[{"author": "Cathie Yun", "text": "<p><span aria-label=\"grinning\" class=\"emoji emoji-1f600\" role=\"img\" title=\"grinning\">:grinning:</span></p>", "time": "2025-07-23T12:30:03Z"}, {"author": "Scott Hendrickson", "text": "<p>will get those revised i-d's out soon. Thanks all!</p>", "time": "2025-07-23T12:34:52Z"}, {"author": "Benjamin Schwartz", "text": "<p>I lost audio from Cathie</p>", "time": "2025-07-23T12:38:23Z"}, {"author": "Nick Doty", "text": "<p>still hear her fine here</p>", "time": "2025-07-23T12:38:33Z"}, {"author": "Benjamin Schwartz", "text": "<p>Reloaded, fixed</p>", "time": "2025-07-23T12:38:58Z"}, {"author": "Christopher Patton", "text": "<p>\"pairings are trivial depending on who you ask\" :D</p>", "time": "2025-07-23T12:42:00Z"}, {"author": "Nick Doty", "text": "<p>if nonce is a number between 0 and rate-limit, then it seems less like a nonce and more like the usage number</p>", "time": "2025-07-23T12:46:51Z"}, {"author": "Christopher Wood", "text": "<p>It's literally a number that can be used at most once</p>", "time": "2025-07-23T12:47:14Z"}, {"author": "Nick Doty", "text": "<p>maybe we think of a nonce as a number used once by anyone (a big random number), rather than a sequential number in a small set used one at a time</p>", "time": "2025-07-23T12:48:12Z"}, {"author": "Benjamin Schwartz", "text": "<p>FYI, the underlying crypto is here: <a href=\"https://datatracker.ietf.org/doc/draft-yun-cfrg-arc/\">https://datatracker.ietf.org/doc/draft-yun-cfrg-arc/</a></p>", "time": "2025-07-23T12:48:18Z"}, {"author": "Watson Ladd", "text": "<p>? i've never seen nonce having to be globally unique</p>", "time": "2025-07-23T12:48:45Z"}, {"author": "Benjamin Schwartz", "text": "<p>Queue is closed</p>", "time": "2025-07-23T12:49:10Z"}, {"author": "Watson Ladd", "text": "<p>Sam said what I was going to</p>", "time": "2025-07-23T12:49:35Z"}, {"author": "Samuel", "text": "<p>You can put the origin in the presentation context, I think is what Tommy is saying</p>", "time": "2025-07-23T12:53:06Z"}, {"author": "Samuel", "text": "<p>Maybe I misunderstood</p>", "time": "2025-07-23T12:53:10Z"}, {"author": "Samuel", "text": "<p>We'd be using something even less far along, nobody has ever specified publicly verifiable CMZ</p>", "time": "2025-07-23T12:55:51Z"}, {"author": "Samuel", "text": "<p>Or implemented it, to my knowledge</p>", "time": "2025-07-23T12:56:02Z"}, {"author": "Deirdre Connolly", "text": "<p>Does going into public key verifiable have any consequences for supporting PQ options?</p>", "time": "2025-07-23T12:56:03Z"}, {"author": "Samuel", "text": "<p>Nope, the deanonymization is just as bad on Q-day for both</p>", "time": "2025-07-23T12:56:44Z"}, {"author": "Christopher Wood", "text": "<p>^^^ is an important point. ARC is a privacy regression against a PQ adversary.</p>", "time": "2025-07-23T12:57:14Z"}, {"author": "Christopher Wood", "text": "<p>(compared to blind RSA)</p>", "time": "2025-07-23T12:57:20Z"}, {"author": "Samuel", "text": "<p>Yep, the anonymous credits I'll talk about and Watson's talk will give ideas towards avoiding Q-day deanonymization</p>", "time": "2025-07-23T12:58:13Z"}, {"author": "Christopher Patton", "text": "<p>+1 ChrisW, let's not add privacy footguns</p>", "time": "2025-07-23T12:59:35Z"}, {"author": "Deirdre Connolly", "text": "<p>Sorry what I mean is do