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
Anthropic Claude Fable 5 and New Google, Microsoft AI Features: Targeted Insights for Business Leaders
In a week dominated by the launch of Anthropic’s Claude Fable 5—now positioned as the world’s most powerful AI model—critical updates quietly surfaced across Google, Microsoft, and OpenAI’s AI platforms. These developments directly impact business workflows, data management policies, and decision-making processes. The following detailed overview translates these advancements into tangible business outcomes, enabling strategic adoption without introducing unnecessary operational risks.
Claude Fable 5 Capabilities and Enterprise Guardrails
Anthropic released Claude Fable 5, part of the Mythos model family, establishing it as the new performance benchmark for a broad range of applications, from 3D world building and software engineering to knowledge work and scientific research. Fable 5 is available to all paid users on Claude’s primary platform until June 22, after which API access is required, at twice the cost of the previous top model 07:14.
Business adoption, however, is directly constrained by a mandatory 30-day data retention policy 05:15, which cannot be opted out of. This creates a compliance tension for large enterprises, especially those in regulated industries (e.g., finance, healthcare) with strict data privacy requirements or zero-day retention mandates. Even Microsoft, despite its financial investment in Anthropic, suspended internal use due to these policy issues 05:44. Organizations must analyze whether the benefits—state-of-the-art capabilities in AI-driven content creation and technical problem-solving—outweigh the significant data governance risks 06:09.
Google Gemini 3.5 Live: Real-Time AI-Powered Multilingual Communication
Google introduced Gemini 3.5 Live, an audio AI model delivering near instant speech-to-speech translations across more than 70 languages 10:26. This model preserves speaker intonation, pacing, and pitch, auto-detecting spoken language and generating translated audio within one to two seconds of the original 10:55. The model works continuously—unlike traditional turn-based translators—creating smoother communication in real-world scenarios, such as multilingual business meetings or customer interactions 11:09.
Gemini 3.5 Live is currently incorporated into Google Meet, the Google Translate app, and Google’s AI Studio, where it can be used directly in browser tabs streaming foreign language audio 13:14. This supports multinational teams engaging in real-time dialogue and expedites onboarding for global clients and partners, positioning Google’s tool as a valuable asset for geographically dispersed organizations.
Advanced Visualization and Email Integration Inside ChatGPT
OpenAI’s ChatGPT quietly received several workflow-accelerating features. ChatGPT now delivers interactive bar, line, pie, and scatter charts directly in conversation threads, auto-generating visuals when users request them or when relevant data is present 15:03. These charts are dynamic and can be converted back to static images as needed, eliminating the need for manual spreadsheet exports and enhancing data comprehension directly within ongoing chats 16:40.
Additionally, ChatGPT enables direct drafting and sending of emails—integrated with Gmail and Outlook—from within a chat session 29:59. After connecting the preferred email provider and granting write access, users can edit AI-generated drafts using the platform’s new "writing blocks" feature before sending. This workflow streamlines outbound communication, especially valuable for high-volume email management and reducing context-switching for professionals managing customer or client correspondence 30:38.
Microsoft Copilot Studio: Streamlined Agent Authoring and Multi-Advisor Roundtables
Microsoft rolled out a fully rebuilt Copilot Studio, delivering a more intuitive, streamlined authoring experience for building AI-powered agents and automations 18:03. The updated Copilot Studio consolidates configuration tabs, strengthens agent instruction adherence, and introduces a new orchestrator for more consistent execution of multi-step tasks. These improvements yield approximately 20% better benchmark performance and 50% lower AI token consumption 19:32.
A new feature—Microsoft Copilot "team of advisers"—is now live for Copilot Online users 26:30. This tool introduces AI-generated panels, each adviser offering distinct expertise and reasoning. Unlike traditional chatbots, this roundtable approach stimulates layered discussion, challenging and augmenting proposed conclusions in real time. Several prebuilt templates are available, including historical figure debates, business decision panels, and product naming workshops, expediting ideation and scenario analysis for business teams 29:14.
