This is the rub, tech is able to track you based on your browser, viewport size, os, location (a vpn still has a location if you aren’t rotating) and more. I use Firefox for privacy and just that measure alone rules out 97% of internet traffic and zeros down who I am within 3%. How private am I if I default to that 3%. 1440p monitor and a half screen Firefox viewport? Now we’re building an advertising profile!

If targeted advertising is your main threat, then you are lucky to _currently_ live in a country whose government _currently_ does not consider you an enemy or potential enemy. Many people are not that lucky and many people will become unlucky despite not changing anything about themselves.

That’s what I do, the jetkvm has its “own” mini hdmi that follows it as I have moved it from different machines. Buying one cable wasn’t the end of the world (I was able to snag the jetkvm for $70 during its kickstarter and a $10 cord) and I use it constantly so it’s never collecting dust

I like the comparison of Meshtastic being the replacement for CB, it fits pretty well since they are both decentralized and made to run without a lot of underlying infrastructure and both work well to communicate in emergency’s.

I think this is the real problem. If it’s unsourced, how can I verify the LLM isn’t hallucinating. That being said I started running open web ui to host models locally and have heard that some will source their content(I don’t know which, I haven’t hosted them yet) so that is promising. I also like hosting deep seek locally and being able to review its logic process so I can assess how it arrived at its conclusions. All that to say, I still use a traditional search (self hosted version of searxng) for 95% of my search. I like llms for bouncing ideas around, but not for finding accurate results quickly