Even though we still don't have a new frontier model from Google, well, they kind of went on a small tear this week in terms of new and helpful AI updates we can actually use today. I mean, their Spark agent now lives inside of Chrome. Google Docs got a huge AI glow up, and their new music model is pretty impressive. But Google Drops weren't the only new AI updates we got this week. We got helpful updates from ChatGPT and Chrome and Codex, a new Vibe design contender, and a new Slack-like platform that's mainly for directing agents from all providers. Yeah, you can't stop following the AI updates. I know this week was full of drama, open source, and you know, is the government gonna slow things down? You know, people from the big companies saying maybe we should slow down, but here's the reality you can't uh because these new updates are rolling out. And sometimes we think it's just the big new AI models that change how we work. I'd beg the opposite. A lot of times it's these smaller uh AI features that you might miss that might actually have the biggest impact. And that's why we do our Friday features show, where we tell you the best new AI features you can actually use today. These aren't waitlists, these aren't leaks, these are coming soon. This is like, hey, hit pause on this podcast, go log in and let's learn about it together. So, on today's show, here's what you're gonna learn. You're gonna know why Google is making AI noise in your browser docs and ears. You're gonna know the new Chat GPT browser magic you can cook up, and you're gonna know the new Slack-like platform that's made for your agents to work together, but with you directing them. All right, let's get into it. Welcome to Everyday AI. My name's Jordan Wilson, and well, this thing's for you. It's your daily live stream podcast and free daily newsletter, helping everyday business leaders like you and me not just keep up with AI, but how you can all actually get ahead to grow your company and your career. So if that's what you're trying to do, make sure to subscribe to the podcast, but also check out today's newsletter. We're gonna be recapping the highlights from today's show, as well as giving you the 101 of everything you need to know that's going on in AI world. All right, so let's get straight into it. A lot to cover today in our Friday features segment. Uh, first, a new contender in the Vibe Design Arena, and it's from Replit. So Replit launched their Replit design, and here's what's new it is a full creative suite that's replacing Canvas, uh, where the tool keeps offering you design variations as you work instead of waiting to be prompted. So it is directly built in. Uh, sorry, uh Mobin is directly built in, which is a reference library of 600,000 real app screens that normally require its own paid subscription. So you can upload a design system once and everything you make just snaps to your brand. So this is uh if you have a paid Replit account, uh, you don't have to pay anything extra, and it's actually free for every user with no tier gate announced, which is kind of unusual for something this model heavy right now. Also, that mobbin access requires no separate account and no additional costs. So here's kind of why it matters and who should care. Well, I mean, I think the target here is kind of the handoff, not like the drawing. So Replit is better at you know, is they're kind of betting that the export and rebuild step between design and engineering kind of disappears, and that's a direct shot at where Figma sits in the workflow. So they're just giving away also a paid reference library for free. That's a distribution play. Um, and the the quieter implication here, though, is the org design. So, like when a project manager can produce something that's publishable, the question of who makes the first version stops being so obvious. So, who should care? Well, founders, project managers who are shipping without a designer on staff. This is big, right? So I I do think, at least for Replit, this is one of the bigger uh missing pieces, right? Obviously, Replit is kind of a full stack vibe uh coding app, uh, but the design piece left something to be desired. A lot of times you had to use uh a separate uh you know project uh product and bring it all into Replit. So here is what Replit says about the new Replit design. They said today we're launching Replit Design, a new creative suite where anyone can shape what they see as fast as they can think. Replit design is the fastest way to go from idea to design. Ambient intelligence guides and inspires you at every step. And because everything lives on Replit, there are no handoffs, exports, or rebuilds that cause you to lose the beauty of what you created. You don't need to be a designer, you just need to know what you want to bring to life. Create with your choice of the world's leading models for design, including Claude, GPT-5, Gemini, Kimmy, and GLM. Uh, so I do hope that something like this becomes the status quo um in Claude Codecs, everything else, because it's actually really cool. Um, so a lot of times, right, the way I like to start kind of vibe uh projects is first I find inspiration, um, and then I'll