[{"author": "Christian Huitema", "text": "<p>And it is not yet Friday!</p>", "time": "2026-03-20T06:01:00.000Z"}, {"author": "Brian Trammell", "text": "<p>that's one way to stop having friday meetings: timezone shenanigans :D</p>", "time": "2026-03-20T06:01:03.000Z"}, {"author": "Marcus Ihlar", "text": "<p>That's the cheat code to avoid Friday meetings!</p>", "time": "2026-03-20T06:01:28.000Z"}, {"author": "Christian Huitema", "text": "<p>If we manage to be done in one hour, we win the \"no Friday\" contest!</p>", "time": "2026-03-20T06:03:56.000Z"}, {"author": "Brian Trammell", "text": "<p>looking at the agenda, if no friday is the goal, you might need to pop over to hawaii real quick</p>", "time": "2026-03-20T06:05:05.000Z"}, {"author": "Brian Trammell", "text": "<p>but let's see what we can do :)</p>", "time": "2026-03-20T06:05:36.000Z"}, {"author": "Christian Huitema", "text": "<p>At a minimum, lt's log Scone in qlog.</p>", "time": "2026-03-20T06:10:17.000Z"}, {"author": "Martin Duke", "text": "<p>I'm sorry is, there supposed to be a slide for what we're talking about?</p>", "time": "2026-03-20T06:15:28.000Z"}, {"author": "Christian Huitema", "text": "<p>Do we have a \"qlog extensions for scone\" defined somewhere?</p>", "time": "2026-03-20T06:16:33.000Z"}, {"author": "Qin Wu", "text": "<p>we have SCONE inter op  slides</p>", "time": "2026-03-20T06:18:05.000Z"}, {"author": "Martin Thomson", "text": "<p>Trying so hard not to do the hand gestures on that one.</p>", "time": "2026-03-20T06:20:13.000Z"}, {"author": "Brian Trammell", "text": "<p>use scone and it'll work</p>", "time": "2026-03-20T06:20:58.000Z"}, {"author": "Lucas Pardue", "text": "<p>what is the WG next steps after the protocol doc is done?</p>", "time": "2026-03-20T06:21:40.000Z"}, {"author": "Martin Duke", "text": "<p>i think I'm with Mirja on this. What possible deployment outcome would cause us to make a material change in the draft?</p>", "time": "2026-03-20T06:22:06.000Z"}, {"author": "Martin Thomson", "text": "<p>We might want to talk about applicability/manageability, but I don't see a lot more for <em>this</em> group.</p>", "time": "2026-03-20T06:22:18.000Z"}, {"author": "Martin Thomson", "text": "<p>Make a go/no go decision in Vienna?  Seems reasonable.</p>", "time": "2026-03-20T06:22:50.000Z"}, {"author": "Matt Joras", "text": "<p>IMO Vienna feels pointlessly long, but maybe that is just me</p>", "time": "2026-03-20T06:23:13.000Z"}, {"author": "Martin Thomson", "text": "<p>I would be OK with earlier, but I don't have any pressure to ship code that needs SCONE.</p>", "time": "2026-03-20T06:23:31.000Z"}, {"author": "Martin Thomson", "text": "<p>So if others have that pressure, I'm happy to be guided by that.</p>", "time": "2026-03-20T06:23:48.000Z"}, {"author": "Martin Thomson", "text": "<p>We can go today if you like.</p>", "time": "2026-03-20T06:23:59.000Z"}, {"author": "Matt Joras", "text": "<p>My other hot take would be that the 67 number actually ... doesn't matter that much</p>", "time": "2026-03-20T06:24:14.000Z"}, {"author": "Martin Thomson", "text": "<p>I agree Matt</p>", "time": "2026-03-20T06:24:34.000Z"}, {"author": "Kazuho Oku", "text": "<p>+1 to Matt</p>", "time": "2026-03-20T06:24:54.000Z"}, {"author": "Martin Thomson", "text": "<p>That's interop</p>", "time": "2026-03-20T06:24:58.000Z"}, {"author": "Christian Huitema", "text": "<p>We made the period long enough to not create too much overhead, but any end to end test has to last at least 5 or 10 periods to have a chance to measure effects. That means 10 minutes per test. Are we fine with that?