Claude and HeyGen Hyperframes: Code-Driven Video Content Creation
The integration of HeyGen Hyperframes into Claude Fable 5 allows business teams to convert prompts and documents into composable video scenes, rendered as shareable MP4s 22:08. This connection provides an open-source, agent-driven framework for rapid video creation without requiring video editing skills or specialists. Use cases range from internal training videos to marketing collateral—particularly valuable for organizations scaling content output or enabling nontechnical contributors to produce high-value multimedia assets 25:20.
Why These AI Developments Matter for Business Operations
The AI features released this week provide targeted leverage points for operational efficiency and competitive advantage:
Adoption of Claude Fable 5 should be balanced against new data retention risks, demanding enterprise policy review before deployment in sensitive verticals.
Google Gemini 3.5 Live empowers real-time problem-solving and customer interaction across language barriers, reducing friction in global operations.
Interactive visualization and in-chat email management in OpenAI’s ChatGPT offer immediate workflow gains for data-driven teams and customer-facing professionals.
Microsoft’s Copilot Studio and team-of-advisors functionality shorten iteration cycles for process automation and group decision-making by creating AI-assisted, multi-perspective panels inside business applications.
HeyGen Hyperframes in Claude accelerates video generation, enabling more agile communication internally and externally, even when creative resources are constrained.
Business leaders evaluating AI investments should focus sharply on these specific upgrades—not as abstract productivity drivers, but as directly quantifiable workflow multipliers, provided potential compliance and oversight issues are fully considered.
Topics Covered in This Episode:
- Anthropic Claude Fable 5 Mythos Model Overview
- Claude Fable 5 Data Retention Policy Controversy
- Microsoft’s Stance on Claude Mythos Model Integration
- Claude Fable 5 Benchmark Performance Analysis
- Google Gemini 3.5 Live Speech Translation Demo
- Real-Time AI Translation for Google Meet & Apps
- ChatGPT Interactive Charts and Visualizations Update
- ChatGPT Inline Email Drafting and Sending Feature
- Microsoft Copilot Studio Rebuild and Workflow Enhancements
- Copilot Multi-Agent Team of Advisers Introduction
- HeyGen Hyperframes Video Plugin for Claude
- OpenAI Canvas and Writing Blocks Feature Notes
Episode Transcript
Jordan Wilson [00:00:16]:
Yeah. The biggest AI feature this week was the world's new most powerful model, Anthropic Claude's fable five of the Mythos family. But while you were squeezing out as many prompts as possible until it goes away on June 22, there were some other under the radar AI updates to other large language models and AI tools that you probably use every single day. Like, did you know that you can actually send emails in chat g b t now? Or that Microsoft Copilot has a multi agent group of advisors who will work on a problem together for you? Or that Google released a new model that can translate any spoken word live even in your browser. So, yeah, if you get too focused on one big model release, you might miss a handful of other useful AI features that make your day to day workflow much easier. But it's pretty much impossible to spend hours a day combing through release change logs, Twitter accounts, and YouTube videos to understand and test what's new. That's why I do it for you all week and then break it down for you on Fridays with our new Friday features series. So welcome.
Jordan Wilson [00:01:32]:
Let's get into it. So here's what you need to know today. On today's show, you're gonna learn why Microsoft isn't even using cloud's new model and why your company might wanna sit it out as well. You're gonna learn how a new Google AI model might change how the world communicates and the new AI update that allows you to send email from a chatbot without leaving. Alright. Welcome to Everyday AI. Let's get into it. My name is Jordan Wilson, and this thing's for you.
Jordan Wilson [00:02:02]:
It's a daily unedited, unscripted livestream podcast and free daily newsletter helping business leaders like you and me not just, keep up with everything that's happening in the world of AI, but how we can actually use all this information to get ahead to grow our companies and our careers. So if you find yourself literally struggling to stay afloat because you have a real job to do, but, you know, maybe you're being told to use all these new AI features and you're like, I don't know what they are. How can I keep up? That's what I do for you on Fridays. So if you want the normal AI news, we're gonna have that in the newsletter, so make sure you check that out. Go to our website at youreverydayai.com. Alright. So a lot to cover. Obviously, the world's most powerful model with the new fable five of the Mythos family.
