# [[AI Weekly Update - 2026-04-28]] <iframe width="560" height="315" src="https://www.youtube.com/embed/qts_fkMtdps" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe> plugins:: "" Related:: "" [Video draft](https://share.descript.com/view/41MNOAEpGpt), sent to [[Mat Ryer]] and [[Tiffany Jernigan]] on [[2026-05-11]]. ## Pre-production ### Thumbnail ![[AI Weekly April 28, 2026.png]] ### Title 1. ### Hook ### Structure ### Outro ### Related videos and resources ### Pinned comment ``` ``` ### Timestamps 00:00 Intro ## Post-production - [ ] Create captions using Rev, then edit them. - [ ] Take a thumbnail photo and create one for YouTube (1280 x 720). - [ ] Add description. - [ ] Include the title and any keywords in the first few sentences, but in a natural way. - [ ] Add related videos and resources as links. - [ ] Add end screen. - [ ] Add cards if necessary. - [ ] Add chapters/timestamps for YouTube if longer than 5 minutes. - [ ] Add mid-roll ads/check monetization settings. - [ ] Add pinned comment. - [ ] Post video preview on Patreon. - [ ] Schedule on YouTube. - [ ] Add video to relevant playlist(s), or create a playlist if necessary. - [ ] Schedule on Mastodon. - [ ] Create a blog post on [site](https://nicolevanderhoeven.com). - [ ] Add to [video-database](obsidian://open?vault=obsidian-playbook&file=video-database%2FVideo%20database). - [ ] Add to other notes in my vault as appropriate. ## Transcript Speaker 1 (00:00): Welcome to the Speaker 2 (00:01): AI weekly, isn't it? Speaker 1 (00:02): Do we have any demos? Speaker 2 (00:04): Yeah. Speaker 3 (00:05): We can start with the demo if that works. Yeah. Yeah. Ben, do you want to dive in? Speaker 2 (00:08): That'd awesome. If I can share my screen. So you probably know we are working on a new feature called Workspace Home, which the goal of which is to allow Workspace to become the landing page of Assistant to replace and augment the experience that you've got at the moment that you can see in front of you. That work is ongoing and I probably won't be ready to show you just yet, but some of the bits that have arisen from it, I've started to integrate and I thought I'd show you a couple of bits from that now. Mostly I've been thinking about how we get this attract mode for the chat into the workspace. And also thinking a little bit about the different modes that Assistant supports. So we've got this dropdown that lets you do different modes and there's a cool little excellent that goes on with some of them. (00:57): So I've been thinking about how we can use that. I'll go into storybook to show you the actual components in isolation to begin with, but we've got a new little pickup for picking between the different modes. And what I've done is I've come up with the idea that these accent hues are now associated directly with a mode. The idea being that we can then use that as a touchpoint throughout the workspace UI. So the purple that we've got here for investigations is then going to be surfaced wherever we reference investigations within workspace. We don't yet surface any of the others yet, but when we get to that, we'll be able to do the same thing. And then scroll up. The track mode, a little quick demo if you're not really seeing the PR. (01:40): His surface is the same way there. Just some nice little sort of glow effects because now you're supposed to know it's an AI unless it's got a rainbow glow halo. I think it's mostly for now. It's mostly just a UI polished thing, but it's kind of giving you an indication of the direction we're going at. It's probably going to be a marmite choice, but there's a new icon for a system, which is this little for investigations where there's just little hat biases. Give me your feedback if you hate it. Give me a feedback if you love it. Yeah. Speaker 1 (02:14): Nice. Thank you very much. Speaker 3 (02:16): I can demo on small thing that I really like that Sonya worked on. She instrumented investigations with such a new dependency feature basically and that lets you see the agents of the conversation as a graph. So here, for example, you can see all the generations that are happening from the investigation lead and then you can go through the start of that basically and see that, for example, over here, the leaders having some issues. I think if we continue this way, it makes it really clear clearer than this one straight view how agents communicate with each Speaker 4 (03:03): Other. Regarding the dependency graph, let me show you a small one because you will see the probation of errors. Let me show you. Speaker 1 (03:16): Yeah. Why were some of those nodes ... By the way, that looks phenomenal. Why were some of the nodes red? Speaker 4 (03:21): Because it was failing. Yeah, this is a small one. So we have here an evaluation that is failing and you can see how this quality preparation is visualizing here. So all the downstream nodes are visualizing a warning with the apps in quality issu. This is because this evaluator is failing. And you can also drill down, for example, to a soup duck here. And here is because it's small, but with the previous one that goes bigger, it's like you have the ability to do something like this. You can move through this map. And for example, you can drill down here and in this case makes more sense and you can also do this. I'm finishing. I found this really useful because for example, you don't know this arrow where is the pattern. So you see here the toolkit, but you can click on here and you go directly. (04:33): You navigate to the parent now and that's it. Speaker 1 (04:37): Well, that's amazing. It is a great kind of visualization because at a high level, you really do just see where problems are and then drill in. It looks great. Okay. Well, that's all the time we have, but thank you very much everyone for another weekly. Please have a think about what you might like to demo next time. Thanks very much. Have a lovely rest of your Tuesday