usually upload that inspiration uh into either Claude Design or um I love uh Google's um Stitch. My gosh, like I I I speak so much AI and use so many. Uh yeah, is it Stitch or is it now I'm gonna have to look it up? Uh it's like, oh, I only use this like all the time. Yeah, Google Stitch, uh absolutely fantastic. Love Google Stitch. Um, you know, it just kind of gives you entire mock-ups. Um in, you know, it's powered by Nano Banana. And I also use GPT image um uh two, uh so images too. So I usually start and it's is usually kind of a multi-step process for me. Um so if you all if you are already a heavy replit user and you just haven't liked the visuals of what you've been able to get with Replit, I think the new Replit design is obviously a no-brainer. It's obviously not an additional cost. Um, you know, with the mob and library, uh, it's it's it's a pretty good kind of no-brainer setup. All right. Our next one, a small one for chat GPT, but at least for me, I'm like, okay, I'm gonna be using this all the time. So this is an update to the Chat GPT Chrome extension. So the Chrome extension now answers questions about any YouTube video, right? That's one thing I obviously do uh a lot of is watching YouTube videos in 2x or in 2.5x. I get a Chrome extension to make it go even faster. Uh so that's great. Uh, you know, that's one of the main reasons that I would still use the Ask Gemini uh tab inside of Google Chrome is just to uh quickly recap YouTube videos. So that is now in the ChatGPT Chrome extension, which is new, right? It's is completely new as of like three weeks ago. Um, also you have the um opportunity to now mention open tabs or pull highlighted text straight into the side chat. So if you don't, if you kind of missed the uh the new chat GPT Chrome extension, uh it's really good. So the new version of this came out um kind of alongside GPT 5.6 and uh ChatGPT work, kind of the super app. But uh essentially, and what I love is it's completely different than the old ChatGPT Chrome extension. Uh it has all of your conversations that are there. So that's something I love, even though I am using the built-in browser um a ton in you know chat GPT workslash codecs. Um, it's actually usually pretty busy, if that makes sense. Like I sometimes have like dozens or hundreds of automations a week that are using the browser inside of ChatGPT workslash codex. So sometimes because of that, and I'm like, okay, I'm really just straight up abusing Codecs ChatGPT. I'll still use Chrome for that reason, right? It's like I don't want to throw anything off. I'm just gonna let, you know, essentially my agents, that's their browser, and I still use Chrome a lot. So uh I do love the new ChatGPT extension, and these new features are really cool. So ChatGPT can obviously search your browsing history so you can ask it to find the page that you read Tuesday instead of hunting for it. Um, so who has access to this right now? So any desktop app users who are obviously you know have the most up-to-date build, uh, but the tab and YouTube features do require the Chrome extension specifically. So uh there's no plan tier restrictions for these. So uh yeah, you don't even have to be on a certain paid plan. Uh so why does this matter? Well, I think this is the uh kind of the assistant moving from a place you go to a layer that sits on top of what you're already doing. Um, so it's a different product category than a chatbot. Even for someone like me, right? Uh my codex is open at all times. Um, I love the new Chrome extension, right? Even just being able to watch in real time, like, hey, what is my uh, you know, where is my codex? This thread I'm at that's maybe an important thread that I'm watching, to just be able to click a button and see it without having to change, right? Especially when I'm on my laptop, um, you know, without having to, you know, kind of switch the flow of where I'm at, I think it's actually super helpful. So who is this gonna impact? Well, I think anyone whose research happens across 20 tabs, and you know, especially if you're watching watching long form video, uh, those kind of things. So anyone that's a researcher, if you're a student or if you're just doing heavy knowledge work uh inside Chrome, um, it's it's gonna be something that I think is gonna be extremely helpful. Um so let's get to our next AI update. And well, this is Meta, and Meta is making that shift slowly but surely uh from an AI chatbot to, well, something a little bit more agentic. So here is what is new uh with Meta's new AI tasks. So Meta AI now runs recurring tasks and daily briefings instead of just answering questions or pulling uh from your calendar to flag conflicts and also summarize what's coming. So you can set something up once, like a weekly plan or uh a restock alert or whatever, and it's just gonna run without prompting. So this is driven by the new Muse Spark 1.1 uh that is meta's uh agentic model built for tool use essentially. So who has access to this right now? Uh well, it's in select markets only, so this isn't a global rollout. Um, and it is just in the meta AI app and at meta.ai. So yeah, this isn't where, you know, in all the other platforms, right? WhatsApp, uh