Jordan Wilson [00:02:51]:
Some big releases from Google, Microsoft, some smaller ones from OpenAI too that I think will make a big difference. But let's just get straight into it. So, I'm gonna be sharing my screen for our livestream audience. But podcast audience, you're probably not missing too much, really just reading some, from some, release, blogs and things like that. So first, yeah, we have the new top of the line AI model with Claude Fable five and Claude Mythos five. So for the most part, you're probably gonna be using Claude Fable five. Let me explain. So Fable five is of the Mythos family.
Jordan Wilson [00:03:32]:
So Mythos preview and Mythos five, these are the models that are only available to project Glasswing. So a, select few enterprises. So, basically, no one, you know, has access to the Mythos family models. But, Anthropic, what they did is they released a version called Fable. So Fable is part of the Mythos family, and it sits on top of, Haiku, Sonnet, Opus, and then you have your Mythos family. And Fable is essentially Mythos with guardrails that's safer. Alright. So this was, just released, and it is available for everyone now.
Jordan Wilson [00:04:11]:
So Fable five is the publicly available version. So in high risk areas, though, like cybersecurity, biology, and chemistry, you can't use it. So, yeah, in those instances, it's going to fall back to Claude Opus 4.8. There's been a lot of chatter about that in the first couple of days since Fable has been available. Anthropic has taken some hits on the chin because I think they didn't think this one through. You know, even if you're someone that is just a researcher, there's a lot of people showing examples of this. Couldn't even start a new chat with Claude with Fable five. And you had entire organizations that couldn't, and they were paying for it.
Jordan Wilson [00:04:47]:
But hopefully, Anthropic will get that figured out because the model is really good. But I think the safeguards were maybe a little too overzealous. So Microsoft as well. And, yeah, I teased this one because I think it's actually important. And I'll probably talk a little bit more about it in Monday's show. But there's a lot of downsides to this model. So, I actually covered them in yesterday's episode. So make sure you go listen to episode seven ninety six.
Jordan Wilson [00:05:15]:
I talked about the four big downsides of Mythos, and one of them is it has a mandatory thirty day data retention policy. That's not a joke. You can't opt out of this. So as an example, Microsoft, which is a big financial investor in Claude or in, Anthropic in the Claude models. Right? So not only are they a financial invest, investor, but they usually offer that to their customers via Microsoft Office. So, reports said that Microsoft themselves stopped using the model for that very reason. And I talked about on yesterday's show about how, hey. If you're a large enterprise, you probably shouldn't be using it just for that.
Jordan Wilson [00:05:56]:
I do expect or would hope that Anthropic will pedal back on this. Like, they've, kind of, I guess, clawed it back, clawed back some of the things with the, the safeguards. Yeah. I don't think they really thought this one out, especially with a thirty day retention policy. Right? Goes against zero day retention, HIPAA, everything. It's, actually bonkers and wildly irresponsible of them to roll that out. But, anyways, that doesn't change the fact that for everyone else, if you don't care about every single thing that you send, to, Claude and every single thing it sends back to you being held for thirty days, and it can be reviewed by an actual human. If you don't care about that, which is the biggest red flag ever, I don't necessarily care.
Jordan Wilson [00:06:43]:
Right? But probably if you're a big enterprise, I'm a small business owner. If you're a big enterprise, I wouldn't be touching it. But if you don't care, let's talk about what's cool, and why it's useful. So Fable five is generally available right now to paid users on claw.ai on the front end. If you're using claw desktop, via the API, you know, AWS, Amazon, Benchmark, all the normal places, Vertex AI, Microsoft Foundry. Here's the other thing, though. It's only included in those plans through June 22. So if you are a normal Claude subscriber, so if you're on the $20 a month pro plan or if you're on the 100 or $200 a month, max plan or a team plan, yeah, at least right now, it's only gonna be available through June 22.
