Ep 753: Anthropic Goes Full OpenClaw, Meta Muse Spark Drops, Google Gets Notebooks and More . 7 New AI Features You can’t afford To Miss

This is the Everyday AI Show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business, and everyday life.

Anthropic continues to add new features that look and function kind of like OpenClaw, the same platform they pretty much shut out last week. Google just dropped a notebook feature that's kind of like NotebookLM, but kind of not at all. But it also adds new features to both platforms. We got new models from Meta, Microsoft, we got AI video updates, and a ton more. But the most meaningful updated AI feature that you may have missed this week comes from a company we haven't talked about in a while. Zapier, and I think that one will be big. All right, let's get into it. If you're new here, welcome to Everyday AI. But on today's show, if you stick with me for the next 20-ish minutes, here's what you're going to learn on our new Friday Features segment. You're going to learn how Anthropic is going full on OpenClaw after banning its models from being used by OpenClaw. If Meta's new Muse Spark model is worth the billions they've invested in, you'll learn what Google's new notebook feature in Gemini unlocks. And last but not least, you'll see why I think Zapier's new SDK might be your favorite tool, even if you're not technical. All right, welcome to Everyday AI. My name is Jordan Wilson. If you're new here, we do this every day. This is your unedited, unscripted guide to keeping up and getting ahead with AI. So if you haven't already, please make sure to go to youreverydayai.com, sign up for the free daily newsletter. We're going to be recapping today's show and all the other AI news that you need to know to stay ahead. All right, so there's so much that's happening in the world of AI, y'all. I have a daily AI podcast from Monday to Friday at least, and I realized after like 700-some episodes that still 80% of all the actual useful stuff that we were all using, I wasn't covering. So that's what this new Friday Features is all about. So on Mondays we do the AI news that matters, on Wednesday, we go hands-on, you know, with a new model or release. Last week we did the— or this week. Wow, time flies. This week we did Google's Gemma 4, the local model. So make sure to go check that out. And then on Fridays, well, we're going to fill in the gap in between because I realized that between the Monday AI news and the Wednesday deep dive, most of the things that we talk about aren't getting covered. And then on Tuesday and Thursday, we do different shows, rotate, sometimes interviews. All right, but let's get into it. There's a ton to cover this week. So, livestream audience, bringing up my screen here. Podcast audience, you can always see the video version at youreverydayai.com. Go click on episodes, not only today's, but literally, you know, 750+ videos, podcasts. You can read it. It's all on the newsletter for free. It's a free generative AI university. All right, first, let's start with the one that I think is I'm the one that was— this is honestly probably the one I was most excited for, even bigger than the Anthropic Agents one that we're going to get to in a little bit. But first, let's talk about the Zapier SDK. So here's the way that Zapier explains it. They say, let your agent connect to anything, authenticated, governed access to the full Zapier catalog in code on behalf of your users. No OAuth flows, no token management. Zapier handles the keys. So here's what that means. Zapier opened its SDK to everyone, including giving coding agents like Cursor Claude and Claude Code and Codex programmatic access to Zapier's full ecosystem of 9,000 apps, 30,000 actions, and raw API access to 3,000 apps. So here's the good thing. If that— if you heard that and you're like, well, what the heck does that even mean? You know, OAuths and raw API, you're like, wait, what? This isn't a show for non-technical business leaders. Yes, it is. So even if you don't know what most of that means, well, you can just use natural language to set it all up. So you will need to use a coding agent to obviously tap into this SDK. So that's just a software development kit, which used to be kind of technical, but in the age of AI and large language models and natural language processing, it's not actually very hard at all. But what this means is, let's say as an example, if you use Claude Code or if you use OpenAI's Codex, right? Up until, well, before this Zapier kind of announcement, what you could connect to was extremely limited, right? Obviously, Anthropic's Claude Code has a growing list of MCPs. With Codex, they do have some plugins, not a ton, but now essentially you have Zapier. So you have everything, right? So your agent or different agents that you use can now access your everything from your Gmail to your CRM to Slack to your project management tool, like whatever you're working on. So I haven't started to set this one up yet because it literally just came out like like 48 hours ago. But this one I think is pretty big. So who has access right now? I actually know, I will