Instagram, all that stuff. So why does it matter and who's gonna use it? Uh well, Meta has spent a year like telling everyone it wasn't gonna chase productivity tools, but it just kind of did. Um, right. But I think ultimately, uh, you know, Meta realized that this is what people want, uh, right? People want uh to be able to set it and forget it, right? They want agentic scheduling, uh, right. This is for me one thing I do every single day, right? I'm always looking to get things on a routine or a schedule. Uh, literally, like my first thing I do is well, I check my overnight runs, you know, when I have my agents working third shift. Um, so that's what I do in the morning. Um, and then at least recently, right, I'm using ChatGPT voice and getting outside more, right? Like the weather's nice that right now in Chicago. So taking like more walks in the morning and just going through all my threads and making decisions right there with ChatGPT voice. But for the most part, if I'm not um, if I'm prompting, I usually think there's a problem, right? So if I am prompting, well, after I prompt, and you know, if I'm calling on multiple skills or whatever, you know, I'm usually going to turn that into a skill or a plug-in and then schedule that. So it is good to see Meta here uh, you know, finally offering something in terms of scheduling. So um Meta's actual asset is you know, it's personal context that nobody else has, um, you know, mainly your social graft. Um, so that's ultimately, I think, what's gonna be their moat. Um, and I think, you know, and who's gonna use Meta AI? I'm not sure. Um, right. Even when they, you know, came out with a new model and had it for free and all these features, it's like, okay, I go in there and they're like fine, they're okay. Um, but even, I mean, I don't know. Uh, just with the amount of uh free uh that you get from Chat GPT, uh, you know, everyone else is cutting the subsidies a little bit, right? Google used to be very generous, not as generous anymore, but still, I don't know. If you have a paid subscription, I don't think you really necessarily need to use the new meta AI, but let's at least uh talk about what's new. So here's what they said: you've been using meta AI to find answers, think through ideas and create content. Now it can go further and actually get things done for you. Meta AI can now make plans, follow through on next steps, and keep you on track without needing to be reminded or reprompted. Tell Meta tell Meta AI you're renovating your kitchen and it will learn your style, scout marketplace for furniture and fixtures that fit your budget, and send you a mood board so you can see how it all comes together. Ask it to help you build a running plan for your first half marathon and it will create a week-by-week training schedule, adjust for your availability, and share your plan for the week every Monday morning, or ask it to plan a birthday dinner and it will find restaurants, check your calendar for night that works, and suggest options. All right, so pretty cool. Am I gonna use it? No. Uh, but I don't know if you're a heavy uh Facebook user or if you don't have any paid subscriptions, maybe, right? The new Muse Spark 1.1 is actually a really good model. Um, and if you are, I don't know, maybe you're you know, you work in social media marketing, advertising, so you're always using the meta platforms, then it might make sense. All right, next. Yeah, now we're getting into the Google portion. Uh, three straight Google updates for everyone. So the first, a big Gemini, uh Google Gemini in docs glow up. So Gemini can now generate and edit images, diagrams, and infographics inside of Google Docs using what you've written as context. So it also works your comment uh threads, summarizing them, drafting replies, leaving new comments for you, and rewriting sections based on review or feedback. Uh so who has access to this right now? Well, you have to be on a paid plan. So either a uh work uh workspace plan, standard or plus, or just an individual, you know, plus or AI Pro Ultra plan. So image generation, though, is web only, and the comment features require edit access to the document. Uh, the rollout is gradual, although it is already starting to roll out to those tiers. So um, why does it matter and who should care? Well, this is like the boring productivity win that actually produces ROI, right? Drafting was never really the bottleneck in document work, especially when we have like AI everywhere. It was more of like the review cycle and you know, those 40 unresolved comments, right? That's where the project sometimes stalls. So it's good to see that uh, you know, Gemini is starting to, you know, infuse more of its AI into Google Docs. I would say, um, you know, of all the Google products, I'm very heavy in docs and sheets. Um, so I do think Sheets has been a little bit more ahead of the game in terms of how they've worked Gemini into the fold. I think Google Docs has been a little bit behind. So this is pretty uh, you know, a pretty good uh update in my mind. So who should care? Who's gonna use this? Well, I mean, uh, especially if you are a heavy uh Google Docs organization, leaving