<monerobull:matrix.org> where are we on checkpoints?

<monerobull:matrix.org> if monero now gets more attention because of the fiat ATH, someone might attempt a very deep reorg to discredit it

still blocked on the same bug within the monero core, no one has worked on it further much (either working on rust or FCMP++ stressnet, or fixing other old code smell so we can scale to 1 GiB blocks)

we know how to repro it, but not an exact repro case (doesn't trigger all the times the steps are followed)

<monerobull:matrix.org> great

<monerobull:matrix.org> so we are at the mercy of the powers that want to wipe us out

the rest of programs was tested and run on testnet, which is where the issue was identified

<bikrambiswas:matrix.org> > <plowsof> bikrambiswas, i misspoke / tried to enforce a double standard : we can talk about our AI slop projects in MRL rooms, my apologies, you will not be removed, the CCS sponsors vibe coding projects if you are interested

<bikrambiswas:matrix.org> Can anyone tell me more about the process in short. I am not an experienced programmer..I know basic stuff..But surely I can give you far better output than this AI generated " Hey Kyle" thing.

<bikrambiswas:matrix.org> How to participate in Vibe Coding things ?

One research related vibe coding CCS for reference is ccs.getmonero.org/proposals/gingeropolous-netsim.html (although this has a high bar of entry due to the hardware required) perhaps MRL could benefit from some other vibe coded projects? Follow meetings and see if any opportunities present themselves

<boog900> uhh LIONLINK has some nodes again: moneronet.info

<boog900> not many though, at least when that scan was done.

<redsh4de:matrix.org> DataHoarder: is there a github issue regarding the bug?

it's commented as part of monero-project/monero #10075 but mostly talked about here in lounge and MRL

<boog900> @boog900: they are also not showing the proxy fingerprint

<bikrambiswas:matrix.org> plowsof: Well I can build this type of simulation tool...But i used free services..with that hardly I can deliver a basic simulator like a prototype. I hope you guys know..to deliver any successful output.. initial funding is essential. So ge gets funds for prototype ( POC) or after the 100 % delivery of product?

<ofrnxmr> He did not receive any payouts

<bikrambiswas:matrix.org> But if someone can deliver a simulator with low funding..around $8K..it makes no sense to judge them based on how they accomplished it, whether they used AI assistance or not. I see many people in crypto acting like super geniuses, while behind the scenes they secretly outsource work to Indian or Filipino developers for $100 p [... too long, see mrelay.p2pool.observer/e/9fe33dsKUUJ6b3R1 ]

<ofrnxmr> Nonsense

<jwinterm:matrix.org> Many people?

<jwinterm:matrix.org> I mean, we all know Satoshi paid some guy on Fiverr to write the whole white paper and codebase, but do you some specific examples of people you're referring to @bikrambiswas:matrix.org: ?

<bikrambiswas:matrix.org> @jwinterm:matrix.org: You think satoshi is a guy ? Not agency? Well good to know that . I guess you know what happened with the Bitcoin inflationary bugs.

<bikrambiswas:matrix.org> I believe 99 % crypto is a scam.. including the Top 10. It's a fake bubble.. running behind 1 narrative financial freedom...but in reality they are just setting up a trap.

<jwinterm:matrix.org> So...that's a no I guess, you're not going to accuse anyone specifically just setting up strawmen then?

LL

<bikrambiswas:matrix.org> @jwinterm:matrix.org: I’m not accusing anyone. My point is simple: building a working product on $8K is an achievement, and we should judge results and transparency rather than gossip about how it was built.

<bikrambiswas:matrix.org> You can detect AI work in 2026..it will become impossible for people to detect Automation work after 3 to 4 years... Judge the quality of work, and prototype and POC .... But fighting against the people who are using AI tools is nothing but pure waste of time..No matter what they do " Hey Kyle" is never gonna achieve the quali [... too long, see mrelay.p2pool.observer/e/mJ-f39sKUWN1UEhY ]

<jwinterm:matrix.org> you make a lot of categorical statements wholly unrelated to monero research bubba...

The point about AI is that it's a low effort to create and high effort to review, and it's likely there's something totally broken in non obvious ways somewhere.

So trying to judge it by its merits is a DoS on the reviewers.

Now, you could make the argument that something written manually by some coder is also likely to have something totally broken somewhere, but at least it's not a potential deluge of such things.

Once LLMs can do better, it becomes easier to avoid rejecting outright for this.

However, you then start getting into the problem of people instructing the LLM to purposefully do something subtly broken, and you end up in the same place.

It's like spam or ads. Receiving an email from a company is not bad per se. Receiving loads from many is shitty.

