its crazy to me that following up on a claim makes me evil. I'm still seeing <10% knots connections even with tor and clearnet+asmap on.
once I finish my bitcointap script I can collect data over a longer period of time to get a more accurate picture.
nostr:note1n0ee8efw2eyt0fyqqqa65sj9je89cuystf9ytvxwsupqrezr0yjqk53hdk
i still think a non-tor test would be good to, but it seems most knots nodes are over tor, which would hide any form of sybil attack. so there's efffectively no way to know if they are real or not.
I had to look up what a Sybil attack is. In the context of bitcoin, how is it different than me just setting up loads of nodes? Itβs not like there is a 1 person 1 node rule.
Sorry if this is a stupid question.
The real crime here is naming it bitcointap script when you ain't even playing with merkle trees.
to be fair there is a space in between
FIAT system has Cantillonaire
On BITCOIN there is an emergent group of Crypto-Cantillonaires emerging
OGβs, Podcasters an Influencers that are now running βFor Profitβ initiatives and whose incentives have shifted apart from the ones regular Hodlers and Pelbs have
Nothing to be afraid about this healthy tension, just be mindful that being and OG or wearing an Orange tie does not make them fair or holders of the truth
this uses ebpf btw, nix-based, might suit your purpose https://github.com/0xB10C/peer-observer
You claimed 2-3%.
Show your work.
please send me the post where I claimed that
i said it appears this way, I never said for sure. it seems most were because of tor (where sybil attacks can't be detected)
π€£
My question is How do you know that most of the TOR nodes are fake?
its impossible to know that, which is why tor wouldn't be a reliable data source. someone could create 20 tor endpoints in front of a single node relatively easily
Ok. So why did you say it appears to be 2-3% instead of 20%?
i was quoting the guy who said that from the data he was collecting. it took me time to write my script to replicate his findings, which I am running now.
Oh, ok.
Are you crawling it yourself or using snapshots from bitnodes
crawling myself
Cool. I was going to do it but it will take like 40 hours.
Whatβs the ETA on yours?
going to look into creating a visualization of all of the nodes i collect grouped by AS.
Emotional people being emotional
nostr:npub1xtscya34g58tk0z605fvr788k263gsu6cy9x0mhnm87echrgufzsevkk5s rides for the culture. Heβs definitely not evil. This coming from someone on the opposite side of the debate from him.
It seems to me you're being straightforward about your findings. I haven't done enough research on this debate to make up my mind and pick a side, but I do see all the names people are getting called and it is infuriating.
As far as I can see the actual knots nodes count is irrelevant & I see no way it can be used to attack Bitcoin.
A node is not a vote it's just a user's way to find the longest chain of blocks that meet their definition of Bitcoin. Calling it a sybil attack is deceptive at best.
I dislike where Core is moving & I dislike Knots. I dislike the division & the bickering.
Grow up. You're both intelligent but limited humans. Find a better enemy for your ego to squabble with.
nostr:nprofile1qqs0m40g76hqmwqhhc9hrk3qfxxpsp5k3k9xgk24nsjf7v305u6xffcpz4mhxue69uhhyetvv9ujuerpd46hxtnfduhs3mlty2 himself is very strict about how he counts nodes. He doesn't count tor nodes at all, he counts all cloud nodes as one. He has the knots node count at ~15% of the network.
Iβm not counting aws nodes or tor nodes, asmap on. iβm getting around 8%. But i only have like 300 samples. Will continue collecting
His problem is his insistance on using asmap. It doesn't make sense for this use case. Using asmap means everyone with Comcast as an ISP counts as 1 node. Everyone on Spectrum, just 1 node. It's ridiculous.
What a lower percentage with asmap would indicate is a lack of diversity among Knots nodes. Which would make sense, since the bad actors in Core stole the Transifex repo a year ago to prevent people from translating Knots... So non-English-speaking countries would understandably be underrepresented at this point.
