Episode Categories:
Resources:
Join the discussion on LinkedIn: Got something to say? Let us know on LinkedIn and network with other AI leaders
Upcoming Episodes: Check out the upcoming Everyday AI Livestream lineup
Connect with Jordan Wilson: LinkedIn Profile
Start Here Series in our Inner Circle Community: Join for free access
AI Advancements in Business: Key Updates Driving Practical Value
In a landscape where transformative AI news is constant, several overlooked updates now provide immediate, tangible value, especially for competitive businesses. Recent changes by OpenAI, Meta, Alibaba, and other innovators are not just headlines—they offer new levers for cost control, creative productivity, and strategic flexibility. Below are the essential insights and actionable updates from the latest AI releases and features.
Unlimited Access to OpenAI’s Latest Free Model Elevates Productivity
OpenAI has quietly granted unlimited access to GPT-5.6 Luna for all free users—a significant upgrade from previous limitations. Previously, free users were served combinations such as GPT-5.5 or 5.4, which were notably less robust for day-to-day business workflows. Now, anyone on the free plan can benefit from a model comparable in strength to paid offerings, enabling robust and reliable business use cases without additional costs. This drive toward greater accessibility not only improves decision-making and workflow automation but also sets a competitive benchmark for AI utility at zero software cost, as referenced at 00:22.
Meta's Muse Code: High-Performance Coding at Deep Discounts
Meta's Muse Code, powered by Muse Spark 1.2, introduces a novel “contributor tier.” By allowing Meta to train on user data, coding costs decrease by approximately 95%—from $1.25 to $0.10 per million tokens in and from $4.25 to $0.20 per million tokens out, as described at 07:10. This pricing model targets independent developers, startups, and cost-sensitive teams unencumbered by strict data privacy requirements. For businesses with non-proprietary projects, AI-assisted coding can now be executed at a fraction of historical costs, facilitating long-running agentic tasks and background automations which previously required dedicated engineering resources.
Open Source Standards Streamline AI Agent Integration
A newly launched open, vendor-neutral standard for packaging AI agent extensions enables single-plugin deployment across major platforms such as ChatGPT, GitHub Copilot, Google, and Microsoft’s ecosystems 20:27. By housing agent skills, server code, and configuration files within a universal plugin format, companies can eliminate the need for platform-specific customizations. This efficiency is crucial for IT leaders managing mixed AI environments and seeking to avoid vendor lock-in. Notably, except for one prominent holdout, leading industry players have adopted this standard, further ensuring future-proof interoperability for businesses scaling AI-driven operations.
Alibaba Qwen 3.8 Max Offers Enterprise-Grade Multimodal AI
Alibaba’s Qwen 3.8 Max enters the upper echelon of large language models, ranking alongside the world's most advanced solutions and offering a context window of one million tokens and support for text, images, and video inputs 11:46. While immediate utility is skewed toward enterprises with the infrastructure to handle trillion-parameter models, the API’s compatibility with OpenAI's standards means rapid onboarding for organizations with established AI integration. The prospect of open weights, combined with upcoming accessible versions, may soon democratize access for mid-size businesses as well. Enterprises hedging against US big tech SaaS lock-in and seeking customizable, frontier-level AI capabilities now have credible alternatives.
Gemini Notebook Delivers Agentic Workflows for Knowledge Workers
Gemini Notebook (the successor to Google’s Notebook LM) is now agentic by default for paid users, meaning it can autonomously run code, generate charts, spreadsheets, and even slide decks from user-provided data 16:32. This upgrade empowers business analysts, project managers, and research professionals to automate data-intensive workflows, reducing manual manipulation and context switching. The system can also pull in online data with user clearance, maintaining source-grounded analysis and transparency. Broader feature parity is expected soon, but the current rollout already addresses common workflow bottlenecks for knowledge-driven teams.
Cloudflare OS: Open Source AI Workspace for Internal Teams
Cloudflare has released its internal AI workspace platform, Cloudflare OS, as open source. This self-managed solution offers browser-based agentic workspaces, controlled access to internal systems, and streamlined low-code automation, as described at 25:00. Unlike SaaS solutions priced per seat, this approach allows full customization in-house and taps into Cloudflare’s robust security and governance. Technology decision-makers can introduce advanced AI-driven collaboration tools without recurring licenses, giving IT departments direct oversight of integrations and deployment.