we have good PQ options anywhere on the horizon that would also fit in the public key verification box</p>", "time": "2025-07-23T12:59:56Z"}, {"author": "Samuel", "text": "<p>Good is maybe subjective, but we have options on the horizon</p>", "time": "2025-07-23T13:00:16Z"}, {"author": "Deirdre Connolly", "text": "<p>Yeah in PQ having anything at all is good</p>", "time": "2025-07-23T13:00:28Z"}, {"author": "Deirdre Connolly", "text": "<p>Small, fast, well, can't have everything <span aria-label=\"sweat smile\" class=\"emoji emoji-1f605\" role=\"img\" title=\"sweat smile\">:sweat_smile:</span></p>", "time": "2025-07-23T13:00:40Z"}, {"author": "Christopher Wood", "text": "<p>Gus is small and fast</p>", "time": "2025-07-23T13:00:51Z"}, {"author": "Christopher Wood", "text": "<p>It is possible</p>", "time": "2025-07-23T13:00:52Z"}, {"author": "Deirdre Connolly", "text": "<p>ooo link to Gus?</p>", "time": "2025-07-23T13:01:03Z"}, {"author": "Christopher Patton", "text": "<p><span class=\"user-mention silent\" data-user-id=\"603\">Deirdre Connolly</span> <a href=\"#narrow/channel/259-privacypass/topic/ietf-123/near/173316\">said</a>:</p>\n<blockquote>\n<p>Sorry what I mean is do we have good PQ options anywhere on the horizon that would also fit in the public key verification box</p>\n</blockquote>\n<p>In the realm of fancy crypto, I would rather be able to scope our efforts to the fewest possible number of variants possible. Private verifiability is sufficient for many use cases, so let's solve the PQ storry for that first.</p>", "time": "2025-07-23T13:01:16Z"}, {"author": "Deirdre Connolly", "text": "<p>I also support adoption</p>", "time": "2025-07-23T13:01:17Z"}, {"author": "Christopher Wood", "text": "<p>Can't doxx the dude</p>", "time": "2025-07-23T13:01:18Z"}, {"author": "Christopher Patton", "text": "<p>@chairs we remote folks hear a lot of echo in the room</p>", "time": "2025-07-23T13:04:49Z"}, {"author": "Lorenzo Miniero", "text": "<p>Is it better now?</p>", "time": "2025-07-23T13:05:21Z"}, {"author": "Christopher Patton", "text": "<p>yes :)</p>", "time": "2025-07-23T13:05:28Z"}, {"author": "Benjamin Schwartz", "text": "<p><span class=\"user-mention\" data-user-id=\"6359\">@Cathie Yun</span> Do I understand correctly that ARC for some limit N is functionally equivalent to issuing N VOPRF tokens, apart from efficiency?  Do you have a sense of what value of N is required for ARC issuance to be more efficient than Batched VOPRF (which also improves efficiency for large N)?</p>", "time": "2025-07-23T13:05:54Z"}, {"author": "Deirdre Connolly", "text": "<p>Love that Fiat-Shamir paper</p>", "time": "2025-07-23T13:05:57Z"}, {"author": "Christopher Wood", "text": "<p>@Ben this was asked during the last meeting, where I said we'd do some performance benchmarks to run this comparison. I didn't do that yet, but we still can do it</p>", "time": "2025-07-23T13:06:57Z"}, {"author": "Christopher Patton", "text": "<p><span class=\"user-mention\" data-user-id=\"6359\">@Cathie Yun</span> to refine my take on public vs private verifiability: I think it's totally fair to focus your draft on one cryptographic paradigm (algebraic MACs + Sigma proofs) rather than having to deal with two in the same draft.</p>", "time": "2025-07-23T13:07:03Z"}, {"author": "Samuel", "text": "<p>ARC can produce N tokens for an arbitrary context, so N * 2^256 or something, which you cannot get from issuing N VOPRF tokens.