point out, I watched the video with Wade Foster, their CEO, and he called it the most powerful thing that they've launched in years, which I think was pretty telling anyways. So this is an open beta right now and it's free during early access with no billing charges. So that's pretty cool. And then enterprise and teams plans are off by default and they do require manual opt-in. Or contacting Zapier. So if you have a personal plan, this will be a little bit easier to enact right away. So here's why it's useful. It just eliminates the single biggest friction point for AI agents, and that's managing all the authentication right across, you know, thousands of apps. You're probably not going to be using this across thousands of apps, but you'll be using it maybe across a dozen or dozens of apps, right? It's difficult to manage all of that, but Now Zapier just kind of does that. And that's where Zapier's MCP gives agents a curated menu of pre-built actions. And then the SDK lets agents write loops, handle edge cases, and chain complex logic across apps. And your existing— the cool thing is your existing Zapier connections work immediately. So you don't have to re-auth anything. So once you do authenticate the Zapier SDK in your coding agent of choice, anything that you've previously connected It works right away, right? So I'm like, even as I'm saying this, I'm like, oh frick, I should have prioritized this over other things because, you know, with Zapier, you know, I have our WordPress website set in there. I have our Beehive email newsletter. I have our, our Circle community, right? It's all there. So now I can instantly start working with those things inside of Claude Code and inside of Codex are the two that I use the most. So Here's who I think is going to find it valuable. Well, developers building AI agents, that's going to be extremely valuable. You know, automation experts. So if you are someone or, you know, anyone in marketing that has used Zapier for years, you're going to find this extremely helpful. I also think solopreneurs, entrepreneurs who maybe don't have the time to usually do all of these things. Well, now it made it a lot easier with this from Zapier. All right, let's go to our next one. Doing something a little different. OpenClaw updates, right? And hey, let me know, livestream audience, um, podcast, uh, you know, if you're listening on Spotify, if I should make the Friday feature, right? If we're going to do this permanently, should I just always— like, we've been doing essentially 7 things. I think that's easy, right? It's not too many, it's enough. But should one of them be OpenClaw every week? Um, if so, right, because OpenClaw ships updates like every day. So let me know, just say OpenClaw in the comments if you think I have a number, right? I actually do this. I always have a number and I say between LinkedIn and Spotify, if we get enough comments on this, we'll do it. If not, we won't. So if you want to always include the weekly OpenClaw update, let me know, but I will do it this week regardless because I think it's pretty big because now in OpenClaw, we got built-in video generation, built-in music generation. We got some new memory features and an experimental kind of dreaming mode. All right, so, well, if you don't know OpenClaw at all, I don't know, you must literally be a lobster sleeping under the sand somewhere. So OpenClaw is the world's, technically the most popular piece of open source software ever. It is an autonomous agent that acts on your behalf, right? You can talk with it, communicate with it via various channels, but let's get into the actual new update. So we got 3 major releases. So one is built-in video and music generation. So these are obviously all things that you pay for, right? You have to connect, you know, your API keys unless you're using an open model. So if you do have a, you know, super fat like Mac Studio like I do, you'll be able to do some of these things for free or your agent will be able to do them for free. But otherwise you will be using APIs, so you'll be paying for your usage. Keep that in mind. So the new things, we got the built-in video and music generation just from text, right? Video, music, and editing tools, and then the persistent memory, kind of like a wiki knowledge system. So the video generation, the agent can create videos and music tracks directly using configured providers like Runway, Google's Lyria, et cetera, without even leaving the chat. Then music and video editing. Agents can trim, mix, splice, and refine existing video. That's really cool. And music files through natural language commands. So that's pretty sweet, right? Turning OpenClaw into a creative editing assistant. You know, this is the first versions, right? I'm not going to try that one, you know, until it's gone through some updates. You know, some of these things also keep this in mind. The OpenClaw team is amazing, right? There's a reason why it's literally the most popular piece of open source software ever. But for some of these things, when they first are released, they're usually rough around the edges. That's how it goes, right? And