MiniMax H3: AI Video Generation for Rich Media Campaigns
The MiniMax H3 video model, now the highest-ranked open model for video editing and text-to-video generation, enables the production of lifelike, multi-modal video clips from text, image, and audio prompts 27:37. While higher resolutions are currently cloud-only, local quantized versions are now available and can be run on high-spec consumer hardware—albeit with longer generation times. Video marketing teams and content strategists can exploit these capabilities to produce campaign assets rapidly and at low cost, albeit within licensing constraints in the US, EU, UK, and South Korea.
Competitive Takeaways for AI-Driven Business Strategies
Several key themes have become clear:
Free users now have access to AI capabilities previously reserved for paid enterprise clients, lowering the adoption barrier and broadening the user base.
Cost structures for coding agents and generative AI are dropping dramatically when businesses are willing to share non-sensitive data.
Standardization efforts are reducing integration friction and enhancing scalability across platforms.
Open source and open weight offerings from major tech firms are pushing proprietary vendors to accelerate innovation and lower prices.
Generative AI is driving a new wave of internal automation platforms that improve productivity without increasing SaaS spend.
Strategic leaders monitoring these developments can extract immediate cost savings and workflow improvements while positioning their organizations to act on further advancements as they’re released. The pace of change makes ongoing attention to these “Friday features” crucial for maintaining a competitive edge.
Topics Covered in This Episode:
- GPT-5.6 Luna Free Unlimited Access
- Meta Muse Code Contributor Pricing Explained
- Alibaba Qwen 3.8 Max Model Release
- Adobe Unified Plugin Updates for ChatGPT
- Cloudflare OS Open Source AI Workspace
- Open Agent Plugin Standard Launched
- Minimax H3 Open Weights Video Model
- Gemini Notebook Agentic Features Expanded
- GPT-5.6 Soul Model Update & Comparison
Episode Transcript
Jordan Wilson [00:00:15]:
What if you could make a billion people smarter? Yeah. It's kinda what OpenAI just did. There's usually never been a free AI plan where I actually think it was good enough or reliable enough for everyday business use, but OpenAI just announced that as they quietly kind of announced that free users, so users on the free plan, get unlimited use of GPT 5.6 Luna, a really good and capable model. So previously, free chat GBT users were usually served a combination of GBT 5.5 and GBT 5 0.4 depending on the thinking level they chose. Actually, free and low price AI really had a week in the limelight. So between that announcement from OpenAI, there was also a new OpenWay powerhouse model from Alibaba and a new model from Meta that's basically free if you let them train on your data. So you probably missed some of it. That's because between all the big model releases, there's actually dozens of smaller feature updates each week that can move the needle for your business.
Jordan Wilson [00:01:23]:
And if you miss one or two, it's no big deal. But if you really don't pay attention to the movements that happen between the big splashes, you'll miss out completely. And that's why on Fridays, we bring you the Friday features, our weekly update of the latest AI updates that you can actually use today. These aren't news or leaks or coming soons. These are just new AI features you can use this Friday. So let's get into it. So on today's show, here's what you're gonna learn. You're gonna learn how you can get essentially free frontier coding or at least for free 99.
Jordan Wilson [00:01:57]:
If you let a certain company named Meta train on your data, you're gonna know what new open source standard there is for better agent orchestration and how basically all the big players in the game got in on it except for one. And you're gonna know the new updates that give free ChatGPD users the best deal in AI right now. Let's jump into it. Welcome to Everyday AI. My name is Jordan Moulson, and we do this every day. This is your daily livestream podcast and free daily newsletter helping business leaders like you and me keep up with these AI updates because they quite literally don't stop. I tell you what's good, what's not, how to use it to grow your company and your career. So if that's exactly what you're trying to do, do me a favor, stranger.
Jordan Wilson [00:02:40]:
Click that subscribe button on Apple Podcasts or Spotify, wherever you're listening, and then make sure you go to our website at youreverydayai.com. It's in the show notes, so make sure you go check that out and sign up for our free daily newsletter. We're gonna be recapping the highlights from today's show. So if you miss anything, don't worry. It's in there as well as all the other AI updates you need to know to get ahead. Alright. Let's get into it and look live. Alright.
Jordan Wilson [00:03:04]:
So, yeah, if you are watching, or sorry, if you're listening to this, there's always a video version, not too much that you need to see, today if I'm being honest. But, you can always check that out on our website @youreverydayai.com. Alright. Here we go. Eventually. Come on, screen. There we go. It's unedited, folks.