</p>", "time": "2026-03-20T06:25:09.000Z"}, {"author": "Martin Thomson", "text": "<p>I think that the sort of testing that Meta did in the past was over weeks.</p>", "time": "2026-03-20T06:25:46.000Z"}, {"author": "Martin Thomson", "text": "<p>As in, you ship an A/B test to users and see how the various metrics track.</p>", "time": "2026-03-20T06:26:01.000Z"}, {"author": "Matt Joras", "text": "<p>That is correct!</p>", "time": "2026-03-20T06:26:20.000Z"}, {"author": "Martin Thomson", "text": "<p>A: users who get SCONE signals and follow them; B: users to get SCONE packets and ignore them; follow metrics for a good period.</p>", "time": "2026-03-20T06:26:43.000Z"}, {"author": "Lucas Pardue", "text": "<p>is the idea to publish these two docs together?</p>", "time": "2026-03-20T06:26:58.000Z"}, {"author": "Martin Thomson", "text": "<p>I don't think we need to go out together.</p>", "time": "2026-03-20T06:27:10.000Z"}, {"author": "Christian Huitema", "text": "<p>OK. I guess I will have to rig the simulator so I can do CI tests that take seconds to simulate 10 minutes...</p>", "time": "2026-03-20T06:27:18.000Z"}, {"author": "Martin Thomson", "text": "<p>Our simulator runs tests that simulate decades in milliseconds.</p>", "time": "2026-03-20T06:27:43.000Z"}, {"author": "Martin Thomson", "text": "<p>It's very useful.</p>", "time": "2026-03-20T06:27:52.000Z"}, {"author": "Martin Thomson", "text": "<p>For some reason, we only have a 6 week one-way delay in that test now.</p>", "time": "2026-03-20T06:29:25.000Z"}, {"author": "Christian Huitema", "text": "<p>I know, could not have developed picoquic without that kind of tech. But simulation load remain proportional to number of packets. So I guess I will need to simulate some very low rate video to minimize the packet load...</p>", "time": "2026-03-20T06:30:12.000Z"}, {"author": "Marcus Ihlar", "text": "<p>Early opsdir review is the appropriate action here.</p>", "time": "2026-03-20T06:31:46.000Z"}, {"author": "Martin Thomson", "text": "<p>I guess that if you are looking to simulate usage, that's going to take an awfully long time.  Our simulations only do things like handshakes when they have very large latencies.  Even then, the basic handshake retransmission leads to many, many packets.  Adding actual usage to that would be quite expensive.</p>", "time": "2026-03-20T06:31:46.000Z"}, {"author": "Lucas Pardue", "text": "<p>SG thanks chairs/AD</p>", "time": "2026-03-20T06:32:45.000Z"}, {"author": "Martin Thomson", "text": "<p><span class=\"user-mention\" data-user-id=\"811\">@Mirja K\u00fchlewind</span> those parts got removed.  It's a lot cleaner now.</p>", "time": "2026-03-20T06:33:28.000Z"}, {"author": "Christian Huitema", "text": "<p>I will think about that... Maybe integrate Scone with the implementation of QPERF that is used for performance testing.</p>", "time": "2026-03-20T06:33:59.000Z"}, {"author": "Christian Huitema", "text": "<p>Add an app-level message to pass the advice to the sender...</p>", "time": "2026-03-20T06:34:28.000Z"}, {"author": "Christian Huitema", "text": "<p>25 minutes to Friday!</p>", "time": "2026-03-20T06:36:13.000Z"}, {"author": "Martin Thomson", "text": "<p>One thing that I will observe is that this document is now so clean that it basically contains <em>less</em> applicability text than the main protocol document.