It's maybe unfair for the person who honestly used a LLM and ended up producing good stuff, but it is unsustainable for the other side, because you are not the only one who'd submit something and we can't tell the chaff from the wheat at a glance.

Now, I do agree there are some with an ideological rejection of LLMs.

<ofrnxmr:xmr.mx> using AI isnt a sin in and of itself anymore than searching google flr answers. The problem is that probably 99% of people using AI cant review the code that they are submitting

<ofrnxmr:xmr.mx> if the "author" cant even see the trash slop, then it makes more sense to throw away the baby with the bathwater than to comb through 50k lines of prompt slop

<ofrnxmr:xmr.mx> Its no better than plagiarism where you dont even understand the work that youve stolen

<bikrambiswas:matrix.org> moneromooo: I agree..But people generate codes using AI ..But a successful final output is always hard to deliver.

<ofrnxmr:xmr.mx> @ofrnxmr:xmr.mx: And that 99% of people using AI just yeet stuff at the wall until something sticks

<bikrambiswas:matrix.org> @ofrnxmr:xmr.mx: You can create new things, visualize new ideas, and innovate .. then convert those ideas into code. Plagiarism is a completely different issue. In science, every innovation comes from building on existing work and theories; plagiarism is simply copying and pasting someone else’s work. All I’m saying is [... too long, see mrelay.p2pool.observer/e/lID539sKa2Y2ZmUw ]

<ofrnxmr:xmr.mx> Its entirely vibe coded through trial and error

<ofrnxmr:xmr.mx> its not even copy-pasted, the ai is making commits directly

<gingeropolous> yeah, i've seen other vibe coded stuff be presented as "job done", like the i2p gui or whatever. that seems different than what im doin.

<ofrnxmr:xmr.mx> @gingeropolous: Like damn 20 of them

<ofrnxmr:xmr.mx> 20000 lines of absolute garbage

<ofrnxmr:xmr.mx> monero-project/monero-gui #4445 monero-project/monero-gui #4513 examples

<gingeropolous> but yeah i don't have a good answer for the human reviewer ddos for ai-generated code.

<gingeropolous> but hey, fixing memory leaks is fixing memory leaks: moneroexamples/onion-monero-blockchain-explorer #341

<bikrambiswas:matrix.org> moneromooo: But that’s exactly why the answer shouldn’t be a blanket rejection, it's unfair to throw away useful work because we are judging the creator (AI creation is easier). We can innovate in this problematic area... We have to figure out a better way to verify and credit contributors ... that can protect reviewers [... too long, see mrelay.p2pool.observer/e/67jV4NsKVUtteFVO ]

You have not understood what I said. Please re-read.

<bikrambiswas:matrix.org> I am not good at English..I hope you understand my point. This place is not for Convo. I rest my case .

Either that, or your definition of fairness is not like mine I guess.

based monero devs rejecting AI slop

<gingeropolous> its a tool. all tools can be used improperly.

<bikrambiswas:matrix.org> True, Private crypto can be used saving life in Gaza, ukraine, Africa..also it can be used for terror funding.. Nuclear energy is actually 1 of the best solutions for the energy crisis..but people created bombs.

<ofrnxmr> @gingeropolous: Not sure what normal use looks like, but mine is using 3gb after like 2weeks uptime

<gingeropolous> well, there's normal usage vs. being a highly trafficked explorer. i dunno how many users you get

<ofrnxmr> @bikrambiswas:matrix.org: "discourage new people" absolutely nobody wants the next generation of developers and maintainers to not have clue what theyre doing

<ofrnxmr> @gingeropolous: Its the stressnet explorer, so not sure

<ofrnxmr> @ofrnxmr: LLMs have to learn from somewhere, and learning from themselves is like photocopying a photocopy. It gets worse.

<gingeropolous> yeah this gave me 2026 as output: netstat -anp | grep :443 | grep ESTABLISHED | wc -l .

<rucknium> @boog900: @boog900:monero.social: I'm looking into this. In the latest date (2026-01-11), all of the LionLink IPs in good_peers.txt have the same peer_id:

<rucknium> mrelay.p2pool.observer/p/3e734tsKOUJFeHh6/1.txt (code snippet, 5 lines)

<rucknium> But I only have those 5 IP addresses in good_peers.txt, instead of the full 20 in the main database. And they don't appear in bad_peers.txt. Maybe the others in LionLink don't respond to the pings after the handshake.

<rucknium> I'm running the simplified crawler now to investigate that hypothesis.

<rucknium> @boog900:monero.social: In the README of github.com/Boog900/monero-ban-list , maybe provide a link to github.com/Boog900/p2p-proxy-checker so readers know what the criteria is.