then the open source community, you know, finds fixes. You know, they ship them out pretty fast. But, you know, if you need something like this for production for like a work project, don't think that you're going to get anything, you know, usable or great. But it is important to know what they're working on and what they're shipping. All right. So that is the music and the video editing. And then there's the Memory Wiki. So This kind of replaces the old fuzzy recall system with a more structured kind of persistent knowledge base that works a little bit more like Wikipedia. And it does support Obsidian-compatible export. So if you do use the Obsidian app, which I know a lot of people do, that will be helpful as well. And then last but not least, there's a new dreaming experimental mode. So this is opt-in. And this is essentially now where the agent processes conversations during idle time through light, deep, and REM sleep. Style phases to consolidate short-term memories into permanent knowledge. So pretty, pretty cool. And then it kind of has a diary timeline user interface. So obviously, who has access to these? Well, everyone, right? So as long as you're updated your Open Claw to the latest version, you will have all of these features. And the cool thing is obviously they added GPT-5.4 support. Recently Gemini Claude only via the API because obviously Anthropic cut off access to using your Claude paid plan via Open Claude, which we talked about on the show last week. So I mean, here's why it's useful and who will find it valuable. Well, I don't think some of these things are going to be super useful now, but they will be very useful soon. So I think specifically if you work in anything creative, right? The fact that you now have video in music generation is pretty big, you know, Runway support, Suno, you know, ComfyUI, whatever you use. And then the same thing with Google's Lyria. So if you work in any creative industry, I think it's going to be good to just start, you know, toying around with this. I'm sure it's going to get much, much better. But then the memory in the the memory, the memory wiki solves a lot of, you know, the common problems that people have been having with OpenClaw. Even if you have your, you know, your SoulMD file set up, your heartbeat set up, everything correctly, still the memory piece is pretty tricky. So the new memory wiki is the next attempt to solve that. So who's going to find this valuable? Like I said, content creators, great. If you just want, you know, quick and dirty, you know, videos, developers building AI agent workflows, obviously, and then power users and tinkerers who are kind of using OpenClaw as a primary assistant, right? And you want it to remember more. The new memory features are their attempt to do just that. All right, next, speaking of remembering things, we have a new AI moves too fast to follow, but you're expected to keep up. Otherwise, your career or company might lag behind while AI-native competitors leap ahead. But you don't have 10 hours a day to understand it all. That's what I do for you. But after 700+ episodes of Everyday AI, the most common questions I get is, where do I start? That's why we created the Start Here series, an ongoing podcast series of more than a dozen episodes you can listen to in order. It covers covers the AI basics for beginners and sharpens the skills of AI champions pushing their companies forward. In the ongoing series, we explain complex trends in simple language that you can turn into action. There's 3 ways to jump in. Number 1, go scroll back to the first one in episode 691. Number 2, tap the link in your show notes at any time for the Start Here series, or you can just go to starthereseries.com, which also gives you free access to our inner circle community where you can connect with other business leaders doing the same. The Start Here series will slow down the pace of AI so you can get ahead. Notebook integration in Gemini. So it's technically a NotebookLM integration, but not the same as the one that they added at the end of 2025. All right. So let me explain like what that means. So what they added at the end of 2025 is you could add your NotebookLM kind of notebooks as a source inside of Gemini. So this is a little different. The new notebooks feature is more of kind of like a folder, but not the folders that you would normally use inside of like ChatGPT or Claude. These are more like a NotebookLM folder, but that you can use in Gemini. NotebookLM, and then it syncs. All right, let me kind of read the official word here from Google. So it says notebooks in Gemini help you organize your chats and projects in the Gemini app with our AI-powered research partner NotebookLM for easier learning and working. All right, so here's kind of the gist of what it does. So it says with notebooks, you can keep conversations about a topic organized in one place. Just click new notebook on the side panel of the Gemini app to get started. You can move past chats into notebooks. So that's, that is good. You can organize your past chats into notebooks and then give Gemini custom instructions and add