Cryptoni0x
327 posts
@Cryptoni0x
Cypherpunk As reliable as MD5... Oh, you like crypto? Name every output of a sponge function. Cryptography Engineer @Nethermind
Digital assets are entering a new phase. What once ran in parallel to existing financial systems is increasingly being applied to solve practical, real-world needs — often behind the scenes – from cross-border remittances to B2B money transfers. This creates new opportunities to add value in how money moves globally. Today, we introduced the Mastercard Crypto Partner Program — a global initiative that brings together more than 85 crypto-native companies, payments providers, and financial institutions. Together, we're creating a forum for meaningful dialogue and collaboration as this space continues to mature.
AI watermarking schemes force a choice between public verification & private detection keys. Nethermind Research introduces zkWM by @Cryptoni0x (Antonio Larriba), using SNARKs to prove watermark presence/absence without revealing detection mechanisms. 305-455ms prove+verify.
This is a very interesting take from @dabit3 and I agree with him. "With AI APIs you're trusting that the provider is running the model they say they are and not messing with outputs, and there's no real way to check." Our @Nethermind Research team have been developing a zkML-based watermarking technique, with a working implementation. The approach enables cryptographic proofs that the model in use is the authentic, unmodified model, and it provides a verifiable method to detect and prove any tampering or unauthorized alterations. If this is of interest for the wider Crypto x AI projects and model builders, please reach out and I would be happy to schedule a meeting with my team to discuss further. There are many use cases for LLMs where this capability is especially valuable.
Obol and @Nethermind are joining forces to strengthen Ethereum. Introducing Pluto: a new client built on the DV Spec that brings client diversity to Distributed Validators.
Just published a simpler proof of the RBR soundness of FRI! Work with @0xAlbertG and Benedikt Wagner I also wrote a blog post explaining the high level ideas of the proof. All you need to know is how to colour a graph! Links to paper and blog below 1/3