</p>", "time": "2026-03-20T06:36:21.000Z"}, {"author": "Alan Frindell", "text": "<p>are you suggesting we should merge them in some way, or move some applicability text out of protocol?</p>", "time": "2026-03-20T06:36:54.000Z"}, {"author": "Martin Thomson", "text": "<p><span class=\"user-mention silent\" data-user-id=\"1495\">Alan Frindell</span> <a href=\"#narrow/channel/403-scone/topic/ietf-125/near/216624\">said</a>:</p>\n<blockquote>\n<p>are you suggesting we should merge them in some way, or move some applicability text out of protocol?</p>\n</blockquote>\n<p>I'm not sure that the main document is good if it doesn't contain at least some of that applicability information.  Because we use that to establish important context.</p>", "time": "2026-03-20T06:37:30.000Z"}, {"author": "Alan Frindell", "text": "<p>agree</p>", "time": "2026-03-20T06:37:41.000Z"}, {"author": "Martin Thomson", "text": "<p>I'm maybe coming to the view that there isn't much value add from the applicability document now.  I might even suggest dropping it entirely.  After checking that we didn't miss anything.</p>", "time": "2026-03-20T06:38:17.000Z"}, {"author": "Lucas Pardue", "text": "<p>sounds like we all need to review and consider that question</p>", "time": "2026-03-20T06:38:48.000Z"}, {"author": "Brian Trammell", "text": "<p>that might be an explicit question to the opsdir reviewers</p>", "time": "2026-03-20T06:38:50.000Z"}, {"author": "Alan Frindell", "text": "<p>having only one place to look does help fix the problem of not knowing which doc to look at</p>", "time": "2026-03-20T06:38:51.000Z"}, {"author": "Martin Thomson", "text": "<p><span class=\"user-mention\" data-user-id=\"781\">@Brian Trammell</span> are you going to ask for early review from ops?  That would be helpful.</p>", "time": "2026-03-20T06:39:15.000Z"}, {"author": "Marcus Ihlar", "text": "<p>A single document might be an improvement here given the size of this protocol and the issue mirja raised (people might not read the companion documents).</p>", "time": "2026-03-20T06:39:29.000Z"}, {"author": "Gorry Fairhurst", "text": "<p>I think the WG can write BOLD text in this spec saying at the start that specific stakeholders should read the apps/manageability. ... The WG could of course ask to publish them together, but then both need to be ready at the same time!</p>", "time": "2026-03-20T06:39:44.000Z"}, {"author": "Brian Trammell", "text": "<p>yep that's an AI for the chairs from this discussion, probably after we read the opsdir expectations doc gorry will send to the list and maybe have an interim cycle on it.</p>", "time": "2026-03-20T06:39:56.000Z"}, {"author": "Martin Thomson", "text": "<p>Mirja makes a good point here.  Let's have that detailed discussion of relation to L4S/AQM/CC.  Let's have a bit of a discussion of how this might be deployed in some network architectures (see Sri's presentation).</p>", "time": "2026-03-20T06:47:48.000Z"}, {"author": "Martin Thomson", "text": "<p>Right now, the applicability/manageability draft duplicates material from the main document, but with fewer words and less precision.  That's not helpful.  But if someone wants to add some of the stuff Mirja suggests, that would make this a very useful document.