relevant files like documents and PDFs to give Gemini more context. So in that regard, it is like standard, you know, ChatGPT projects and Claude projects. But here's where it's not at all, because all of those notebooks technically sync to Notebook. LM, which is really cool. I probably, which is crazy to think, um, outside of Google Gemini's deep research, because Gemini is everywhere now, right? I use AI Studio a ton and I use Notebook LM a ridiculous amount. It might be one of my most used tools aside from like GPT-5 Pro. Um, and I am using like most people Claude more and more, uh, you know, with all these new features that come out, but I actually am not using the I'm not using the normal Gemini app to chat as much as I used to be, right? I'm using it a lot more for deep research. I think Google's deep research is still goated. It's so good. But now with this, I'm like, okay, now I'm going to have to start doing more of my chats inside Gemini, right? Because this feature to sync all of those, put all of those in a notebook and have that be synced to NotebookLM, right? You have all of these sources that you bring into NotebookLM, if you're a power user, or maybe if you don't know, NotebookLM is a very unique product from Google. It is powered by Gemini, but it's grounded. So that means the hallucination rate is essentially zero, right? Where all other large language models, you know, you're constantly having to be vigilant against hallucinations, right? So this is so cool now that your chats inside Gemini can become your source material for Notebook LM just by adding it to a notebook. So that's really cool. So like I said, you do have that persistent project space where you can organize chats, upload documents and PDFs, and give Gemini custom instructions for those ongoing products. And the notebooks sync automatically, which is really cool. So if you add a file, you know, in Notebook LM, that file will be added in Gemini, which is so, so cool. I'm pretty sure I messaged one of the NotebookLM leads something like this, like a year ago. They did add one of my features that I requested finally, getting on the mobile app, you know, plus and minus, you know, plus and minus 10 seconds. I asked for 15, they did 10, but You know, can't complain. So who has access? So right now this is rolling out to paid subscribers first. All right. So if you're on a free plan, you know, maybe this will roll out eventually. I'm guessing it probably won't. I could be wrong because I'm guessing this is going to be a more of a compute-intensive feature, especially if it's popular. So it is rolling out to paid users on the web this week and then mobile access and European expansion in the following weeks. And free users said they'll get access after all the paid rollout completes. So no clue when that is because sometimes it does take a little bit longer for all of the paid customers in Europe to get access. So sometimes those, you know, free users getting access gets delayed a little bit. So here's why this is useful and I think who will find it valuable. First of all, literally anyone is going to find this valuable, right? I don't care if you're a researcher, student, general knowledge worker, if you're not using NotebookLM literally every single day, I don't understand why not, unless you're not able to because of your job. But I mean, use it for your personal life. If you have a personal Gmail account, even the free version of NotebookLM is can't miss. It's so good, right? But here's why I think that this new update's useful. It solves that scattered conversations problem, right? Obviously, if you've used projects in ChatGPT, or if you've used projects in Claude, especially if you've used like project memory inside of ChatGPT, you'll understand how useful it is to essentially— that way you're not wasting all of that back and forth, right? So I think sometimes when we're working with large language models, you know, ultimately we just use whatever that last piece is, right? It's the last thing that you copy and paste, right? Or maybe you're going through a lot of iterations of something just to get a couple of facts. But what about the other 70%, the other 80%, the other 90% could all be really good stuff. So now this lets you automatically use that inside of NotebookLM. That's probably the way I'll be taking advantage of it most is, you know, moving from Gemini to NotebookLM, not vice versa necessarily because I am a power NotebookLM user. But regardless, this is pretty freaking sweet. All right. We have a couple more. All right. Next we have Muse, Spark, Meta, Meta after billions of dollars, quite literally, right? They just had a single $14 billion acquisition of Scale AI, essentially an acquihire. And then they reportedly were spending hundreds of millions on individual researchers. And then over the past year, they've done nothing with Llama. A year of silence in 2025 and 2026, it's like a decade. I think people forgot Meta existed. They came back with what I will say to their credit is a fairly impressive model in Muse Spark. So here is Muse Spark. This is their new model, but the first model from