{"id":"BXHLGwpUJKI","title":"Illia Polosukhin on how NEAR AI proves your AI is private.","description":"\"The goal is how to enable private, verifiable and secure AI, for people and companies.\" Illia Polosukhin, NEAR AI.\n\nNEAR AI runs open-weight models \"fully end-to-end encrypted\", with verifiability.\n\nYou get certifications with hashes on the prompt and the response, plus signatures from the \"exact hardware that it was run on\".\n\nYou can check it all independently, even when you use it through other apps like Brave.\n\n#NEARProtocol #NEARAI #PrivateAI #VerifiableAI\n\nNEAR is the Blockchain for AI\nWebsite: https://near.org/\nNEAR AI: https://near.ai/\nIntents: https://near.org/intents\nChain Abstraction: https://www.proximity.dev/\nSharding: https://nearone.org/\n\nFor the devs\nGitHub: https://github.com/near \nDev Docs: https://docs.near.org/\nDevhub: https://devhub.near.page/ @NEARDevHub\n\nConnect with socials\nTwitter:   / nearprotocol  \nBlog: https://near.org/blog/\nMedium:   / nearprotocol  \n#NEAR #nearprotocol #NEARAI","channel_name":"NEAR Protocol","published_at":"2026-11-01T00:00:00.000Z","thumbnail_url":"https://img.youtube.com/vi/BXHLGwpUJKI/hqdefault.jpg","transcript_text":"for for near AI the goal is how do we enable private verifiable and secure AI right for people and companies and so at the core AI starts with inference and so that's kind of where we're spending most of our time enabling running open weight models in uh kind of fully end to end encrypted way with verifiability right like if you try near AI you actually get like literally certifications with hashes on the prompt on the response signature, you know, hardware signatures like each exact hardware that it was run on, right? So, you kind of get the full certification that you can validate and check independently. So, very blockchain blockchaines, right? We even actually just uh launching explorer-l like thing where if you're using it on other places like Brave, you can then go and double check that indeed everything is correct, right? So that's kind of like the fun fundamentals,","transcript_synced_at":"2026-11-03T04:15:46.542Z","segments":[{"start_seconds":0,"text":"for for near AI the goal is how do we enable private verifiable and secure AI","url":"https://www.youtube.com/watch?v=BXHLGwpUJKI&t=0s"},{"start_seconds":7,"text":"right for people and companies and so at the core AI starts with inference and so that's kind of where we're spending most","url":"https://www.youtube.com/watch?v=BXHLGwpUJKI&t=7s"},{"start_seconds":14,"text":"of our time enabling running open weight models in uh kind of fully end to end encrypted way with verifiability right","url":"https://www.youtube.com/watch?v=BXHLGwpUJKI&t=14s"},{"start_seconds":24,"text":"like if you try near AI you actually get like literally certifications with hashes on the prompt on the response","url":"https://www.youtube.com/watch?v=BXHLGwpUJKI&t=24s"},{"start_seconds":31,"text":"signature, you know, hardware signatures like each exact hardware that it was run on, right? So, you kind of get the full certification that you can validate and","url":"https://www.youtube.com/watch?v=BXHLGwpUJKI&t=31s"},{"start_seconds":39,"text":"check independently. So, very blockchain blockchaines, right? We even actually just uh launching explorer-l like thing","url":"https://www.youtube.com/watch?v=BXHLGwpUJKI&t=39s"},{"start_seconds":46,"text":"where if you're using it on other places like Brave, you can then go and double check that indeed everything is correct,","url":"https://www.youtube.com/watch?v=BXHLGwpUJKI&t=46s"},{"start_seconds":54,"text":"right? So that's kind of like the fun fundamentals,","url":"https://www.youtube.com/watch?v=BXHLGwpUJKI&t=54s"}],"youtube_url":"https://youtube.com/watch?v=BXHLGwpUJKI","url":"https://near-pedia.xyz/videos/BXHLGwpUJKI"}