09:19:05 An idea to transform the monero network to become something more powerful and counter malicious dump from authorities: Monero already has a lot of nodes node, these nodes can be transformed to become like onion router node in the same time, so monero has its own onion network 09:19:56 All the nodes report their stats to monero blockchain so no central directory is required like the tor network. This does not require the edit the chain protocol since we can just do a dummy transaction for storing the info on chain. 09:21:01 For making you node a valid onion router recognized by the network, you need to lock certain amount of monero, this prevents the sybil attack and correlation attack since it becomes costly to spin up a huge amount of nodes. This needs a little bit of change on the chain protocol code, but just a very simple lock and unlock function so no safety problem introduced. 09:21:10 For making your node a valid onion router recognized by the network, you need to lock certain amount of monero, this prevents the sybil attack and correlation attack since it becomes costly to spin up a huge amount of nodes. This needs a little bit of change on the chain protocol code, but just a very simple lock and unlock function so no safety problem introduced. 09:21:40 This kind of staking reward also bring some sort of buying pressure of monero, so hopefully making monero not a coin really failing on fiat value 09:24:00 And we do not edit the network to print money for nodes, instead we fork tor browser and make two changes, First let it to use our new onion network, our userbase will still enjoy the same browser fingerprinting as tor browsers do, and allow nodes to display ads. We we setup up a ads auction everyday, and anyone can pay monero to let their ads get displayed 09:24:37 And the money will be sent to nodes based on a formula combined their uptime, stability, and bandwidth as the reward for locking monero and providing onion network 09:26:39 And we add the monerod client alongside the tor browser, so users who want to use the network must prove they submitted a certain percentage of their local cpu for doing p2p mining. This also strengthen the mining network, making it consists of more legit users instead of hacker botnet, and beats the centralized mining pools 09:28:46 We can also edited the monero blockchain, so we can have domain function based on that, this is the biggest fix for tor network, since we will not have notorious human unfriendly and phisher happy onion address, we will have some cool .monero domain function in or monero onion browser. 09:30:09 In the same time the money auctioned for domain address can be added back to a grand pool for monero future development, pay more professional devs, spin up more onion nodes to combat network montior, and have more legal resources to combat unfriendly regulatory environment 09:33:21 And last there are many websites basically blocks any onion network exit nodes, so making tor users unable to access most sensitive services, we can also do some snowflake like exit nodes, basically we allow high risk runners to get their phone, insert a sim card, and transmit exit traffic, basically this is like what residential proxies do. As a result our users get low risk exit 09:33:21 ip instead of high risk data center ip. High risk runners can simply get a lot of burner phones and sim cards, and change them everyday, this makes ip based discrimination against onion network users hard or almost impossible 09:35:04 And this is optional, if users want to keep their p2p mining rewards their exit ip is some high risk score datacenter exit node, like what tor users do, but if they want to enjoy some smooth surfing, they need to contribute their mining rewards to the exit node rewards pool, and it is like the onion node reward process, operators get rewarded based on their stability, uptime, and bandwidth 09:35:47 An idea to transform the monero network to become something more powerful and counter malicious dump from authorities: Monero already has a lot of nodes, these nodes can be transformed to become like onion router node in the same time, so monero has its own onion network 09:37:25 Also can make a compatibility layer so our monero onion browser can also access tor network in the same time, and we can expand our domain service to the tor network, so those darknet market will also spend a lot of money trying to setup their domain name on monero chain 09:37:53 And I believe this is going to work great since consider darknet, tor browser and monero users highly correlates 09:39:43 And we do not edit the network to print money for nodes, instead we fork tor browser and make two changes, First let it to use our new onion network, our userbase will still enjoy the same browser fingerprinting as tor browsers do, and allow nodes to display ads. We we setup up a ads auction everyday, and anyone can pay monero to let their ads get displayed in user's browser based 09:39:43 on a probability decided by their contribution in the auction, it is like we have an annoying banner in every user's browser. It is open source and anyone determined enough can just compile it out manually, but we can make this process slightly complicated so most user stick to default setting and view the ad as intended. 09:40:27 We can also edited the monero blockchain, so we can have domain function based on that, this is the biggest fix for tor network, since we will not have notorious human unfriendly and phisher happy onion address, we will have some cool .monero domain function in our monero onion browser. 