</p>", "time": "2026-03-20T06:49:47.000Z"}, {"author": "Martin Thomson", "text": "<p>+1 <span class=\"user-mention\" data-user-id=\"1053\">@Marcus Ihlar</span> !</p>", "time": "2026-03-20T06:50:47.000Z"}, {"author": "Lucas Pardue", "text": "<p>if concerned it should not be read normatively, stick it in an appendix</p>", "time": "2026-03-20T06:51:02.000Z"}, {"author": "Gorry Fairhurst", "text": "<p>I would like to ask at the Mic is it possible to simply publish the two at the SAME TIME - we can have adjacent numbers then:-)</p>", "time": "2026-03-20T06:51:14.000Z"}, {"author": "Martin Thomson", "text": "<p><span class=\"user-mention\" data-user-id=\"730\">@Gorry Fairhurst</span> You can get adjacent numbers even if they don't go out together.</p>", "time": "2026-03-20T06:51:43.000Z"}, {"author": "Kazuho Oku", "text": "<p>I think the question is exactly what additionally need to be said in A&amp;M, current text would fit well in protocol doc.</p>", "time": "2026-03-20T06:52:16.000Z"}, {"author": "Gorry Fairhurst", "text": "<p>We can.. but then we hold publication of one t await the second... which delays the RFC date/number</p>", "time": "2026-03-20T06:52:25.000Z"}, {"author": "Martin Thomson", "text": "<p><span class=\"user-mention\" data-user-id=\"730\">@Gorry Fairhurst</span> ... I don't think that this naturally leads to delays.  RPC are good at managing this stuff.  They can reserve us a number.</p>", "time": "2026-03-20T06:54:01.000Z"}, {"author": "Gorry Fairhurst", "text": "<p>@Martin - we can ask for what is needed.</p>", "time": "2026-03-20T06:54:56.000Z"}, {"author": "Martin Thomson", "text": "<p><span class=\"user-mention\" data-user-id=\"730\">@Gorry Fairhurst</span> yep.  Though I'm not seeing anything novel in this draft.</p>", "time": "2026-03-20T06:55:20.000Z"}, {"author": "Brian Trammell", "text": "<p>hey gorry's using scone now! :party:</p>", "time": "2026-03-20T06:55:37.000Z"}, {"author": "Kazuho Oku", "text": "<p>+1 to Mirja</p>", "time": "2026-03-20T06:58:52.000Z"}, {"author": "Lucas Pardue", "text": "<p>succint, compact, operators and networks explainer</p>", "time": "2026-03-20T07:02:37.000Z"}, {"author": "Alan Frindell", "text": "<p>i see what you did there</p>", "time": "2026-03-20T07:05:06.000Z"}, {"author": "Brian Trammell", "text": "<p>wow that took me too long, more coffee plz</p>", "time": "2026-03-20T07:06:06.000Z"}, {"author": "Martin Thomson", "text": "<p>A new version number would need far stronger justification than what I've heard here.</p>", "time": "2026-03-20T07:22:30.000Z"}, {"author": "Mark Grayson", "text": "<p>apologies for MIC issues - my comment was to re-enforce Sri's point - we heard SA2 integrating SCONE into an E2E 5G architecture, and WFA/802.11 focused on AP-to-device 802.11 specifics, so there is a need to understand the E2E view for Wi-Fi</p>", "time": "2026-03-20T07:32:48.000Z"}, {"author": "Ingemar Johansson", "text": "<p>My (5G view) on throughput guidance that goes into the area of congestion control is that MCS in 5G access typically vary a lot on short time scales due to e.g fast fading. That would require quite frequent SCONE \"packets\". As a comparison, L4S inherently tracks these fast changes well over a low cost signaling channel (ECN bits)</p>", "time": "2026-03-20T07:32:54.000Z"}, {"author": "Matt Joras", "text": "<p>reminder #735 that SCONE is not congestion control <span aria-label=\"smiley\" class=\"emoji emoji-1f603\" role=\"img\" title=\"smiley\">:smiley:</span></p>", "time": "2026-03-20T07:34:38.000Z"}, {"author": "Martin Thomson", "text": "<p>I'm having audio issues, like Mark, but I'm not clear on who would send this signal in MASQUE?