Jordan Wilson [00:03:26]:
Alright. Our first big AI update. If you're creative, you're gonna like this one. That's because Adobe just launched a big update to its unified plug in in chat g p t. Now it can access more than 70 tools from Photoshop, Firefly, Acrobat, Express, Premiere, Lightroom, Illustrator, InDesign, and stock. So, essentially, how does this work? Well, you describe what you want it to do, and it picks the right Adobe tools automatically. So you can edit photo sets, make campaign videos, customize different design templates, turn spreadsheets into PDFs. So it kind of replaces the separate Adobe apps that were available in chat GBT before.
Jordan Wilson [00:04:06]:
So now it's just one unified app where you don't even have to, like, know, oh, should I be using, you know, Acrobat for this? Should I be using, you you know, Lightroom, Photoshop? It doesn't matter. So you just open ChatGPT. If you do have that Adobe Creative Cloud, Creative Cloud subscription, and you just dump in a bunch of files and say do all this stuff to them, and it does it. So Adobe is running this play on every chatbot, not just Chat GPT. Its tools are already on Claude and Copilot, but but I believe this is the first, in terms of these types of features. This is the first one, we've seen that's this robust. So who has access? Well, anyone. If you are on Chad GBT, you got it.
Jordan Wilson [00:04:48]:
And there's also four, many, features that are in included in this announcement. You don't need a plan, to use them. So yeah. Even right now, you don't need an Adobe account to start. Even just using guest mode does give you a broad set of tools immediately, not everything. Because signing in with your Adobe account does unlock, some of the more generative, you know, heavy AI features, and then being able to sync everything in your creative cloud. So it also does work inside of Chad GPT work in codec, so it's not just on the, chatgpt.com release. So who's this useful? Well, anyone.
Jordan Wilson [00:05:26]:
If you're trying to do something creative, if you are creative, if you're not creative but you wanna be creative, this is a great way to do it. So, obviously, OpenAI and chat g p t has a fantastic, literally state of the art, kind of their, GPT images too, which is really, really good. But it can do everything that you might want it to. Right? So in those cases, now you have this tool from, Adobe. Alright. So that's a good one. But here's, maybe even a better one, maybe for a lot of our audience because Meta, another pretty big splash here. And there's actually a plan here that I wasn't expecting from Metta.
Jordan Wilson [00:06:07]:
And I don't hate it as one that I'm just like, Hey, company, take my data. I don't hate this new pricing from Metta Muse, Spark, and their new Muse code and Spark 1.2. So here's what's new. Meta released Muse code, which is their kind of take on Cloud Code or Kodak. So that's a terminal coding agent right now, powered by its new Muse Spark 1.2 model. So, yeah, two different things. It is Muse Code. So that's a terminal coding agent.
Jordan Wilson [00:06:37]:
Right? So not like the desktop version of quad code or codex, like the CLI version of quad code or codex. And then, the next one is the new model, Muse Spark 1.2. So the pricing here, even if you're paying the full API pricing, pretty good rates in terms of performance. It's a dollar 25 per million tokens in and $4.25 per million tokens out. But there is a new, what they're calling, contributor tier. It's an interesting name for it. That drops that price down to 10¢ in and 20¢ out if you let Meta train on your data. So that's roughly 95% off in exchange for essentially your code.
Jordan Wilson [00:07:25]:
Right? So, at first, I'm like, absolutely, no one's gonna do this. But then I'm like, yeah. Like, I don't know. Especially, if you're starting from scratch. Right, let's just say that you have a personal project or, you know, if you're an indie hacker, indie developer, I'm like, that's actually not a terrible idea. Right? Obviously, if you're working with large proprietary, anything enterprise, no one would ever touch that contributor tier. But I'm like, for everyone else that is more of a hobbyist, it's actually, I think, not that bad of a deal. So, every model call tool run approval and edit, does get written to a local event log.
Jordan Wilson [00:08:05]:
So if the agent crashes, it can resume where it stopped. So who has access to this? Well, it's open now in beta. You can literally just go run a terminal command and you're off and running. So the new model though, MetaSpark, sorry, Muse Spark 1.2 is available inside Muse code and to any developer through the meta, meta model API, and it does have a 1,000,000 token context window. So, why is this useful? Well, aside from writing code, if you need persistent background sub agents that can work on hard tasks on their own and decide when to report back, it's pretty good for that. You you know, the scores so far, for what's available on artificial analysis as an example, pretty good. Right? You know, all of a sudden, Meta went from not very relevant for, like, a full year and a half to now all of a sudden they're in the conversation. Right? They're again, Google might change this, you know, any any minute now.
