fighting for any possible data. Heck recently they even started buying antique books, digitizing, and burning them.

They invest a whole lot of resources to get quality data, and they don’t want to share it with competitors for free.

My personal opinion though is FOSS philosophy doesn’t apply that cleanly to LLMs anyway. If compiling Chromium or Linux kernel from source is tedious enough, “compiling” an LLM from source on consumer hardware is straight impossible. It’s something only your competitors could do.

Thus you can’t verify the LLM is the same as it’s source code anyway, so open source label becomes meaningless.

And open weight model is still an improvement over using a cloud model, as it’s local and doesn’t call home with your chats data.

According to the article, this guy’s legal team claims he did, in fact, tell them they should show a search warrant, and he won’t talk without a lawyer.

But they were ignoring those requests, and pressuring him further till he realized it’s a lost cause, and handed them over the Duress pin

Or both! Don’t consume where possible, seek out the most ethical option when you must

Jurisdiction doesn’t matter, it’s an myth spread by Protonmail marketers so everyone thinks their privacy is ensured by some strict Swiss privacy laws.

Although privacy laws are pretty mild everywhere. Even GDPR is quite outdated and is barely enforced.

Proton, Tuta and others do it because “Swiss privacy” or “EU-based” brand is much more recognizable and easier to sell people on, compared to “we do metadata obfuscation” or “here is our 50 page long security audit”.

However, what actually matters more is security architecture. Signal is US based, but they thought through how to retain as little data on users as possible, so when FBI asked them for info, Signal could only give them date of account registration, user ID, and last login.

TLDR ignore country of registration and HQ location, evaluate on the merits

NYM seems promising, they are trying to build a decentralized VPN with multiple nodes, same as TOR but faster, so no one has full data who talks to who and what type of traffic.

However it’s only been around for a year, so we still don’t have enough data if their approach works or not compared to traditional VPNs.

iVPN is a good one, it has been around for 15 years now, and is also privacy and FOSS focused like Mullvad.

However they lack some security features that Mullvad has: DEITA, diskless servers etc.

Ultimately I would say go with:

• NYM if you want to gamble (high risk high reward),

• Mullvad for good security, but yeah your money goes to a far-right party in Sweden,

• iVPN: worse security than Mullvad, but more trustworthy than NYM, and your money doesn’t fund right-wing populism

deleted by creator