a lot of comments, going through working with multiple versions of documents, that's gonna be good. Um, but I I think one thing that could be good is people who are still just writing old school blog posts, right? And maybe you need visuals, uh, you know, working in nano banana into Google Docs, uh, that's pretty nice, right? There was a an AI tool I used to use a lot called Napkin, right? So it's kind of like what I think of, right? Where you can put in a bunch of text and you know make some visuals and um charts and graphics, right? So you can obviously do that with Google's nano banana, but also images as well. So, you know, I think if you um you know you know really are comfortable in the Google Docs interface and you do need some of those other features. Well, now you don't have to do a bunch of copying and pasting and context switching. You can just get it all done in Google Docs. All right, next, staying on the Google train, choo choo, uh, we have their new uh always on 24/7 agent Google Gemini Spark is now in Chrome. So Google's always on Cloud Agent Spark now plugs into desktop Chrome. And if you do give it permission, it can use your logged in accounts and saved passwords. And that's the big thing right there, right? Like that's one reason for me. Um, I absolutely love the new codex slash chat GPT work update because it does integrate all your Chrome passwords, and that's huge in terms of what you can actually actually accomplish um with an agentic browser or an agent that uses a browser, right? Um, because if you're you know, your first go-to is always going to be using a direct integration uh or an MCP, but you know, so many things, at least that I use, don't have a direct integration into you know, Claude Code or um Chat GPT, and a lot of them don't even have an MCP. So then it's well, give the agent your credentials if you trust them and if you're allowed to, and we'll have it do the work for you. Uh so that's a pretty big update there from Google. So let's just talk. Uh, this is from their uh their page. Uh so they say today we are expanding Gemini Spark to more users around the world and adding a new feature to help you handle complex web errands. Here's a look at what's new. So they say handle web errands with Chrome Auto Browse. So they say we're taking Spark's web browsing capabilities even further by integrating directly with Chrome. Now, with your permission, Spark can use your logged in accounts and saved passwords to handle tedious web errands like scheduling viewings for apartments, your uh saved or researching flight options and starting the booking process. This is all done while protecting against threats like prompt injections and keeping you in the loop on sensitive actions such as payments by handling by handing the task back to you. These capabilities are initially rolling out in the US with plans to expand to additional regions in the future. So there you go. So right now it is starting in the US with paid plans. So yeah, if you're a heavy um Chrome user and you have a lot of repetitive tasks, I don't know why. Uh, right, it all the big companies do this, right? People always show the demo. It's like, oh, you know, adding things to your Instacart app or you know, booking tickets. I don't know. For me, for those things that are ultimately requiring a transaction, I don't necessarily want to go agent on those, right? At least not to the point of adding my cart, maybe find me the five best trips. Sure, that's fine. Um, but I don't know for me, like I think it'd be really cool if all the companies would stop using uh you know personal errands, um, because this is actually a big, big step forward uh for Gemini uh in Chrome and the Spark agent. And like we talked about, I think it was either last week or the week before, right? So the Spark agent is now here, uh, right, which is really cool. So before it was only on the Ultra Plan. So now if you have a paid uh Google Gemini plan, well, you have Gemini Spark, which is you know essentially their version of like an open claw um or like a you know just an agent that lives in the browser, right? It's browser-based, not desktop based. Um, but I don't know. For me, I would have liked to see some uh more examples and use cases on using this for work. But anyone that's drowning in scheduling, booking, uh comparison work, right? If you're in security or IT, uh you know, anyone comparing agentic browsers, also, right? Like if you are a heavy Chrome user, or maybe you can only use Chrome, right? This is a great way to marry um, you know, your Google data because that's the other thing. It can access and understand anything in your Google Drive, your Gmail, et cetera, and then use all of that information on the web, right? So, yeah, things like filling out uh forms, right? That's something I would love to use something like this for, and especially having it directly connected uh to all your Google data. All right, our last Google one is Liria three. So this is Google's new AI um music maker. So yeah, if you know like Suno or Yu Dio, uh Liria now 3. Is out in new from Google. So this is Google DeepMind's newest music model with what they say is better melodies, lyrics