Jordan Wilson [00:09:06]:
Right? Watch Google. There's been, rumors that we're getting something, you know, maybe this week or August 10, either Gemini 3.5 pro or if they skip straight to Gemini four pro, who knows? But Meta is actually ahead of Google. Right? Not just in terms of model capabilities, but on pricing as well. So Meta has had a very impressive, last month or two. So I think if you're, you know, a dev working on a large, repo, if you want long running agents, like I said, or if you're in indie dev or a cost sensitive team and you're not necessarily working on, you know, proprietary data, that contributor tier, which is still a funny name for it, not bad. Right? We'll see if Muse code, goes for very long. Right? I don't know. Maybe this is the non developer in me speaking.
Jordan Wilson [00:09:59]:
I've always been able to do the basics in a terminal, but I feel even the people that have been hardcore, like, hey. I'm a terminal coder, you know, in in in terms of being a a software engineer or dev. I've seen even some of the people that, you know, a year ago were like, I'm never giving up the terminal to being like, the terminal just doesn't make sense anymore. So I don't know. We'll see. I mean, you have to start somewhere with, you know, in terms of having a coding agent. But regardless, Meta's showing where they're gonna play. They're gonna try to compete with Anthropic in Google on token pricing.
Jordan Wilson [00:10:30]:
I don't think that's necessarily OpenAI's game. That is where, Anthropic reportedly gets about 80% of its revenue. So this to me looks like a shot at Anthropic, you know, bringing the actual coding agent with Muse code, extremely cheap and affordable, and the performance is there. Alright. Our next piece of AI news, speaking of, well, very affordable, models, well, we have a new, not quite king of the open hill. It's kind of, actually, let me check that because I know artificial analysis just put out a new version, of their four point what is it? 4.11 now. So I'm I'm double checking live to see, okay. Yeah.
Jordan Wilson [00:11:13]:
So they are just a little bit behind Kimmy k three, but, essentially, it is in the same tier. So Alibaba's new, open weights, Quinn 3.8 max, it is in the same tier. So it is in that, you know, GPD five six, Claude Fable five, Claude Opus five, Kimmy k three, and Quinn three eight. That's kind of your top five models in the world now. You know, we have two open weights models from China, in spots four and five. So here is what it is. It is. When 3.8 max, a 2,400,000,000,000 parameter open weight model with a million token context window, it takes text, images, and video.
Jordan Wilson [00:11:57]:
So here's the, interesting thing. So it did become, the top ranked Chinese model for text only. Alright. So for text only, it takes that. And here's the thing though. There was a report literally hours ago, from Reuters that said Alibaba plans to begin charging major enterprise users for Quen 3.8 max to charge them a share of their revenue they generate starting next week. So that one's interesting. So who has access right now? Well, anyone.
Jordan Wilson [00:12:31]:
So if you're working on the hosted API or through Alibaba cloud model studio, so yeah. If you are in Alibaba cloud accounts, you can use it right now. And the API is OpenAI compatible, so switching is a base URL in model, ID change. So it's open weights, although the weights are not out yet, but they did promise them within days of the release. That's essentially what Kimmy did as well, and they kind of hit their target deadline for releasing the weights. But when the weights land, obviously, you can't, like, self host this. Right. And, again, unless you happen to have a dev a data center, you know, a bunch of GPUs that you just weren't using in your basement, you aren't gonna use this.
Jordan Wilson [00:13:15]:
Right? It's kind of this, you know, I don't know, fortunate, unfortunate, current academy of open weight, open source software. Right? I think a year ago, the best open models you could, you know, download. If you had a big expensive Mac, you could do it. You can't anymore. Right? These multiple trillion parameter, you know, open weight models, they're not really for running locally anymore unless you literally have a data center. So well, why is this useful? Well, it's frontier level capability with open weights. So, yeah, if we're bigger companies, you can, you know, fine tune these models, or if you're working with a provider, you can fine tune them. So unlike proprietary models, where it's not always as easy to do something like that.
Jordan Wilson [00:14:00]:
So, also interesting to note, it did debut number four on the front end code arena. So, you know, apparently, it has some pretty good front end, taste to it as well. So who's gonna find this valuable? I think really just large enterprises. That's it. So if you're hedging against, you know, US lab pricing and lock in, or just, you know, people who are already self hosting large models, researchers, they are coming out with a 27 b version. You know, so if you do have, you know, a couple heavy Mac studios beefed together, you can run it, like, super slow or something like that. Actually, 27 b, yeah, you should be able to run it on, like, the most expensive d g x's. But, yeah, Again, nice open model, open weight model.
