tim-clancy.eth

User layer @ethereumfndn, I love @ethereum with a Crusader's zeal; suffer not the weaker souls. Magnifica Æthereum. Was zealot @Sigil_L2 still zealot.

[tim-clancy.eth @_Enoch]

@dumbnamenumbers @joddoss Yes, as is Ethereum Classic. My view of who holds the Mandate of Heaven for the one true World Computer is based on whichever underlying network is fully programmable and is the most decentralzed/has the highest total cost of attack.

[tim-clancy.eth @_Enoch]

@l2beat Still think you should change this category to point out those different licenses. It doesn't matter for underlying security but it does matter for O.

[tim-clancy.eth @_Enoch]

@joddoss I see where you are coming from, but until I am tendered an equally-snappy term alt-L1 will simply remain my perjorative of choice.

[tim-clancy.eth @_Enoch]

That's easy; Ethereum. Anything that is a standalone network not Ethereum or fully-proven by Ethereum is an alt-L1. Anything that is fully-proven is an L2. The term L2 is at times aspirational; per our friends at L2Beat being watchdogs. This implies the existence of alt-L2s but I shudder at the thought.

[tim-clancy.eth @_Enoch]

@z0r0zzz @dxnzn_ @tacitfi the market is always wrong; they've gobbled up multisigfi and trustful solutions for years

[tim-clancy.eth @_Enoch]

@hanni_abu @0xSachinK @0xjei No, zkTLS is gone.

[tim-clancy.eth @_Enoch]

@zkpedrongmi @SpaceComputerIO Since the TEE trust model necessarily collapses to "physical access is death" I used to always joke that putting them in orbit under 24/7 surveillance is the only way to fix them. 😁

[tim-clancy.eth @_Enoch]

@CPerezz19 @RAILGUN_Project @0xprivacypools Zero for RG because closed source license and effectively-permissioned, about one for PP because permissioned, and everything I hold for TC.

[tim-clancy.eth @_Enoch]

Yeah nothing matters because we've only got 2PC and the attestor can always forge proof of payment but I meant we could remove the seller's trust in the faithful downstream evaluation of an otherwise-correct proof; that logic could be proven onchain instead of running offchain. We could get to a point where it would really mean that the only way to screw a seller would be to actually forge a proof. Bigger fish to fry I suppose.

[tim-clancy.eth @_Enoch]

We might be talking past each other. The attestation gets posted onchain but it sounds like the current check is confirming it came from the trusted attestor. What I'm saying is that with zkTLS, the seller should not actually need to trust the attestor because a smart contract should be able to do all of the attestor's verification work by confirming correctness of the zkTLS details directly (or a zkVM consumes the zkTLS data and generates a proof).

[tim-clancy.eth @_Enoch]

Yeah of course, I just mean that once the buyer/attestor finished the 2PC I thought the zkTLS output was being verified onchain, either directly or via wrapping in a zkVM first or something. You can never shake the liveness problem but you could enforce soundness for the seller's end and only release their funds with an actual proof. This seems strictly better than the TEE assumptions.

[tim-clancy.eth @_Enoch]

@z0r0zzz @dxnzn_ @tacitfi you should make a version where we don't need to assume the trust assumptions of an alt-L1

[tim-clancy.eth @_Enoch]

@drahcir @peerxyz Ouch. I'd been under the impression that the zkTLS proof itself was being verified on Ethereum as a precondition for seller release, and a malicious attestor screwed the buyer. What you've described is a lot weaker than I misunderstood Peer to be.

[tim-clancy.eth @_Enoch]

@CarlotaFang77 carlota this is an alt-L1

[tim-clancy.eth @_Enoch]

@NeerajKA How much does this concession hurt future Romans when evil prosecutors try to allege that building unconditional privacy tools means one had _specific intent_? I'm disappointed in this one because I saw the 1960 change as the entire point.

[tim-clancy.eth @_Enoch]

Nooooooo!!!

[Peer (Prev. ZKP2P) @peerxyz]

Peer is now running on TEEs.

Payment verification just got 100x faster, from ~30s to <300ms.

Here's what changed.