Our paper "Unstoppable Wallets: Chain-assisted Threshold ECDSA and its Applications" was presented last month by Guy Zyskind enc/acc at AsiaCCS . The paper offers a lightweight threshold signature protocol (for ECDSA) under a more permissive trust model, assuming a non-colluding

sup fhreaks

Even when someone tries to acknowledge Secret Network leading the pack, it's derogatory gj nonetheless. Doing something before it's popular is 100x more impressive than jumping onto an already saturated narrative

Unrolling of a 1500 year old Persian Haggadah discovered in Iran. Still in great condition.

This is true because like Batman and Superman, MPC holds the secret to destroy FHE

1/ Get compensated to build the future of onchain confidentiality! We're taking the Fhenix ecosystem to the next level by providing builder incentives with the Fhenix Grant and Bounty program. fhenix.io/unlocking-ince…

After 10 years in the space I often get cynical. So it’s amazing to see vitalik.eth still fight for what crypto should become, not what it is. Today I feel better working in crypto, because of this 👇

Is there a nice curated tool to curate interrsting/relevant papers in Arxiv just like on eprint? There are still few enough papers on eprint such that I can check titles/abstracts once a week/month. Arxiv is too massive

I guess the $SCRT is out. Congrats to MakerDAO for merging with Secret Network! What a phenomenal milestone when the most decentralized stablecoin embeds privacy features for all

People old enough remember the ‘blockchain, not Bitcoin’ summer of 2015. ‘Large’ companies popped up like chain.com, R3, and Digital Asset - they all got stuck in the ‘pilot, no ARR’ loop and fizzled out (Stellar bought Chain). Web2 corps will come last.

This design choice will also be important for threshold encrypted transactions as in Fhenix (but elsewhere as well)

Rollups are centralized and that's a feature, not a bug. Decentralization/Security/State is inherited from the L1/DA layer.

TEE projects missed their chance to call it TeeLS

This is a great read. Glue computing = (general purpose, easy, slow) Coprocessor = (special purpose, fast) For FHE-land, glue is TFHE, but for coprocessing other things make sense. For example: MPC protocols are really good for signing; CKKS is much better for AI, etc..

.Secret Network built/proposed half of these years ago

This is pretty amazing if I read this correctly. Doing secure inference over FHE (=queries are encrypted), takes ~50s per token for Llama-3-8B on 4xA100s (eprint.iacr.org/2024/136.pdf). This is way faster than I expected, and while we're not there yet, this is incredible!