Jordan Wilson [00:14:50]:
So if nothing else, it's pushing the other open models, which are pushing proprietary models, which are push pushing prices down. Right? That's ultimately what's happening here, because the people who can afford, to run something like a quen, quen 3.8 max by running it locally are those, you know, fortune 500 types that crap that may have access to compute. So that is the people who are generally, you know, using anthropics models would probably be their first customer. OpenAI, Google maybe as well. So, you know, that's also gonna push, in theory, you know, either push down the prices from the major frontier labs or at least keep them in check. So even if you're like, okay. This model isn't gonna impact me or my business. We can't run it.
Jordan Wilson [00:15:33]:
Well, you can. Right? There's always, you know, providers that you can use. But if anything else, keep models in check. Alright. Let's go next. Next, we already covered this one in-depth on our Wednesday AI working Wednesday show. That's episode eight thirty four, but a ton new with Gemini notebook. Yes.
Jordan Wilson [00:15:54]:
The tool formerly known as notebook l m finally got its big glow up that rolled out to all paid users. So this is technically an older update, but we didn't cover it on the show previously. Well, aside from Wednesday, when it was you know, I think it was just released on Tuesday. We didn't really cover it before because it was only available to Ultra subscribers. So just so you know, on these Friday shows, these are just, you know, tools or updates that are available to your standard, you know, $20 base plans. But now we have that Gemini agentic notebook that is out. So, yeah, has a new name. Notebook l m is gone.
Jordan Wilson [00:16:32]:
Gemini notebook is here, and Gemini notebook, at least for paid users, is agentic by default. So Google finished rolling out the agentic version of Gemini notebook, the upgrade to pro subscribers. So every notebook now gets a secure cloud computer that writes and runs real code against your sources, plus auto generated charts, spreadsheets, slide decks, and PDFs. So it runs on Gemini 3.5, but it also uses the anti gravity harness, and it has a 100 built in skills. So think of, like, code interpreter grounded in your documents. You can also just, you know, start a blank notebook, give it a vague query, and it can also now use the web. Although, it will confirm with you. Right? Because that's the good thing about, you know, previously notebook l m is it saved completely grounded in your source data.
Jordan Wilson [00:17:20]:
So now, Gemini notebook can query the web, which is a little bit different for me. Right? But it will at least let you know, hey. You know, if you're asking about something that's not your source material, it will tell you, I don't have the answer. Do you want me to go search the web? So, who has access? Anyone with a, Gemini Pro accounts, you've got it now. So this is on the web version only. If you're an ultra subscriber, you already have this. Free users, you're still gonna see Gemini notebook. Right? But you're not gonna have these new agentic the code execution or some of these new output formats.
Jordan Wilson [00:18:00]:
So that's the other thing that's new. In Gemini notebook, you can output as a PDF, a doc, you know, different PNGs. Right. So, essentially, before, you were a little more limited in what you could actually output. There was, you know, I think about eight different sources on the right hand side in the studio panel. Now you can just say, you know, output in these different sources. So I actually did what I thought was a pretty good rundown on Wednesday. So if you do wanna check that out, make sure you go do that.
Jordan Wilson [00:18:27]:
Episode eight thirty four. Who's gonna find this valuable? Literally, anyone. Right? If you're, you you know, a heavy, CHGPT work codex, cloud code, user, right, as an example, these new agentic features are pretty good. Right? I'm really looking forward to when, you know, Google does introduce whether it's Gemini 3.5 or Gemini four pro, whichever one, and then seeing those roll out to notebook l m, I'll be really excited for it because they're technically running the Gemini three point flash model. So running it agentically, sure. Good. But the Gemini three five flash model is not that great. So, yeah, it's got, like, a 30 what is it? Mid no.
Jordan Wilson [00:19:11]:
That's Gemini three six flash. Come on. Come on. Where's where's my Gemini three flash? Let's see. Let's see where we have it. It is okay. A 52 on the artificial analysis. So not that good.
Jordan Wilson [00:19:23]:
So it is below, or no. Sorry. It is at the same level as, GBD 56Luna. So okay. From that respect, it's okay ish. Right? So if you think that OpenAI has three tiers, you you know, they have, Soul, Terra, and Luna. Right now, Google's best tier is the lowest tier for, OpenAI's Gen GPT. So take that for, you you know, what it's worth.