<rucknium> @rucknium: It doesn't see to be a general failure of the network crawler. There are only 44 total peers that appear in the handshake database that do not appear in bad_peers.txt nor good_peers.txt, out of 12555 total peers (note that this counts peers on different ports of the same IP address as distinct peers). So the LionLink peers have much greater ping failure rate than average.

<boog900> I'm not sure I understand, does your scanning add them to the database then checks if they are proxies and adds them to the file? > <@rucknium> But I only have those 5 IP addresses in good_peers.txt, instead of the full 20 in the main database. And they don't appear in bad_peers.txt. Maybe the others in LionLink don't respond to the pings after the handshake.

<boog900> your scanner*

<rucknium> @boog900:monero.social: Yes: github.com/Rucknium/xmrnetscan/blob…main/src/rust/src/main.rs#L301-L385

<rucknium> The data from the initial handshake creates the database record.

<boog900> ooo if they are disconnecting if they detect the spy check that would be interesting

<rucknium> These current LionLink nodes only use one port, 18080. Before, most LionLink nodes used multiple ports.

<rucknium> I will check shared peer lists.

<rucknium> Almost all honest nodes are sharing zero LionLink IP addresses in their Levin handshake peer lists. Twelve are sharing one LionLink address in their peer lists and about 25 have a range of 5 to 98 of the LionLink addresses in their peer lists.

<rucknium> The spy nodes are sharing a lot of the LionLink IP addresses in their peer lists. The majority of them have 50-100 LionLink IP addresses in their shared peer lists.

<rucknium> Probably the current spy nodes are just sharing part of the old list. Because the admins are lazy perhaps.

<rucknium> Do you agree, @boog900:monero.social , that if many spy nodes were successfuly hiding from our scans, they would still appear in peer lists of honest nodes, so they would be detectable indirectly that way?

<boog900> yes although FWIW someone has injected unreachable nodes into the network in the past

<intr:unredacted.org> where do the names LionLink and Spruce Creek come from?

<boog900> @intr:unredacted.org: the names of the AS that the IPs are apart of.

<intr:unredacted.org> I see, thank you.

<rucknium> @intr:unredacted.org: You can read b10c.me/observations/06-linkinglion

<rucknium> This is the "response" from LionLink: linkinglion.net

<intr:unredacted.org> Thank you! I can't find anything when looking up spruce creek however

<rucknium> Spruce Creek was registered about a year ago. By the way, @boog900:monero.social , how did you find the date of Spruce Creek registration? Is it on a website?

<boog900> ipinfo.io/AS401476

<rucknium> Probably Spruce Creek was created specifically for surveillance. @ravfx:xmr.mx also found that Tor & I2P node surveillance switched from LionLink to Spruce Creek recently.

<ravfx:xmr.mx> In the last 13 days, I dropped 7 packets from the old LionLink range (aimed at my 18080)

<ravfx:xmr.mx> In the same period I dropped 19903 packets from the new Spruce Creek range

<jeffro256> Spruce Creek is hilariously vague to me for some reason. Its like filling out your name as Mr. Adult Man on a form

<jeffro256> Legitimate business LLC

<rucknium> Groom Lake 👽️

<ravfx:xmr.mx> Some AS resolver site say "Private Customer" fro AS401476

<boog900> @jeffro256: I like LionLink's Tor daemon names: metrics.torproject.org/rs.html#search/as:54098%20

<jeffro256> Perhaps our fav spies are changing their AS and name not to avoid detection on the network, but maybe they made an OPSEC mistake with their filings that would allow someone who was digging in to tie the LionLink operation with a more well known organization

<sgp_> "The data tsunami is coming. Make sure your queries are ready."

<sgp_> "Optimizing BigQuery performance lets us scale to meet this demand without breaking the bank."

<sgp_> trmlabs.com/resources/blog/performance-hints-for-bigquery

<sgp_> Re blockchain analytics scaling

<ravfx:xmr.mx> 2607:7700... is from US (8 connections)

<ravfx:xmr.mx> 2604:5580:21:: are also sus (Canada) (11 coonnections)

<ravfx:xmr.mx> My software don't show any ASN for theses, and only theses

<ravfx:xmr.mx> [... more lines follow, see mrelay.p2pool.observer/e/1cL25dsKUy1nRnJD ]

<boog900> its this address: 38.38.., and it's still being done, I can't connect to these nodes. > <@boog900> yes although FWIW someone has injected unreachable nodes into the network in the past

<boog900> ah matrix formatting: 38.38.#.#

<boog900> its not that whole range but its a few /24 subnets