the Meta, the new Meta Superintelligence Lab, or MSL, led by former Scale AI CEO Alexander Wang. And it is a natively multimodal reasoning model with text, image, and tool use. But Unlike the previous Llama models, this is a complete rebuild. So this is not an open source model that you can download and fork locally. Right now though, it is free. So free, but not open source. All right. So it uses a multi-agent architecture where parallel reasoning subagents tackle part of a problem simultaneously, reducing latency on those complex tasks. So who has access right now? Like I said, it's free to use on meta.ai and in the Meta AI app across Facebook. Instagram, and WhatsApp. There is also a thinking mode that they call contemplative, and then there is a private API preview right now, but it's not public. It's only open to select partners. So I mean, why it's useful, I mean, you get it, right? I don't know if you're, if you're a heavy Meta, Facebook, WhatsApp user and you like AI, that's cool. You're going to have a much better AI assistant to chat with. Will this be a model that's good enough for businesses to replace their stack? I don't know. Probably, probably not. Right. It's actually really good. Right. The benchmarks for technically being a first model like this might sound crazy because, yes, it is technically their first model, but it's not right. They had Llama before, but this is a complete rebuild. So if you give it the benefit of the doubt of being a complete rebuild, it's actually one of the best first models ever. But is it a top 3 model right now? No, it's not because it's still behind Gemini 3.1 Pro, Opus 4.6, and ChatGPT's GPT-5.4. However, I have been using it. It's really good. It's good at coding. Its ability to write is pretty good. It's obviously a huge step ahead from Llama. So like I said, if, if I think you're going to find value out of this if you are already using Meta's products on a daily basis, or if for whatever reason the certain benchmarks really just kind of hit what you're, you know, what you're trying to get out of it. Like I said, we actually— and I will share it, I'll reshare it in today's newsletter— I put together a chart comparing Muse Spark with its biggest competitors. And then also, just for fun, I put the, the Mythos model from Anthropic that no one's going to get access to for the most part. So yeah, if you want to see how it ranks on the benchmarks, I'll have that in the newsletter. All right, next, Meta wasn't the only company with new models today or this week. Microsoft also did. Yeah, this one, I kind of came out right at the end of last week. I'd already, you know, planned the show out. So this one is like 8 days old. All right. But stuff happens fast, right? So this is from Microsoft. I'm just going to read their little blurb here. So they said, introducing MA1, or sorry, MAI Transcribe 1 alongside MAI Voice 1 and MAI Image 2, what they call world-class quality at lightning speeds, now available at the most competitive prices. And it is available now in the Microsoft Foundry and the M AI Playground. So I will say if you want to, you know, compare everything on the image side, right, it's— I think it's a top 3 family of models. It's not a top 3 AI model, image model, but, you know, for their first technical model, it's fairly impressive on the image side. And then on the transcribe or the voice side, it actually got some pretty good benchmarks here. So it was the lowest for mean word error rate, right? So when transcribing, it actually is the top model for that. So if you are building something with voice or transcription, it might be a model worth looking at. But some more details. So this is the first big series of releases from Microsoft's new in-house AI team, now led by Mustafa Suleyman, and they released these three 3 proprietary foundational models. So MAI Transcribe 1 claims the lowest word error rate, 3.8% across 25 languages, beating the likes of Whisper, Gemini Flash, and ElevenLabs Scribe. Then you have MAI Voice 1. I don't know why the MAI is just hard to say. We had MAI Voice 1 that can generate 60 seconds of natural audio in under 1 second on a single GPU. So that's text-to-speech model or text-to-audio model. Very, very fast. And then the MAI Image 2 debuted as top 3 on the Arena AI leaderboard. So these are all available right now, right? So you can access them on the Microsoft Foundry or the MAI Playground. So if your company, for whatever reason, right, if they're not a big Microsoft shop, maybe you're a Google Gmail shop, or you just don't have access to the Microsoft Foundry, Foundry, which is kind of there, you know, formerly the Azure AI Foundry. Well, you can just go sign up for the MAI Playground, right? That's the way I've been kind of playing around with these models in the same way that, you know, OpenAI, Claude, etc., you know, they have backends essentially for developers. That sounds daunting. Trust me, it's not. You just sign in with your credentials. You usually have to connect a card credit card, because then at that point you are charged. Even if you're just playing around with it in the sandbox, you are