Jordan Wilson [00:07:28]:
So Infropic is hoping everyone just gets hooked on this and will then pay for it via the API. And it is expensive via the API. It is twice as expensive as Opus 4.8. But here's why the model is useful. I mean, it's really, really good. Alright? Especially if you do anything with building three d worlds, if you do anything with, individual software engineering, coding, you know, game, video development, any of those things, it's by far the best out there for individual I think for certain individual outputs, individual artifacts, like if you're designing a website, Mythos is gonna be by far the best. Right? If you're needing to build a three d world or a simulation, Mythos is by far the best. When it comes to knowledge work tasks, when it comes to long running, keeping the context, those things, I still don't think it's the best.
Jordan Wilson [00:08:22]:
I I would still choose five five pro for that. But, I mean, engineering teams are gonna find it extremely valuable. Knowledge workers, if you are working in finance data research, you know, anything like that, I think it's good. Also, if there's enterprises that are already standardized in AWS, Bedrock, Vertex, or Microsoft Foundry. You know, you do have state of the art best model in the world by most benchmarks available now. Again, if you don't mind handing literally everything over to Anthropic for thirty days and knowing that a human may review said everything. So, yeah, pretty big consideration. Here's what Anthropic said about it.
Jordan Wilson [00:09:02]:
They said today we're launching Claude Fable five, a mythos class model we've made safe for general use. Fable five's capabilities exceed those of any model we've ever made generally available. It's state of the art on nearly all tested benchmarks of AI capability showing exceptional performance in software engineering, knowledge work, vision, scientific research, and many other areas. The longer and more complex the task, the larger fables fives lead over our other models. Alright. So, yeah, like I said, on most benchmarks, it took the cake, but not on all of them. On some important benchmarks, like the ability to keep, context over a long period of time, it loses out. In terms of instruction following, it loses out.
Jordan Wilson [00:09:46]:
Anthropics models are usually pretty bad at the ad instruction following, over some you know, there's a new benchmark that I think is actually really cool. If you've heard of humanities, last exam, there's a new agents last exam. Didn't win out on that. Five five went out on that. So for the most part, one of the best models we've ever seen. Is it dangerous? Maybe. Right? I I I think Intrepid wants you to think that. But for the most part, if you're an average user not working at a large enterprise, I highly suggest you use it.
Jordan Wilson [00:10:13]:
I've been having a lot of fun with it. So, you know, overall, pretty good. Alright. Let's go next. Next, we have Gemini 3.5 live. I think this could change how we all communicate. So here's what it is. It is a new audio model from Google for near real time speech to speech translation across more than 70 languages.
Jordan Wilson [00:10:34]:
So it auto detects the language that's being spoken, and it generates translated speech that preserves the speech or, the speakers, intonation, their pacing and pitch, everything. So unlike turn by turn systems that wait for a speaker to finish, it actually translates continuously, staying only like a second or two behind, which is really cool. So right now, it's available for developers through the Gemini Live API and inside Google's AI Studio. So, yeah, you don't even have to be super technical. If you use Google's AI Studio, which I absolutely love and I use a lot, it's pretty easy to use. And it's also right now, which is cool, it's in preview for businesses who are using Google Meet, and also for all the users in the Google Translate app on Android and iOS. So yeah. The translate, section has been available for a little bit longer, but in terms of this being like an actual model that you can go use inside Google's AI Studio, this is new.
Jordan Wilson [00:11:31]:
So why is it useful? Well, in Google Meet, I mean, that expands the languages you can talk to people from. I think before is, like, a handful. Now it's more than 70, and from English only pairs to only 2,000 language combinations in a single meeting. So it kind of removes the awkward stop and wait rhythm of older translation apps because, like I said, it's only about a second and a half behind. And it works also in noisy environments, and it stays synced to the speaker, without any lag. So the cool thing is and actually, just let me, read this part here from Google. So they say, today, we're taking our next steps with the release of Gemini 3.5 live translate, our latest audio model for live speech to speech translation. So the it says the model automatically detects 70 plus languages and generates smooth, natural sounding translated speech that preserves the speaker's intonation, pacing, and pitch.