09:41:40 And this is optional, if users want to keep their p2p mining rewards for themselves, their exit ip is going to be some high risk score datacenter exit node, like what tor users do, but if they want to enjoy some smooth surfing, they need to contribute their mining rewards to the exit node rewards pool, and it is like the onion node reward process, operators get rewarded based on t 09:41:41 heir stability, uptime, and bandwidth 09:43:54 Finally more services can be provided by onion network nodes, we can add them as plugin function in the main gui wallet, for example reputable long lasting onion nodes can get the right to provide storing file as a service, basically like what amazon buckets do, and of course through our own onion network 09:45:53 The file service can be either public which serves some website like a cdn cache service, or private just like a webdrive for people(with some redundancy of course) 09:46:55 We can also mandate servers to use intel sgx or amd secure enclave during the hosting process, so they acquire more legal plausible deniability, when someone do some pedophile content or terrorism content server admin can simply say due to secure enclave we are unable to know what is happening inside the processor 09:47:12 Of course all these services are going to be exchanged through monero 09:48:07 It is like making monero some real currency instead of other crypto hype which pump and dump alongside btc, when a currency is tied to more real goods and services it naturally stablize 09:49:56 We can also have something like password recovery as a service in the portal where people with degen powerful gpu can make money on monero, when the eth pow good old days are gone gpu can still be something profitable 09:51:56 These are just some thoughts from me, critics welcomed 09:53:52 And we add the monerod client alongside the tor browser, so users who want to use the network must prove they submitted a certain percentage of their local cpu for doing p2p mining. This also strengthen the mining network, making it consists of more legit users instead of hacker botnet, and beats the centralized mining pools. This also makes ddos against monero onion network and nodes harder 09:57:28 Also domain can also work for personal users address, so we can have joe.monero to pay for and easy to remember(and more suitable to use in real day life! Like pay for grocery or weed when you meet up!) 10:01:09 Also the file services and accumulated monero can become a combo to do something benefit human race as a whole, for example a grand online library projects, we can store all the books human race have distributed in the monero blockchain, and we can put rewards or even hire people working as a profession to digitalize(maybe illegally) or the books they can find. Also for rare books 10:01:09 special bounty can be setup for people who find and digitalize them. So eventually we will turn out having the largest online and censorship immune shadow library where every school kid and scholar knows, with monero logo on it of course, and they need to do some monero p2p mining everytime they want to search/download something pretty much like how eth gas work, but instead of c 10:01:11 hanging state of evm you are searching our grand knowledge base 10:01:30 Also the file services and accumulated monero can become a combo to do something benefit human race as a whole, for example a grand online library projects, we can store all the books human race have distributed in the monero blockchain, and we can put rewards or even hire people working as a profession to digitalize(maybe illegally) all the books they can find. Also for rare book 10:01:31 s special bounty can be setup for people who find and digitalize them. So eventually we will turn out having the largest online and censorship immune shadow library where every school kid and scholar knows, with monero logo on it of course, and they need to do some monero p2p mining everytime they want to search/download something pretty much like how eth gas work, but instead of 10:01:33 changing state of evm you are searching our grand knowledge base 10:02:45 Maybe too much talk and too less action, better gear up myself right now for doing something first 12:47:55 Distributed AI leveraging Monero signatures could be a more efficient option. Just a suggestion—miners already have their GPUs ready to run neural models. What’s needed is distributed data with privacy, similar to how Matrix secures our chats. 🙂 12:55:45 Distributed AI leveraging Monero signatures could be a more efficient option. Just a suggestion—miners already have their GPUs ready to run neural models. What’s needed is distributed data with privacy, similar to how Matrix secures our chats 🙂 . Users can seamlessly sign with their wallets and request a specific book from the model. 14:57:29 Distributed AI sounds like something not very applicable, does this even exists? 14:59:04 Like if you simply run a model on remote computer this is going to be very sketchy, since AI can be used on something very sensitive especially LLM, and there on numerous unfixable local attack method on secure enclave, so nothing stop the computer owner from really peak inside what people are doing 15:01:00 And if it is something much more complicated like real distributed computation, like you formally verify something with haskell, and you just send everything to different nodes, and wait for them to come back, because it is purely logically programmed and formally verified, so you can expect it comes back in random order but still works correctly, but does thing like this even exi 15:01:01 sted for AI? This is like two very very different fields 15:01:22 And if it is something much more complicated, like real distributed computation, you formally verify something with haskell, and you just send everything to different nodes, and wait for them to come back, because it is purely logically programmed and formally verified, so you can expect it comes back in random order but still works correctly, but does thing like this even existed 15:01:23 for AI? This is like two very very different fields 15:01:34 And if it is something much more complicated, like real distributed computation, you formally verify something with haskell, and you just send everything to different nodes, and wait for them to come back, because it is purely functionally programmed and formally verified, so you can expect it comes back in random order but still works correctly, but does thing like this even exis 15:01:35 ted for AI? This is like two very very different fields 15:02:00 And plus what if some nodes make a mistake, doesn't seems working for AI on a high latency network for me 15:54:50 well @not_a_money_printer there's stuff like this https://github.com/b4rtaz/distributed-llama not saying what karuga said it's a good idea but the answer is yes it's possible. May not be ideal in some cases. Definitely hard and/or impossible to verify the work but my 2¢ 16:37:48 not_a_money_printer: From another perspective, I was referring to what m-relay just shared. We're conducting research on an on-chain security mechanism powered by AI. We've developed a model that verifies access authority by checking for an existing on-chain signature before allowing prompts. The training data is securely stored in distributed storage. Notably, the model only proc 16:37:49 esses prompts, querying the distributed data without modifying it, and returns a response. I considered the library idea within this context. 22:25:22 Test