charged. So, well, here's why it's useful. I think if you have already built or your company has already built transcription or voice products, it's worth looking at it just because at least I do think on the transcribe and voice side, it's fairly impressive. Or maybe for whatever reasons, maybe your company has wanted to use AI images, but you are locked into Microsoft. In this case, this is their first image model that I think is definitely worth, worth using. So who's going to find this valuable? I think enterprise developers who are building voice agents. If you need accurate, fast, and affordable transcription at scale, that's one great use case. Or, you know, marketing teams who need high-quality AI image generation and you only have access to Microsoft, right? I obviously wouldn't choose this, at least not now, over the Nano Banana, over the GPT-1.5 image. We talked earlier this week that, you know, there's rumors OpenAI is going to be coming out with their next image model. So it's definitely not the best, but I have seen some of the examples so far and it looks fairly impressive. All right. We have two more quick ones for you here. So our next one, Google Vids update. Stick with me here. Why does this matter? Well, you probably know Google Veo, right? And if you're on a paid plan, then you have access to Google's extremely impressive AI video generator in Veo 3.1. Now it's free. Well, a version of it. And if you use Google Vids, so here's what's new. So Google Vids now includes free AI video generation powered by VO3.1 for anyone with a Google account. All right. You do only get 10 clips per month, but y'all, if, if, if you don't remember, like when VO3 came out or even VO2. Right? But when VO3 came out, it's like everyone's like, oh my gosh, I'll pay a million dollars, you know, for this. Because at the time it was so good and it was, you know, so far ahead of, you know, the original Sora and so far ahead of all the other, you know, Chinese models. Now that gap has been closed. But the fact that now we're saying that you can get VO3.1 for free, uh, it's pretty impressive. So let's talk a little bit more about how this works. Well, one of the big features of Google Vids— I did do a dedicated show on that, uh, couple of months ago, but it has AI avatars that can be directed via text prompts, placing custom scenes, and dress to match your brand and interact with products and props, which is really, really cool. It also has a new screen recorder and— sorry, a new Chrome screen recorder and extension to direct YouTube publishing as well. So here's who has access. Like I said, free, free. VO 3.1 free. How many times can I say that? So up to 10 clips a month, even if you have a non-paid account. Also, you have custom music in Lyria 3 and AI avatars, but that is for paid subscribers only. And then Ultra subscribers get 1,000 VO video generations per month. That's obviously on the $200+ plan. Plan. So here's why it's useful, and I think who's going to find it valuable. Well, it's useful because it creates really good, uh, VO 3.1. It generates 8-second video clips from text prompts or photos at no cost, right? So, uh, this is one of those— I think if you've been, uh, looking at some, some ugly visuals on your company's website for many years, which, be honest, this is probably most of us, y'all liven it up. Or if you have this old, you know, corporate training video and maybe you can't replace all of it, you could at least splice it up and breathe some life into it with VO 3.1. All right. And then our last feature, which may be the one that's most talked about, although at least for me personally, I think I'm most excited about the Agents SDK, but Here we go. We got, uh, managed agents from Anthropic. All right, so more or less, if it sounds like every week that Anthropic is releasing one or two things that sound very much OpenClaw-esque, well, that's because this is very OpenClaw-esque. All right, so this is technically a little bit more technical than using Claude AI on the front end because you do have to use Anthropic's Council, right? So that's their kind of backend, but it's super easy. I did mess around with this a little bit. The good thing is, is you can do it all in natural language. You can do it all just by chatting with Claude and it will walk you through how to set it up, you know, authenticating things, etc. But here's essentially what's new. So they launched the Anthropic, launched the Claude Managed Agents right now in public beta. And it's a platform that handles hosting, scaling, monitoring, and failure recovery for AI agents so teams can focus on agent logic instead of infrastructure. So you just define agents in natural language or you can use YAML, then you can set guardrails and then Anthropic runs them. Them. All right, so this isn't like you have to set up your own infrastructure, or you don't even need to necessarily run these on a website or, you know, set up a local instance on your computer or anything like that. It is all running on Anthropic's instance. So this has built-in orchestration that handles tool calling decisions, context management, and error