Jordan Wilson [00:19:50]:
So not the best model in the world, that's powering it, especially when you're a paid user. Right? We're gonna talk about the free, kind of upgrade for, GBD $5.06 Luna, which I think is great for free users. Right? But for being a paid user right now, it's kind of not a great time, to be a, you know, Gemini first company, but I will expect Google any day, any minute, any week now to come out with a good 3.5 pro or Gemini four pro, and I hope it gets rolled out to Gemini notebook. Alright. Next, we have a new open standard for agent plug ins, and this is pretty cool. So here's what's new, and, yes, it's available now. There's a new open vendor neutral standard for packaging AI agent extensions. So now it's a one plug in, a folder with a plug in JSON file bundling in agent skill and MCP servers.
Jordan Wilson [00:20:41]:
And right now, it works across chat g p t, codex, cursor, Google, GitHub, Copilot, Versus Code, and others. So, essentially, you can build an agent plug in once and run it anywhere. So this is a little confusing, but let me just break it down, and hopefully, this is, like, 99% correct because I haven't actually gone out and built this because it's only, like, in a couple hours old. So previously, think. You have skills. Right? And skills are essentially just a bunch of text files in a folder. Right? And then you also have things like plugins, and plugins are, you know, multiple skills or a skill in an app together. Right.
Jordan Wilson [00:21:24]:
And then you have, like, JSON files. So it's kind of like the new agent plugins. You know, and then, you know, you have your skills or markdown files. Right? It's essentially pulling all of these things together, but in an open standard. Right? So that's great because you have the skills MD. You know, skills are pretty universal, right, because they are just text files. They're markdown files. So those are pretty universal.
Jordan Wilson [00:21:52]:
Also inside the, agent plug ins sorry. I forgot to mention this. It's MCPs as well. So it's essentially stacking, you know, your skills, your MCP servers, you you know, your JSON files, you you know, all in one package. So this is nice, because before, you would have to load all of these things in individually, which takes a lot of time. And when, you know, it's slightly different, across to providers, you don't always have that from a business strategy perspective. You get a little locked in at times, which you don't want. So this is why it was great to see almost every single big name instantly support the Asian plug ins.
Jordan Wilson [00:22:33]:
Right? Anthropic, obviously, not on the list. I don't know why Anthropic just won't get on board with this. You you know, obviously, they have their claude.md. Right? But it's like, okay. Anthropic, just change it to agent MB, like, literally every other company. Right? Google, Microsoft, Cursor, OpenAI. Everyone else is doing this. Anthropic, get on board.
Jordan Wilson [00:22:56]:
It's just making, it's I think it's slowing the pace of AI down. I think it's a maybe a bad look on anthropic. Like, are you just trying to lock people in? Right? May I don't know. Maybe they are. Maybe that's part of their business strategy. But for me, right, I love, kind of having this this open, you know, new agent plug in standard because it does make it easier for businesses. The like, the other thing too is, yes. In most cases, I think it's best for companies to kind of pick one AI kind of operating system and stick there.
Jordan Wilson [00:23:27]:
But, you know, you might maybe your finance team needs a different one. Right? Maybe 90% of your company can be on, you you know, JWT work, but maybe 10% of your company needs to be on Gemini for whatever reason. Right? And in this case, when you have these agent plug ins, it's very easy, to have that standard and to be able to transfer that over. But, you know, if Anthropic doesn't get on board, it's not gonna be as easy for them. Anyways so OpenAI announced this, but it's not OpenAI's. So Vercel initiated it, as well as, Amazon, AnySphere, which is Cursor, GitHub Copilot, Microsoft, etcetera. So, who has access right now? Well, anyone. So you can just build this.
Jordan Wilson [00:24:11]:
Right? You can just point your coding agent to this, the website. The website is just agent-plugins.org. Right? All of the documentation is there. So this is just an easy way, especially if your team is becoming more agentic in your work. Right? You have multiple MCP files. You have multiple skills. You have JSON files holding it all together. It's just an easier way to port this across server to server.
Jordan Wilson [00:24:36]:
Alright. Or or sorry. Service to service. Alright. Next one. Speaking of being open. Right? It's a big week for open and free. This one was not expecting this.
Jordan Wilson [00:24:46]:
Cloudflare announced their open source Cloudflare OS. It's an internal AI workspace that it built to actually run its own company. So they essentially said, we built this to run our company so everyone else can have it as well. So, essentially, every employee gets a browser based agent workspace with govern access to internal systems, research documents tied to live data, automated workflows, and shareable internal apps with, well, no developers really needed. So thousands of Cloudflare's own employees across every team already use it daily according to the company, so seems pretty battle tested, not just like some demo software. So it competes technically, right, with Chad GPT work or Claude Cowork or Gemini Enterprise, but it's free, open source, and runs in your own CloudFlare account instead of the kind of per seat SaaS. So, who has access right now? Well, it's open source. So anyone.