</p>", "time": "2026-03-20T07:34:41.000Z"}, {"author": "Alan Frindell", "text": "<p>the masque proxy injects capsules?</p>", "time": "2026-03-20T07:35:02.000Z"}, {"author": "Kazuho Oku", "text": "<p>a node that can inject capsules can also modify datagram capsules no? So why use a different mechanism?</p>", "time": "2026-03-20T07:35:42.000Z"}, {"author": "Martin Thomson", "text": "<p>If this is intended to communicate from a proxy to a client, they have many options, including updating SCONE packets.  Or using flow control.  Or using ECN markings.  Each has a different applicability.</p>", "time": "2026-03-20T07:35:43.000Z"}, {"author": "Martin Thomson", "text": "<p>Why would you use a capsule when the content of datagram capsules can be tweaked?</p>", "time": "2026-03-20T07:36:08.000Z"}, {"author": "Lucas Pardue", "text": "<p>yeah a proxy is a netowrk element</p>", "time": "2026-03-20T07:36:16.000Z"}, {"author": "Martin Thomson", "text": "<p><span class=\"user-mention silent\" data-user-id=\"5132\">Lucas Pardue</span> <a href=\"#narrow/channel/403-scone/topic/ietf-125/near/217009\">said</a>:</p>\n<blockquote>\n<p>yeah a proxy is a netowrk element</p>\n</blockquote>\n<p>exactamundo</p>", "time": "2026-03-20T07:36:27.000Z"}, {"author": "Alan Frindell", "text": "<p>I think they are leveraging the capsule protocol channel to send richer information.</p>", "time": "2026-03-20T07:37:12.000Z"}, {"author": "Kazuho Oku", "text": "<p>it'd be much more interesting if a masque server generated and attached a new SCONE to the outer datagram, so that the client can receive the outer signal and the inner signal.</p>", "time": "2026-03-20T07:38:00.000Z"}, {"author": "Martin Thomson", "text": "<p><span class=\"user-mention\" data-user-id=\"1164\">@Kazuho Oku</span> yeah, that seems sensible.  Though I'd caution against matching the inner and outer packets (you shouldn't just add SCONE every time that the tunneled packet has SCONE).  Otherwise, that's a privacy leak.</p>", "time": "2026-03-20T07:39:19.000Z"}, {"author": "Martin Thomson", "text": "<p>Independently adding SCONE to the outer tunnel would be fine.</p>", "time": "2026-03-20T07:39:36.000Z"}, {"author": "Martin Thomson", "text": "<p>A good idea, even.</p>", "time": "2026-03-20T07:39:44.000Z"}, {"author": "Ted Hardie", "text": "<p>Suddenly, Mijra is very quiet compared to before.</p>", "time": "2026-03-20T07:41:19.000Z"}, {"author": "Martin Thomson", "text": "<p>On quic-aware:</p>\n<blockquote>\n<p>QUIC long header packets MUST NOT be forwarded. These packets can only be tunnelled within HTTP Datagram frames to avoid exposing unnecessary connection metadata.</p>\n</blockquote>\n<p>That suggests to me that we're OK making an exception for QUIC-aware forwarding and SCONE.</p>", "time": "2026-03-20T07:42:29.000Z"}, {"author": "Martin Thomson", "text": "<p>forwarding SCONE through isn't ideal, because it reveals the scone throughput advice, but the way we'd manage that is to extend the scrambling protection to cover that signal.</p>", "time": "2026-03-20T07:43:33.000Z"}, {"author": "Martin Thomson", "text": "<p>That would be pure optimization (which quic-aware is already).  The default quic-aware behavior is perfectly fine.