</p>", "time": "2025-07-23T13:07:25Z"}, {"author": "Deirdre Connolly", "text": "<p><span class=\"user-mention silent\" data-user-id=\"3787\">Christopher Patton</span> <a href=\"#narrow/channel/259-privacypass/topic/ietf-123/near/173369\">said</a>:</p>\n<blockquote>\n<p><span class=\"user-mention silent\" data-user-id=\"6359\">Cathie Yun</span> to refine my take on public vs private verifiability: I think it's totally fair to focus your draft on one cryptographic paradigm (algebraic MACs + Sigma proofs) rather than having to deal with two in the same draft.</p>\n</blockquote>\n<p>agreed</p>", "time": "2025-07-23T13:07:33Z"}, {"author": "Samuel", "text": "<p>2^256 = infinity anyways :)</p>", "time": "2025-07-23T13:07:38Z"}, {"author": "Deirdre Connolly", "text": "<p><span class=\"user-mention silent\" data-user-id=\"603\">Deirdre Connolly</span> <a href=\"#narrow/channel/259-privacypass/topic/ietf-123/near/173365\">said</a>:</p>\n<blockquote>\n<p>Love that Fiat-Shamir paper</p>\n</blockquote>\n<p>(<a href=\"https://eprint.iacr.org/2025/118\">https://eprint.iacr.org/2025/118</a>)</p>", "time": "2025-07-23T13:07:48Z"}, {"author": "Benjamin Schwartz", "text": "<p>Thanks Samuel</p>", "time": "2025-07-23T13:07:50Z"}, {"author": "Watson Ladd", "text": "<p><a href=\"https://eprint.iacr.org/2025/1030\">https://eprint.iacr.org/2025/1030</a> is the one from the slides</p>", "time": "2025-07-23T13:10:20Z"}, {"author": "Christopher Patton", "text": "<p>Agree with ChrisW and Thibault. This is cool, but is more researchy. Perhaps once there's a concrete problem that's solved and you have a concrete protocol (implemented in as much detail as ARC, say), then come back and tell us about it?</p>", "time": "2025-07-23T13:13:46Z"}, {"author": "Watson Ladd", "text": "<p>yeah, that's useful feedback</p>", "time": "2025-07-23T13:14:45Z"}, {"author": "Benjamin Schwartz", "text": "<p>If PQ security is the topic, I would also be interested in an analysis of PQ threat models for Privacy Pass, which doesn't seem trivial.  It seems like there are a few different important threat models, like \"CRQC exists but is more expensive than other attacks on the protected resource\".</p>", "time": "2025-07-23T13:16:55Z"}, {"author": "Christopher Patton", "text": "<p><span class=\"user-mention\" data-user-id=\"6359\">@Cathie Yun</span> <span class=\"user-mention\" data-user-id=\"128\">@Christopher Wood</span> aside: should we be calling the ARC tag the ARC \"nullifier\"? I'm wondering if these have different semantics in your mind.</p>", "time": "2025-07-23T13:18:32Z"}, {"author": "Christopher Patton", "text": "<p>Doesn't ARC have statistical (i.e., PQ) anonymity?</p>", "time": "2025-07-23T13:22:10Z"}, {"author": "Christopher Wood", "text": "<p>No -- computational</p>", "time": "2025-07-23T13:22:17Z"}, {"author": "Cathie Yun", "text": "<p>Yes you can think of the ARC tag as a nullifier</p>", "time": "2025-07-23T13:22:21Z"}, {"author": "Nick Doty", "text": "<p>'you must go back to the issuer with this particular signal I want to communicate to them' sounds like verifier-issuer linkability</p>", "time": "2025-07-23T13:22:29Z"}, {"author": "Cathie Yun", "text": "<p>And +1, ARC doesn't have PQ unlinkability</p>", "time": "2025-07-23T13:22:40Z"}, {"author": "Dennis Jackson", "text": "<p>I missed the first part of the meeting and this might have been covered (Sorry!). Do the authors of ARC have an opinion on how hard it would be to provide PQ unlinkability?