The brief · 2026-10-31__rc1-paper-review · briefs/2026-10-31__rc1-paper-review/batch-08-pages-34-to-38.md

Batch 8: pages 34 to 38: images that belong, control flow, timelines as components, two kinds of trigger

The editor, in the conductor's session, 31 October 2026. Kept as he gave it.

Okay, now on page 34. Yeah, so this is a perfect example where the user suddenly is reading about hold, right? He's reading about this technique and then sees an image that have nothing to do with hold, right? You know, they're cool images, but they actually mix multiple things. So now we have to make sure that when every time there's an image on a page, it makes sense, and it makes sense to the reader, and he understands why it's there, and it's not just almost a padding image, which this almost feels like it, right? Now on page 35, I think this whole action section here, again, going back to the hold, needs to be in a section about control flow, about how we are starting to now use this sort of state machine and control flow, right? Because that is super critical, and that's kind of what we're talking about it, right? And then the stop is one of the elements that we have in our control flow, which is how we keep our agents controlled, but also like the agent behavior policy and all that stuff, right? That's also quite important. Page 36, this is important, but needs to be in a section about the logistics, right, of doing this. Like, you know, in a way, the limitations in agents are hard. We talked about it, right? The power of reuse, you know, cloud session. So we explained why we do this, like the logic that eventually we arrive on having a cloud session that we can reuse, but also the limitation of 1 hour schedule trigger. So you cannot have more than 1 hour, you know, sort of less than an hour schedule, which, you know, it's okay. It's a, you know, it's a good, again, constraint, right? Because it gets you that cadence that we want, right? So that's kind of what we need to do, right? And now on page 38, I think, now I said here, we finally get some meat, right? And the thing about this, I like the timeline. I think we should be using timelines left, right, and center. Again, it's probably a good example of mini tools that you can start to create to make sure that then you just use JSON or you connect to it. But I think we need better time entries here, right? I think we should explore multiple ways to show the, you know, and the content, and also we should be exploring multiple ways to even show this timeline, right? Because these timelines are very easy to understand. So I think it's worth, again, probably doing a whole round. And I think for some of these, you think of it as also as what the, you know, like the CSS frameworks have, and, you know, Bootstrap and Foundation, all the other ones, where they have this sort of almost toolkit, right? Or have this example of multiple options. And remember that we want this to be reused by other agents, even our own agents. So, you know, we should have these reusable components, you know, which, you know, again, you can code them now. So even like a lot of the little infographics or things you do, that can actually be a piece of code that then we update, right? And that will allow us to scale a lot more on these, like, for example, these sort of timelines that we have, right? These timelines, you know, it's, we should explore the multiple ways to present it, you know, horizontal, vertical, you know, different modes, etc., right? And different visualizations. And again, this is, we can give again briefs like this to Claude, to ChatGPT, to Claude Design, to other agents, and see what they come up with. Yeah, so I think what's more interesting, right, when we talk about this timeline stuff, I think it's more interesting to show the work that happens automatically, right? And the work that happens that is triggered by me. So if you think about it, like, now you now have two very distinct ways of triggering work, right? So when I talk about the loop, what triggers the loop, right? So one of them is me, when I send a chat message that triggers a loop, and then we have an opportunity to work on a live agent that will continue for a little while, okay, until it reaches a state where it thinks there's nothing to continue, or there's a natural pause, right? So that's what the agent provides. But what I think is interesting is to map those out, right? Map like the work that was happening automatically, and then the work that happened when I provide. In fact, there should be almost some kind of noticeable difference, right, in some of the work, some of the activities. But also, can you try to again capture some stats, capture like the token consumption, right, and capture like the network calls and stuff like that? Because I think it's that stuff that happens automatically, right, and the stuff that happens then, you know, when I get involved, is very important. A human gets involved, is very important, right? And again, this touches day or night, right? So we should go there, right? Yeah, so I still think that on our website, one of the best images that we have of all of this, in fact, I had it shown on the screen before I even presented, was the view that shows every agent, every row, and then a color code, a little dot inside of it to show how many times, or the sequence of events, but that was pretty amazing, right? I think that view really gets you a sense of the size of what we're having here. And sometimes it's important, to show a lot of stuff in the same time, as long as easy, you know, you can consume it, but you see a lot of momentum, right? Cool. Okay.