cryptography

Great thread on mpc-tls

Sometimes there are better solutions than ZK 😉😎

@Euler__Lagrange @worldcoin @reclaimprotocol @OpacityNetwork Ah yes, now I remember that constraint - this is because of the HMAC, right? I think DECO did a lot of optimizing for the garbled circuit to stay under the limits but all that only for 3pc. 

So is there a way to get around? 

How about at the systems-level: Let's say you get…

@vanishree_rao @worldcoin @reclaimprotocol @OpacityNetwork I’ll respond with 2 replies.

&gt; I think DECO did a lot of optimizing for the garbled circuit to stay under the limits but all that only for 3pc. 

So anyone reading has clarity, handshake is 3-party, but the mpc is 2 party. Only thing server knows thats its slower than usual.…

EulerLagrange.eth - bank/acc (Hersh)

Great thread on mpc-tls

Sometimes there are better solutions than ZK 😉😎

https://x.com/euler__lagrange/status/1875046837708554434?s=46

Working on building zkTLS infra/products. A ruthless empiricist and meme connoisseur.

/opacity
/eulerlagrange

TEE would solve that problem under that security assumption.

One thing to note is you can still fake proofs with a TEE without breaking the TEE.

You can configure the network adapter outside the TEE to route traffic through a malicious node that will swap out the transmission after the TEE signs.

https://x.com/euler__lagrange/status/1875616265441562996?s=46

Some tweets linked in this thread that explain

> For the proxy architecture, some node signs what ciphertext they saw go through the pipe. This node can sign whatever they want.

can’t you restrict what proxy signs using TEEs?