last time i used those it didnt circumvent cloudflare blocking properly so guess they have some secret sauce here not sure

Krolden is taking issue that they can see the vpn traffic at all, as in they can see that youre using a vpn (re: ISP knows who you are) and can sell that info to third parties who will now use your identity associated with that IP address (because ISP sold them info that the VPN connection came from you)

theres ways around this of course and its actually somewhat easy to measure how big your fingerprint is. most sites of course wont have access to this paid info so reducing fingerprint is only useful to hide from google or cloudflare, and even then they wont know specifically what data you have, just what IPs you accessed. also since so many people use the same VPN it makes it very difficult to track you, you need to have a large fingerprint which can be mitigated by using linux (or with piracy, having a separate server that doesnt fuck with anything but torrents and so on)

ISPs can also tell if youre using i2p or tor so ultimately its just a masturbatory argument, of course nothing is perfect, we all know that the only way to be truly secure is to steal internet from someone else without them knowing who you are, and you also need to be sure they dont have cameras and you have a sophisticated setup with qubes + vpn + (maybe) tor depending on what you want to do. this is of course an absurd amount of security for looking at pig poop memes on hexbear

And now we’re getting back to the original point of this new secret protocol, it hides that its vpn traffic, at least to current systems. I’ve been able to visit sites that normally block me or limit me. From what I can tell it reduces speed by half or so but if your speed is already good its usable, whereas something like TOR drops it all to a crawl, like 300kb/s if youre lucky. Dial up shit. This new protocol implementation good for daily driver obfuscation

as to 2 i do use wireguard and so on on occasion but its not really a good way to obfuscate who you are to some random website or bypass a firewall (unless you own a server in china or something)

ive heard of i2p but afaik it has similar issues with slowness as tor, hosting your own exit relay from that list of isps is still going to result in excessive latency

im talking stuff like daily driver privacy, not ‘organizing a revolution’ privacy, which really shouldnt be done on the internet to begin with

If you read about how it was chosen it was elected by a community poll and they mentioned they didnt really verify anything. Yeah, ridiculous thing to do, but possible for a lib that doesnt know what bellingcat is, and they later apologized take it as you want, theyre obviously libs and the other alternatives arent any better, and are probably worse

you could always host your own mail of course but that has its own issues too, you obviously cant host your own vpn so the only option is tor which can be very messy to use regularly

Are there any other decent alternatives for mail that arent just ‘store in plaintext on our server’. Of course it will never be perfect but my biggest concern is always weird sysadmins and ransoms, theyre always the most likely threat to an average person

edit: I also feel like proton is fairly ‘normie’ and most people wont flinch at the email address

Yeah I edited to say that, if you are an edward snowden or something you really need to just not have any computers and sneak into libraries with a usb with qubes on it or something if its absolutely necessary to get on the internet

Theres a lot of reasons to use a vpn that arent just privacy related anyways, like I vpn into chinese servers to get around the firewall (it blocks outside requests or shows different results to westerners) so I can read specific chinese news articles. Proton is probably the most secure choice for that (vpn into hong kong seems to do this fine). Tor traffic seems to not yet be fully obfuscated a lot of sites block it its also slow as fuck

Fundamentally one company gets all your information whereas before it was all companies get all your info, its also been audited and shown to only store your ID in a visible way so

Of course if you’re using an OS or browser that has a ton of telemetry that can be used to track you, I’m personally not a privacy absolutist, I’m more concerned about weird sysadmins stealing nudes and knowing my IP or something than a nation state cracking down on me (which they can do anyways)

Synthetic data is basically a fancy way of saying ‘I’m properly formatting data and reinforcing the ai’s good outputs’. Rearranging words, fixing / adding tags, that sort of thing. This is generated with various tools that usually have an LLM or VLM plugged in, though some are as simple as a regex script.

This is accurate, though I am actually going to explain why. These big model companies (Google, ClosedAI, etc) parasitize the open-weights/open-source community that actually makes good Loras, fine tunes, and research papers. Consumer hardware simply hasn’t gotten good and cheap enough for very good fine tune training, and thats why this is all slowly petering out. In a couple of generations of consumer GPUs, which will be when we get consumer GPUs geared towards AI (re: super high VRAM counts of like 70gb+ for an affordable sub 700 usd cost), we might see another leap forward in this tech. Though I will say that this mostly pertains to LLMs, generative AI models like Stable Diffusion have a lot of tricks up their sleeves that can still be explored. Most of recent research and tweaking has been based around building a structure for the AI to build on, to sort of guide it rather than letting it take random stabs at things, in order to improve outputs. Some people have been doing things like hard coding color theory, framing a photograph, etc, and interpreting human language to trigger that hard code.

We’ve had statistical models like these since the 50s. Consumer hardware has always been the big materialist bottleneck, this is all powered by small research teams and hobbyist nerds. You can throw a ton of money at it and have a giant research team, but the performance you squeeze out of adding 400b more parameters to your 13b model or having a gigantic locked-down datacenter is going to be diminishing.

Also, synthetic data can be useful, people are hating on it in this thread but its a great way to reinforce good habits in the AI and interpret garbled code and speech that would otherwise confuse the AI. I sometimes feel like people just see something about ‘AI bad’ and upvote it and don’t try to understand it, where it is useful and where it is not, and so on.

gpt4all is the easiest way in, hands down, no contest

im at 30TB uploaded but i swapped setups last year and idk what it was at before. so 30tb 1 year