that follow the prompt, more expressive vocals, and direct control over tempo and length. So vocal delivery is the specific target, which has been the thing that kind of gives AI music away. So who has access? Why does it matter? Who should care? Well, it's available in Flow Music. All right. So uh I think I hope eventually they'll bring this to Gemini. Um, you know, it previously when they rolled out, I think uh Liria 3, it was available in Gemini uh on launch day. So this time it's not, it's only in Google Flow Music, which is actually a really cool program. It's kind of like a vibe music, right? You just kind of chat with the chat bot and you can make iterative edits, which is really cool. Um, so it's available in Flow Music. Uh, you do have to have a paid uh Gemini account to use it right now, and then also uh it has commercial use rights that ride on those paid tiers. So uh Google says it does not claim ownership of the original content that you generate. So the story here is licensing, not necessarily audio quality, because Suno and Yu-Gi-Oh are both in active Sony music litigation and Google trains, uh they say on licensed data. Uh so this is the third, though, meaningful Liria release in like six months. Um, that tells you that Google is treating AI music as an actual product line, not just you know, something they're toying around with. Um, so who should care? I mean, if you're a video team, a podcaster, right? A creative, uh, you know, marketer, advertiser. The the cool thing is that I didn't know is well, the songs are like three minutes, uh, right. And the songs are actually pretty, yeah. Okay, I won't say good. The lyrics aren't good, all right, but you can, you know, use a different program to to bring the lyrics. Um, before I started recording here, I I tried, I don't know, five, six, seven songs, and like the cohesion was great, right? I was having it blend like two extremely different music genres, and I'm like, okay, it actually did a really good job, right? And the vocals were um on beat, they were on point, even when I was throwing in these mixed genres. But sometimes I won't say the lyrics were nonsensical, but the lyrics were just lyrics, right? Um, and I think Suno does a much better job. Like I said, obviously, for any AI music platform, you can provide your own lyrics, but a lot of times you don't necessarily want to. Um, you know, with a little bit of prompting, they get a little bit better, but I don't know. To me, the lyrics were still just kind of um slop um or slop-ish, right? Better than they were, uh, but still not great. Although they sound really good, right? Like the different, you know, I tried out all kinds of different genres, like it sounds like good music. And then when you listen to the lyrics, you're like, this is a little sloppy. Uh, but overall, uh, if you want to do your own lyrics and if for whatever reason you need uh, you know, audio uh music generated for something, uh, pretty, pretty good to check out. All right, last but not least, and this is one aside from like mentioning it once in our newsletter. I haven't talked about this, and I've actually been enjoying uh kind of toying around with this. It hasn't, at least for me, gotten uh, I don't think production ready yet, but it very well could be just because the company that's behind it. So here's what's next. This is Buzz from Block, uh, the parent company uh for Square. So this is kind of like a Slack lookalike that's built for teams to essentially talk with agents from any provider. All right. So yeah, you can bring your whole team in, both all your humans. So it is like Slack, it has channels and at mentioning and all those things, but essentially you can bring all your different agents in too. And then you can, you know, like as an example, you know, just like you would on Slack or Microsoft Teams, you can, you know, put in, you know, you can bring in your skills, which is really cool. And then you can, you know, have, you know, as an example, codecs start on one thing, hand it off to Claude, they go through, then they hand it off to, I don't know, an open source model, have them work, collaborate, talk amongst themselves. So the thing here it's really nice. I don't know if I'll actually use it. And here's why. Maybe it's more of like I've spent so much time getting this very thing set up in Codex, right? I use like I've created bridges, you know, from Codex to Claude and vice versa to other, you know, other model providers. Um, so at least for me, I'm usually working with multiple model providers in Codex or Cloud Code. Um, but I mean, I still have been having fun experimenting uh with Buzz. It is an open source project. So uh it is, you know, every participant, whether it's a human or an agent, they get their own cryptographic identity, which means that every message, code change, and approval kind of leaves a signed trail. So agents join as members with scoped permissions, so that's important, not as bots like waiting on a slash command. So it's free and open source under uh Apache 2.0. So you can use blocks uh hosted version, you can download the desktop app, or you can self-host it entirely. Uh, I do believe they also