recovery with session tracing, right? So, um, the easiest way to explain this Think Opus 4.6 is agentic by default, right? And I think that we've all seen, I've said this many times, right? Anthropic is crushing it in 2026. They are winning the year. And one of the things that's most impressive is, as an example, you know, adding skills, you know, inside of Claude, adding all of these different MCPs inside of Claude, right? Interactive ones as well. These more interactive apps. Now imagine kind of similarly to Zapier being able to control that harness a little bit more, right? Because if you didn't have, you know, as an example, a connector, you know, for whatever program that you want to use, or if it wasn't available as an MCP, yes, you could go set up a custom MCP server, but this is just a different way to set it up. By doing it just in natural language. And then the thing that I love is that their kind of agent builder walks you through it and it'll do the authentication and everything for you, right? It'll walk you through it. It'll give you options, right? It'll say, hey, you can do it via, you know, OAuth where you can just click the screen or you can do it via API key as an example. So it's going to walk you through it. It's going to give you these different options. So who has access right now? It is a public beta. So anyone, all right, for API users and subscribers. So like I said, you do have to be on the backend. The cool thing is, well, you technically don't need to even be a paid Claude user. You do though obviously have to connect a credit card, go into the backend. I think the minimum, you know, put in like $5, right? You can like prepay $5 to go like mess around with this, right? so here's why it's useful. Because you can abstract months of infrastructure work, right? The sandboxing, the state management, credential handling, uh, permission systems, all these things that typically delay agent projects and just let Anthropic manage that piece, right? So Anthropic claims 10 times faster time to production compared to building the agent infrastructure. Yourself. All right. So here's who's going to find it valuable. I think enterprise engineering teams who want to deploy cloud-powered agents without building infrastructure from scratch, you know, project managers and non-developers who can define agents in natural language, or companies that are already using Claude's API who want managed scaling and monitoring for their agent workloads. So we will see how, you know, how popular that this is from Anthropic. On the surface, it seems like it will be fairly popular, although it is a little bit more technical. But then at the same point, I thought that OpenAI's, uh, their version of this with their agent builder, it seems like that never really took off. And it's very similar to what, uh, Claude just released here in managed agents. So I could be wrong. We'll see. Maybe most of the momentum goes into things that you use on the front end, which is something I've been saying for many years, right? I've always said You know, when you're talking about buying your building, right? I'm like, well, you should technically be buying and using these systems on the front end because the front ends are becoming more and more powerful. And I do think, you know, even as we see OpenAI eventually go to the super app, I think then it'll become apparent to people why using these things on the front end is probably gonna become more prominent. Because, well, number one is it's gonna share and keep your context, right? So yes, I think things like the managed agents from Anthropic topic are really good. I don't know if it's going to take off as an example the way that Claude Cowork has, the way that Claude Code has, because when you do it there, right, you have it all kind of under the front end, quote unquote, front end system, whereas sometimes moving on the back end, yes, obviously for developers, that's not who I'm speaking to. You all can see the, the promise of this. But for everyone else, I don't know if this is going to take off, although I do encourage you to go play around with it just like I did. I think you're going to find plenty of use cases. All right, so that is a wrap. There are the 7 features that I don't think you can afford to miss. So whether you are a Zapier power user, or if you want a managed agent inside of Anthropic, or if you've really been wanting some more content creator capabilities inside of OpenClaw, this week brought a whole lot and a whole lot for from Google and Microsoft as well. So I hope this was helpful. If so, let me know about it. If you could leave me a rating on the podcast, I'd really appreciate it. Make sure to follow and subscribe to the show. Then go to youreverydayai.com, sign up for the free daily newsletter. Thank you for tuning in. Hope to see you back tomorrow and every day for more Everyday AI. Thanks, y'all.

And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going. For a little more AI magic, visit youreverydayai.com and sign up to our daily newsletter so you don't get left behind. Go break some barriers and we'll see you next time.