Heads up, for the `Backend Tracing` screenshot there's a typo, it reads "enviroment" with the missing 'n' and on the AI debugger page on narrower screens the bubble for the "Learn more about JamGPT" text doesn't fit the text.
I have to admit, I once agreed with this position, before I got to a ~5g a day habit. While I was a heavy user (6+ years) I was a strong proponent of legalization etc. "Better than booze", "It's only a gateway because it builds connections with people that have other stuff available".
But there is arguably a physical dependence at that routing usage. First thing every morning, every 1-2 hours another J. I wanted to quit almost every day. I'd have conversations with my partner about quitting every few days. But neither of us could. If we had none we wouldn't sleep for 48+ hours, just didn't happen. Anxiety sets in, stress about everything. You just want to sleep, so you smoke. You go broke you scour the house to find enough coins or bottle to return (EU here) to get .2g to roll once to just get to the next day. You leave the house to do anything and you're so anxious and paranoid you become misanthropic. Only way to deal with it is to go home and "relax" to manage the borderline panic. Just for reference, before I was maybe a bit awkward sometimes but had absolutely no issues with the outside world.
I also don't agree that afterwards you're in the clear. I used to be borderline photographic with my memory. I'd pull random statistics from papers I'd read years earlier in conversation with references that I could use to validate when challenged. I no longer can even remember what I decided to go to the grocery store for, even though it's a ~2 minute walk. I forget what I started a sentence to express while typing. I'm a systems engineer, when I change tabs in an IDE I lose nearly all the context I took with me to the new tab. Sure context switching breaks flow yadda yadda, but it's just different. I don't even remember there was something to remember, just... "why am I here again?"?
I do agree that the behaviour trigger is very strong, so I replaced the urge with League, an arguably far more self-destructive tendency.
Increase tolerances to reduce cost at scale, sometimes doesn't fit perfectly or requires a little jiggling to line up right.
But now you can't do the jiggle or try to manually quickly realign it, instead you need to call a tech to do it for X bucks and a wait time of longer than a few seconds.
Having your own address in most places is a part of a "dedicated business line". My ISP in Switzerland literally refuses to issue so called "static" addresses at all, business or not.
I miss the days of sending someone a letter with some cash for them to associate address A with line B. All I'd have to do to stay essentially anonymous is finding a someone with bad record keeping.
Suddenly someone shows up with address A and threats and then drowns trying to interpret that persons mappings. While that's happening I can find 5 other someones and suddenly I have 6 addresses all of which essentially ephemerally link to my system. Someone else does that for their mapping system and you get to Dijkstra levels of working out how to block connections.
After like 3 levels of middlemen even centralized authorities just struggle to do the actual work of blocking, outside of just issuing the order.
To me that was the real strength in IPv6. (I know I know innefficient protocol with complex upgrade path lead to near negligible adoption)
NAT "fixed" the problem of address exhaustion, but it killed the old internet. You cannot run your own network anymore. In the "old" times, I gave you a phone number or IP address and that's it, direct connection. All anyone could do was show up and take the computer to stop that. Sure there's a phone company or ISP involved, but they just powered the pump, you completely controlled what went through it.
Now I can't do that. They ran out of addresses and I share an address with X unknown others. So I can't give you a home address, just to a bank of doors. I could give you an apartment number, but that's also shifting transparently, so num X to you is num Y to someone else.
IPv6 would have solved the problem of exhaustion while preserving the right to an address. I could be some number permanently and you could reliably find a connection to my system using it. In that world I could set up a private DNS service in my house no one can alter without physically plugging in. Then have that store records to other addresses. Every part of that chain requires someone finding you and showing up at your door to disrupt.
Instead now I have to pay digital ocean 5 bucks to keep an address for me so anything can find me via them. A bunch of servers in my home effectively an island without a coordinate until DO points me out on request. Like having all mail addresses be to the local town hall for them to forward to me. Sure maybe you trust your local town hall, but they are fundamentally beholden to someone else.
With IPv6 support and adoption a whole network could be set up independent of any other authority besides BGP. Which requires nation-state levels of mobilization just to block an address, with fallout affecting literally thousands of others. They'd have to nuke a block to suppress any site, only for that site to find another address and be back to normal within minutes. Instead they do a WHOIS, send a scary email and boom, you're unknown, unfindable and disconnected. Hoping that word of mouth brings people to your new "address" exactly like losing your phone (and SIM) while abroad.
I know it sucks as a protocol but v6 to me is a massive extremely important development that would change how the internet, and from that all communication, works.
Can you explain how you'd use Webtorrent to synchronize a large dataset that's updated in realtime? If you mean to get a P2P transport wouldn't WebRTC be what you're aiming for?
I'm genuinely curious but isn't Webtorrent just using WebRTC to join a Torrent Swarm? Torrents are fundamentally immutable, the identifier is a static hash of the content of the torrent. That would mean producing a new torrent for each new data point or chunk of data points only to then submit that hash to a WebRTC based connection to again fetch torrent content?
Genuinely curious, I'm interested in how torrent swarms can be used for novell applications.
Heads up, for the `Backend Tracing` screenshot there's a typo, it reads "enviroment" with the missing 'n' and on the AI debugger page on narrower screens the bubble for the "Learn more about JamGPT" text doesn't fit the text.