just came out um with an iOS version and an Android version. So there's mobile support and uh out of the box, it works well uh with Claude Code, Codecs, and uh Blocks agent as well. But there's different ways I think you can use like through open router that essentially you can bring any model in. So I think this is like an answer to kind of not an answer, but a potential answer to agent governance um for smaller teams at the infrastructure layer rather than at the policy layer. And like that distinction's kind of like the whole point, um, because most companies can't just tell you which actions in their logs came from a person and which ones came from an agent. So this one is kind of like leaving that agentic agent trail, right? It's leaving saying, hey, these agents did this, the human said this, right? So obviously, if you're using like Microsoft Copilot, I think they're actually kind of uh a leader in the space in terms of that um uh kind of large teams working with a lot of uh co-pilots and being able to trace, uh, you know, whether it's through intra ID or something else, being able to trace every single thing. Um, I think some of the other players have a little bit more to keep up on. So it's nice that a new open source platform like that has that built in and as a core feature. So uh here is uh what Buzz says. Uh so uh technically Block says about Buzz. So Block said today we're releasing Buzz, a free open source collaboration platform where humans and AI agents work together in a shared workspace. It's available now at Buzz.xyz. Over the past two years, we've built and open sourced a series of AI tools at Block. Through all of it, we've come to the realization that the most productive work doesn't happen when someone asks AI for help. It happens when humans and agents are in the same room working on the same thing with shared context. There was no platform design for that, so we built one. Um, so it's I think, right, it's open source if you already have like someone like me or larger teams that they're generally using a lot of different subscriptions uh and you have different skills or preferences of using certain models for certain purposes. I think it is worth exploring to see if it can maybe be a stopgap. I think eventually, um, I think we're gonna see this type of um interaction probably in codex first. And I would assume that when Microsoft comes out with their super app, which would be uh, you know, any day, any week now, I uh I would think that uh Microsoft will eventually have some collaborative features. Um, I would say the same thing, obviously, for Claude and Gemini. So I don't think this is like where you're, you know, you should throw away your current harness or anything like that. But for smaller teams or for project sprints that you know that you're gonna have to work with multiple teams or multiple team members that are using different harnesses, different models, it actually might be good for instances like that. Scope it out first. This is brand new, but it looks promising so far. Like I said, for me, I don't think I'll be using it. We're a small team, um, and I'm so codex pilled, right? And you know, I like I said, I've built the bridge from Codex to Claude and Claude to Codex. Um, so I don't necessarily, aside from like being like, oh, this is fun, this is cool, I don't necessarily have a good use case for it, but I think for uh many of you out there, it's gonna be something definitely worth looking at. All right, that is a wrap for the seven new AI features you need to know. So a quick update. Replit design is out, it looks really nice. Uh, chat GPT desktop and Chrome extension updates, allowing you to uh summarize YouTube videos or mention at mention uh other tabs and pull info from there. Um Meta shifting from just an AI chatbot with Muse Spark 1.1 to just schedule a genetic model that gets work done for you. Uh Google with a ton of new updates, uh, some new updates with Gemini in Google Docs, um, their new uh Gemini Spark integration in Google Chrome and the updated Liria 3.5 uh model. And then uh last but not least, we just have Buzz from Block. A good way for uh you and your team to talk to many different agents and I don't know, just have them fight it out and watch it and eat popcorn and have fun. All right, that's a wrap, y'all. So uh if this was helpful, yeah, we do these kind of themed shows a couple of times a week. On Friday, we bring you Friday features. These are things, AI tools and updates you can actually use now. On Monday, we go over the AI news, y'all. I'm telling you, this week was crazy in AI news, right? Not just like, you know, it wasn't necessarily about model releases, it was about agents going loose, you know, laws, regulations, nuts. And then on Wednesdays, we do hands-on demos called uh putting AI to work on Wednesdays, and Tuesdays and Thursdays, we rotate our shows. So I hope this is helpful. If it was, please do me a favor. It takes you like 10 seconds. Uh, make sure to subscribe or follow uh the podcasts on Apple Podcasts and Spotify. Then go to your everydayai.com. Thanks for tuning in. We'll see you back next week and every day for more everyday AI. Thanks, y'all.