</p>", "time": "2026-03-20T07:44:30.000Z"}, {"author": "Kazuho Oku", "text": "<p>yeah the outer SCONE can be generated by masque proxies as UNLIMITED B/W, then the client can merge inner and outer. It just looks like the marking is leaky.</p>", "time": "2026-03-20T07:44:48.000Z"}, {"author": "Martin Thomson", "text": "<p>Well, by default, the marking wouldn't be leaky.  It's only if you tried extra extra hard to optimize that you might create a leak.  But we know how to fix that leak as well if that is what we wanted to do.</p>", "time": "2026-03-20T07:45:27.000Z"}, {"author": "Lucas Pardue", "text": "<p>but also to answe the question, don't do anything in the SCONE WG. send it to masque</p>", "time": "2026-03-20T07:45:41.000Z"}, {"author": "Ted Hardie", "text": "<p>Didn't we already dispatch this in QUIC to a BoF?</p>", "time": "2026-03-20T07:45:53.000Z"}, {"author": "Martin Thomson", "text": "<p><span class=\"user-mention silent\" data-user-id=\"40\">Ted Hardie</span> <a href=\"#narrow/channel/403-scone/topic/ietf-125/near/217062\">said</a>:</p>\n<blockquote>\n<p>Didn't we already dispatch this in QUIC to a BoF?</p>\n</blockquote>\n<p>quic wg doesn't dispatch, but yes, that was the feedback there.  It should be the same here.</p>", "time": "2026-03-20T07:46:26.000Z"}, {"author": "Qin Wu", "text": "<p>based on earlier offline discussion, both WG chairs offer chance to introduce this proposal , @Ted</p>", "time": "2026-03-20T07:47:57.000Z"}, {"author": "Ted Hardie", "text": "<p>Given the previous history, I'm very surprised at Brian's agreement.  He must heal faster than me.</p>", "time": "2026-03-20T07:48:24.000Z"}, {"author": "Brian Trammell", "text": "<p>heh</p>", "time": "2026-03-20T07:48:46.000Z"}, {"author": "Martin Thomson", "text": "<p><span class=\"user-mention silent\" data-user-id=\"40\">Ted Hardie</span> <a href=\"#narrow/channel/403-scone/topic/ietf-125/near/217074\">said</a>:</p>\n<blockquote>\n<p>Given the previous history, I'm very surprised at Brian's agreement.  He must heal faster than me.</p>\n</blockquote>\n<p>Brian still holds a candle for SPUD.</p>", "time": "2026-03-20T07:48:49.000Z"}, {"author": "Brian Trammell", "text": "<p>spud x scone yields LATKE (Loss and Timing Knowledge Exposure)</p>", "time": "2026-03-20T07:49:24.000Z"}, {"author": "Ted Hardie", "text": "<p>I still think my slides from PLUS were a solid explanation of the implicit path layer, but that doesn't mean they persuaded anybody much.</p>", "time": "2026-03-20T07:49:44.000Z"}, {"author": "Martin Thomson", "text": "<p>Oh, you've clearly been working on that one for a while.</p>", "time": "2026-03-20T07:49:47.000Z"}, {"author": "Brian Trammell", "text": "<p>say what you will about llms, they're really good at backronyms.</p>", "time": "2026-03-20T07:50:12.000Z"}, {"author": "Ted Hardie", "text": "<p>@MT He gave this time in the session entirely to spring that on us.</p>", "time": "2026-03-20T07:50:13.000Z"}, {"author": "Martin Thomson", "text": "<p>It's starting to feel like that.</p>", "time": "2026-03-20T07:50:27.000Z"}, {"author": "Martin Thomson", "text": "<p>Don't they have open mic nights where you live Brian?</p>", "time": "2026-03-20T07:50:43.000Z"}, {"author": "Brian Trammell", "text": "<p>yeah but the venn diagram of \"gets ietf humor\" and \"speaks swiss german\" is a bit lonely</p>", "time": "2026-03-20T07:51:18.000Z"}, {"author": "Ted Hardie", "text": "<p>I'm sure Eliot would crack up.