</p>", "time": "2025-07-23T13:23:52Z"}, {"author": "Christopher Patton", "text": "<p><span class=\"user-mention silent\" data-user-id=\"6359\">Cathie Yun</span> <a href=\"#narrow/channel/259-privacypass/topic/ietf-123/near/173469\">said</a>:</p>\n<blockquote>\n<p>And +1, ARC doesn't have PQ unlinkability</p>\n</blockquote>\n<p><span class=\"user-mention\" data-user-id=\"128\">@Christopher Wood</span> pointed me to a refinement of CMZ that seems to claim statistical anonymity <a href=\"https://eprint.iacr.org/2024/1552\">https://eprint.iacr.org/2024/1552</a>. Is that true? If so, would it be worth implementing that improvement?</p>", "time": "2025-07-23T13:24:01Z"}, {"author": "Christopher Patton", "text": "<p>(I'm still digging into this paper a bit.)</p>", "time": "2025-07-23T13:24:12Z"}, {"author": "Cathie Yun", "text": "<p>I don't think the \"revisiting KVAC\" paper gives statistical anonymity for unlinkability between presentations. It only has statistical anonymity for unlinkability between issuance and presentation</p>", "time": "2025-07-23T13:24:45Z"}, {"author": "Christopher Patton", "text": "<p>ah OK</p>", "time": "2025-07-23T13:24:57Z"}, {"author": "Christopher Wood", "text": "<p>@Dennis Cathie knows that best -- Cathie, can you say more?</p>", "time": "2025-07-23T13:25:00Z"}, {"author": "Christopher Patton", "text": "<p>and indeed the theorem about the tag generation bit is based on a computational assumption</p>", "time": "2025-07-23T13:25:26Z"}, {"author": "Watson Ladd", "text": "<p>the problem with ARC isn't the credential show. It's the PRF that's involved</p>", "time": "2025-07-23T13:25:27Z"}, {"author": "Cathie Yun", "text": "<p>Best attempt at PQ unlinkability would be what Watson presented on. Seems nontrivial and fairly high overhead, but down to explore that more.</p>", "time": "2025-07-23T13:25:33Z"}, {"author": "Cathie Yun", "text": "<p>(PQ unlinkability of presentations, to be precise)</p>", "time": "2025-07-23T13:26:58Z"}, {"author": "Christopher Patton", "text": "<p>public metadata makes me really nervous</p>", "time": "2025-07-23T13:31:08Z"}, {"author": "Christopher Patton", "text": "<p>I don't think we've done a good job wrangling this</p>", "time": "2025-07-23T13:31:16Z"}, {"author": "Christopher Wood", "text": "<p>+1</p>", "time": "2025-07-23T13:31:22Z"}, {"author": "Dennis Jackson", "text": "<p>Ok, thanks Cathie! I think the PQ aspect might be a good discussion for the list ahead of adoption.</p>", "time": "2025-07-23T13:31:36Z"}, {"author": "Dennis Jackson", "text": "<p>+1 Chris P</p>", "time": "2025-07-23T13:31:41Z"}, {"author": "Christopher Patton", "text": "<p>FWIW I don't think PQ should be a requirement for adopting ARC.</p>", "time": "2025-07-23T13:32:03Z"}, {"author": "Deirdre Connolly", "text": "<p><span class=\"user-mention silent\" data-user-id=\"3787\">Christopher Patton</span> <a href=\"#narrow/channel/259-privacypass/topic/ietf-123/near/173514\">said</a>:</p>\n<blockquote>\n<p>FWIW I don't think PQ should be a requirement for adopting ARC.</p>\n</blockquote>\n<p>Definitely not, I'm mostly just curious</p>", "time": "2025-07-23T13:32:13Z"}]