Jordan Wilson [00:25:45]:
So you can go download the repo. You do obviously need a CloudFlare account to run it because it deploys into your own account, using your own access policies and integrations. But your IT team, essentially, yeah, they'll probably need to help you a little bit. But there is also a managed version through CloudFlare plus implementation partners, that, are coming, but no date is given yet. So, why is this useful? Well, you can just essentially bring your own model, and have an open source agentic platform to work on. So, yes, this isn't free, to use. It's just a free operating system. Think of it that way, but you still have to bring your apps a k a.
Jordan Wilson [00:26:33]:
You still have to bring your own model through the CloudFlare AI model gateway. So but there's, you know, you can do spend caps, and you're not locked in. So think of it more like that. Think of it, you know, and there are other versions of this. Right? There are versions of open, you know, open versions of, you know, desktop agents. Right? But this is from a CloudFlare, one of the literally biggest names, in the Internet ever. So a very trusted source. So, you know, a lot of companies run their, you you know, web operations through CloudFlare.
Jordan Wilson [00:27:08]:
So trusted trusted name, love this because it is giving people more options. Again, I think more options, more modularity in being able to build, the better for people involved. Although for me, I'm still gonna stick to my codex, because it's just part of my everyday workflow. Alright. Next. Yes. This this model has been going absolutely viral. It is Mini Max h three.
Jordan Wilson [00:27:37]:
So this is a new video model, and they did Mini Max did just release the weights for h three, and it became the first open model to ever top a major AI video ranking as its number one in video editing on artificial analysis and number two in text to video. So artificial analysis called it the strongest open weights model by far. That's why this thing has been blowing up online. If you've seen any, like, really good AI video, over the last, like, couple of days, like, on Twitter or whatever, it's probably Mini Max, especially and the reason why you're probably seeing it is because, obviously, with this being, in OpenWeights Chinese model, there's, like, no copyright restrictions. I don't know why it's been, like, at least on my feed, it's been, like, these fifteen second Seinfeld clips, you know, of people talking about AI or, you know, fifteen second clips from the office, people talking about AI, and they're obviously really good. Right? Because it's just, you know, ripping the copyright. It's, you you know, Michael Scott's face and Michael Scott's voice, but talking about, you know, agentic swarms or something like that. So, here's a little bit more about the model.
Jordan Wilson [00:28:55]:
So the model reads text, images, video, and audio as one context, and it outputs four to fifteen second clips at up to two k with native stereo audio. So that's if you're using their hosted service. But if you are, using the, actual, you you know, downloadable version, the the the quantized version. You're not gonna be able to run the two k, resolution. It is gonna be a standard, standard definition, and it's gonna take a while. Right? So let's just say you have, like, you know, 16 or 32 gigabyte MacBook. Let's just say that. Yeah.
Jordan Wilson [00:29:36]:
It's gonna take you maybe twenty or thirty minutes or longer, to make a fifteen second clip. But the thing is, well, it's an option. Right? Now we literally have models that are the state of the art. Right? So, you know, number one, number two, and the two big categories, and you can download them and use them for free. Right? So absolutely wild world we're living living in. And for the most part, you cannot tell that they're, AI videos. Right? Like, obviously, you can't because, you know, if Jerry Seinfeld is losing his mind, over an agentic swarm, obviously. Right? But, I mean, everything else about it, it looks real.
Jordan Wilson [00:30:21]:
It sounds right. Jerry Seinfeld's voice, how he does his voice, he throws it all over the place. It it gets it right. It nails it. Right? Obviously, every once in a while, you're gonna get some some bad generations, but the stuff I've seen online, I'm like, this is fairly impressive. But also to me, it's extremely scary because you know what takes like fifteen seconds. Yeah. Like campaign videos.
Jordan Wilson [00:30:44]:
Right? This is gonna absolutely wreck the upcoming US election because here's the thing, it's out in the wild. So it's not like any amount of, you know, new US policy is gonna stop this thing. So, yeah, I can see this thing causing a ton a ton of trouble in a couple of months, with the elections coming up. So who has access? So, yeah, the hosted version, anyone, anywhere, right, through the minimax API. So, yeah, you don't need the fat, you you know, $10,000 computer to, you you know, eke out a fifteen second video for that one. So you're paying the API prices, but the downloadable weights are on hugging face, but the license excludes right now, The US, EU, UK, and South Korea from running it locally. So if you're in those regions, it says they're API only. It seems like most people are skirting around those, restrictions somehow.