</p>", "time": "2026-03-20T07:51:45.000Z"}, {"author": "Brian Trammell", "text": "<p>more seriously: I do think that the problems here don't go away with time, and it's useful to review which parts of the problem we have how much appetite to solve as time goes on</p>", "time": "2026-03-20T07:52:43.000Z"}, {"author": "Brian Trammell", "text": "<p>as ted pointed out in the chat in QUIC, SCONE wouldn't have flown ten years ago.</p>", "time": "2026-03-20T07:53:44.000Z"}, {"author": "Ted Hardie", "text": "<p>@Brian hence the BoF suggestion</p>", "time": "2026-03-20T07:53:47.000Z"}, {"author": "Lucas Pardue", "text": "<p>I've still not experienced a case where loss bits would have solved a problem</p>", "time": "2026-03-20T07:54:09.000Z"}, {"author": "Ted Hardie", "text": "<p>But we don't need to hear it multiple times to say \"take it to a BoF\"</p>", "time": "2026-03-20T07:54:16.000Z"}, {"author": "Brian Trammell", "text": "<p>that's useful feedback</p>", "time": "2026-03-20T07:54:26.000Z"}, {"author": "Brian Trammell", "text": "<p>we put this on both agendas because it wasn't how loud and clear that feedback would be from quic</p>", "time": "2026-03-20T07:54:44.000Z"}, {"author": "Lucas Pardue", "text": "<p>also QUIC was a short session and we might have bumped marcus</p>", "time": "2026-03-20T07:55:11.000Z"}, {"author": "Matt Joras", "text": "<p>These kinds of things usually aren't useful for endpoints, they're disproportionately useful to operators</p>", "time": "2026-03-20T07:55:41.000Z"}, {"author": "Brian Trammell", "text": "<p>there's also another point: you only really need this if you think QUIC will tend to 100% transport presence on links of interest. if we stay stuck at 33% forever then you can just keep looking at TCP</p>", "time": "2026-03-20T07:55:44.000Z"}, {"author": "Martin Thomson", "text": "<p>Ted just spun a bit!</p>", "time": "2026-03-20T07:57:04.000Z"}, {"author": "Lucas Pardue", "text": "<p>seems nobody is chomping at the bit</p>", "time": "2026-03-20T07:57:36.000Z"}, {"author": "Christian Huitema", "text": "<p>SPUD was so much fun... Especially if you did not care too much about it.</p>", "time": "2026-03-20T07:57:59.000Z"}, {"author": "Martin Thomson", "text": "<p>but then we are done, right? right?!?</p>", "time": "2026-03-20T07:59:17.000Z"}, {"author": "Mirja K\u00fchlewind", "text": "<p>We are never done.</p>", "time": "2026-03-20T07:59:35.000Z"}, {"author": "Ted Hardie", "text": "<p>It's the weekend, @MT, on to better things!</p>", "time": "2026-03-20T07:59:42.000Z"}, {"author": "Martin Thomson", "text": "<p>never say never... ?</p>", "time": "2026-03-20T07:59:46.000Z"}, {"author": "Mirja K\u00fchlewind", "text": "<p>Next step AI-native scone</p>", "time": "2026-03-20T07:59:57.000Z"}, {"author": "Ted Hardie", "text": "<p>It's four o'clock, but yeah.</p>", "time": "2026-03-20T08:00:33.000Z"}, {"author": "altanai", "text": "<p>1 am in Pacific timezone </p>\n<p>see you in next IETF</p>", "time": "2026-03-20T08:00:45.000Z"}, {"author": "Brian Trammell", "text": "<p>ahhhh</p>", "time": "2026-03-20T08:00:54.000Z"}, {"author": "Brian Trammell", "text": "<p>utc+8, i'm utc+1.</p>", "time": "2026-03-20T08:01:07.000Z"}, {"author": "Brian Trammell", "text": "<p>math is hard</p>", "time": "2026-03-20T08:01:09.000Z"}, {"author": "Brian Trammell", "text": "<p>there's a a plus joke in there somewhere</p>", "time": "2026-03-20T08:01:20.000Z"}]