Jordan Wilson [00:12:27]:
Unlike turn by turn systems that wait for speaker to finish before responding, 3.5 live translates generates speech continuously balancing the trade offs between waiting for context to improve quality and translating immediately to stay in sync with the speaker. So the cool thing that I like is it can translate speech, it processes speech as it's streamed. So the cool thing is and I actually saw, a live demo of this actually, with someone on the Google team going over this. You can just choose something that you are watching in a tab. So, you know, when you're using this model inside Google's AI Studio, you can literally select your inputs, device as a tab that's open if you're using a Chrome browser. So let's say you're, I don't know, watching a, a a YouTube video or you're watching a live stream or maybe this actual podcast. Right? Maybe you're watching this, on YouTube or LinkedIn or something like that, and English isn't your preferred first language. So you will still be able to you know, you could check check that as your tab.
Jordan Wilson [00:13:32]:
You know, let's say it's Spanish, and you'll still hear my, kind of pacing. It's not my actual voice, so it doesn't do voice cloning, which is maybe a good thing. So it'll dub out with other voices, but it does it pretty much lie. Right? It's about a second or so delayed. Whereas most other translation apps, you you have to wait until the speaker is done talking, which if you listen to this podcast, it's like never. I don't stop talking. Right? So that's really cool. And, I mean, when when I first heard it, you know, I heard multiple people speaking other languages, not English.
Jordan Wilson [00:14:03]:
Right? And it translated everything to English as other people were speaking different languages to each other. So really cool. It's definitely worth checking out. And I do think in theory, this might be one of those things that could change the way that the world communicates. So I think who's gonna find it valuable? I mean, if you're a global team running multilingual meetings or customer calls, this is a no brainer. Also, if you're traveling right to be able to use this in the app, that's huge. Also, developers building voice translation into apps, you you know, no brainer. This was an extremely, impressive model from Google.
Jordan Wilson [00:14:42]:
Alright. Let's get to some smaller updates you might have missed. So next, let's go to interactive charts. Pretty cool update here, from chat g p t. So chat g p t can now turn some answers into rich interactive bar line pie and scatter charts directly into the conversation. So you can ask for a chart or chat g p t can just maybe add one on its own when it's, pulls in relevant data, that goes with your query. So at any time, you can switch any interactive chart back to a static image if you want or other chart types at that are returned, you can go back as static images. So this is available now to any user, and it's bundled in the same update as some others that we're gonna get to.
Jordan Wilson [00:15:32]:
But as full screen writing edits, table of contents for long chats, and editing messages with attach, with attachments. So I mean, here's why it's useful. Well, it's gonna make your experience in chat GPT a lot better. So, I do wish some of these things had a toggle because I don't think the average user will know this. So like I said, at times, chat GPT will invoke this on its own, but in other times, you can ask. Right? So, let's say you're doing some some research about, I don't know, benchmarks for AI models. Right? And there's, you you know, throughout the course of a long, response from Chatt GPT, there's kind of some scattered stats. Right? So in that case, you can ask Chat GPT for an interactive chart to make it a little bit easier to digest.
Jordan Wilson [00:16:16]:
So, actually, I think Google has been a leader in this space. Anthropic had a similar feature that they rolled out, but I love this feature in, Google's, AI mode. It kinda has that interactive canvas. So pretty cool. I like that this came out with CHED GPT as well. So in terms of why it's useful, I mean, that's simple. So it turns just large blocks of text. Or if you upload your own data, it can do that as well.
Jordan Wilson [00:16:40]:
So it just turns it into visuals without having to leave the chat or exporting anything to a spreadsheet to, you know, generate those charts. You can save them as well. And it just lets you kind of eyeball trends and data and comparisons in line instead of just having to read them through text. For someone like me, this is big, and this is why I like all the visualizations. Like, I love Claude artifacts. I love Gemini Canvas. Chachapiti got rid of their dedicated canvas mode, which kinda stinks. It's still kind of there through something they call writing blocks, but it's not as easy, to invoke as it used to be when canvas was just a, a dedicated button.