Jordan Wilson [00:31:41]:
Alright. So why is it useful? Well, it's one model that can replace a lot. So it does text to video, image to video, first, last frame, subject reference, motion reference. So it it is much more than just a video generator. There's a lot that's going on under the hood. Alright. Last but not least, there is a new version of GPT five six Soul. So, yes, OpenAI made a, pretty notable update to its flagship model.
Jordan Wilson [00:32:10]:
Here's what's new. Well, the big one is OpenAI made the free chat g p t the best of free AI has literally ever been because now you have unlimited generations even if you do not have a paid plan, on GBD 5.6 Luna. So, again, kind of the tiers, GBD $5.06 Soul is the best, GBD five six Terra is in the middle, and then you have GBD five six Luna, still a very good capable model. Actually, GBD $5.06 Luna is pretty much on par, with, quad SONNET five. So just if you need a comparison. So, for paid users, OpenAI merged the instant and thinking models into one. So now you have a slider that controls how much reasoning it puts into each answer. So if you're on chedgvt.com, it's gonna seem a lot more familiar if you were using the desktop version.
Jordan Wilson [00:33:09]:
So along with this, OpenAI claims 68% fewer responses containing factual errors versus g p d 5.6 instance on areas such as financial, medical, and legal prompts. So Zalt just dropped. It's new. So who has access? Well, literally, 1,000,000,000 users. Yes. So, 1,000,000,000 people have better models altogether. So this is not only free plans getting unlimited GBD 5.6 Luna, which I still don't understand how that's possible. Right? How is it possible that, like okay.
Jordan Wilson [00:33:45]:
Let's just say anthropic. Right? Let's just say for whatever reason, obviously, they the, you know, Opus five, and Fable five are, like, you know, one a, one b, or one b, one c depending on how you look at it with the g p d 5.6 soul. But after that, right, it's it's g p d five's, g p d or sorry, Sonic five. Right? On SONNET five, you on a paid plan, right, in your five hour window, you might get 10 to 20 prompts, right, on a paid plan. So now on a free plan on Chattopty, you get unlimited. So this, again, I, like, I know I sound like a broken record, and, you know, there's plenty of stories, you know, chronicling this online. This has been a bad couple of weeks for Anthropic. Right? Between all these new open models coming out, DeepSeek comes out, essentially cheap, like, free.
Jordan Wilson [00:34:42]:
Right? OpenAI, we talked about this last week. They essentially made their g b d 5.6 Luna essentially free. So, yeah, it hasn't been a good couple of weeks for entropic. So, we'll see where their number lands. They're expected to go public maybe in, like, as soon as two months. So that one, should be interesting to see. So, who's gonna find it valuable? Why is this useful from OpenAI? Well, a couple things. Well, there's no more to having to choose between instant in thinking, for paid users.
Jordan Wilson [00:35:12]:
It's just one model with a dial. And then, obviously, the unlimited free chat, absolutely bonkers. Right? This is, I think, OpenAI, starting to ink, inch closer to, you you know, being able to provide this level of intelligence to, you know, everyone in the world, which it seems like, you know, that's what their goal is, you know, bringing AGI to the to the masses. Right? I still cannot get over like, when I read that, when I saw this announcement, I'm like, this can't be right. Like, I'm like, an unlimited g b d 5.6 because usually free users, right, because it's all about compute. You know, usually free users are one tier below. Right? Which is how it was, you know, yesterday. Right? You know, free users were on GBD 5.5, and that's how it's usually been because free users make up the majority of that billion users.
Jordan Wilson [00:36:07]:
So, OpenAI had in in terms of post training and inference, they've cracked something. Right? I had kind of my, recursive self improvement show. It's like, yeah. We we no one has full RSI, but it seems like OpenAI has got something figured out. The fact that they lowered the price of a g p d 5.6 Luna by 80%, and now they are giving a, frontier level model, a way to a billion people, unlimited unlimited for free? Doesn't make sense. Alright. Anyways, I'll stop losing my marbles and let you get to go using all of these new AI features. I hope this was helpful.
Jordan Wilson [00:36:46]:
If so, let me know. Subscribe to the podcast on Apple and Spotify, then go to youreverydayai.com. Thanks for tuning in. I'll see you Monday and next week for more everyday AI. Thanks, y'all.