Jordan Wilson [00:17:18]:
So I kind of miss and wish that OpenAI didn't make this change, with Canvas or even in this case with the interactive charts. Sometimes, I think users want to be able to click something and to get this versus knowing, oh, I can just ask for it. I know in theory, it makes it better to not have to click a button, but I would say 99% of users have no clue something like Canvas or these writing blocks or these interactive charts exist in chat GBT except the times when it brings it up on its own. So, pretty cool update there from open AI. Alright. So let's go next. We have a completely new, improved, rebuilt from scratch Microsoft Copilot Studio. So here's what's new.
Jordan Wilson [00:18:03]:
This is hot off the presses. It is a rebuilt, revamped, and updated Copilot Studio with a streamlined authoring experience in a more modern AI core. So So this kind of goes along with the update that Microsoft made, just to Microsoft copilot, the actual, chat experience. So now the studio, which is kind of their builders, playground just for, you know, non developers, it's no code, low code. Alright. So this got a complete face look face lift in rebuilt from the ground up as well. So Microsoft says it strengthened how agents also follow instructions and handle complex multi step tasks for more consistent results. So it's more than just a kind of a rebrand or a refresh.
Jordan Wilson [00:18:45]:
It is just rebuilt and even how it works. Another big update is this also works with the new Microsoft IQ, which is just how your data your company's data follows you around the Microsoft three sixty five Copilot ecosystem, and that's now generally available as well. So who has access to this? Well, everyone. Because this is also generally available worldwide, and it is ready for production use. So it's aimed though at makers and admins who are building agents on the power platform or the Copilot studio stack. And this rolls out into existing Copilot studio environments rather than as a separate product, which is probably a good thing. So well, why is it useful? Because I think it targets the most common failure mode, which is agents breaking down partway through multi step work. So kind of the secondary coverage, the new orchestrator reports that Microsoft citing roughly 20% better evaluation performance with this new version and about 50% lower net token usage in the upgraded AI stack, plus the new redesigned workflows in the visual designer.
Jordan Wilson [00:19:50]:
So who's gonna find this useful? Well, if you are the one in charge of building, agents, you know, inside of Copilot Studio, obviously, you will likely, enjoy this. Or if you went into Copilot Studio in the past, you're like, I don't know why something with the UI, UX isn't adding up for me. I'm finding it a little more difficult. I think this new authoring authoring experience, might help you. So, obviously, aside from that, enterprise IT and ops teams who are deploying agents are gonna really like this. And just low code makers and business analysts who are building internal workflows inside of Copilot or Copilot Studio. So here's what Microsoft says about it. They said starting now, customers can dive into the rebuilt, revamped Microsoft Copilot Studio.
Jordan Wilson [00:20:37]:
The new streamlined authoring experience in modern AI core make it easier to build agents and workflows that deliver more consistent and reliable outcomes. Teams can move faster and create solutions that hold a better in real world use. So the big thing they said is we're making it easier to create agents that work the way you expect from the start. We strengthen how agents follow instructions and handle complex tasks, leading to more consistent results for your business scenarios. The new building experience streamlines Copilot Studio to bring essential configurations to the forefront, reducing configuration tabs from nine to four. So, yeah, just less clicking around, just an overall better UI UX for building these agents. Also, they say the new experience includes a new agentic orchestrator, a new agent building interface, and support for skills. Right? So, yeah, that one is big, especially if you are using, Claude, if you're using, you know, codex skills, if you're using skills in, I don't know, for the few people who are using, like, anti gravity two point o out there.
Jordan Wilson [00:21:41]:
Right? For anyone that's using skills, this is a great way to be able to bring those in, and work with your preapproved tech stack. So next. All right. This one kind of niche, but I really like it. Alright. So this is actually a new plugin or a new connector for Claude from, Hey Jen called Hyperframes. So Hyperframes itself is not new. Alright.
Jordan Wilson [00:22:08]:
They announced this back in April, and Hyperframes itself is just an open source HTML, the video framework. But now this is just being, released as a clawed connector. And I think this, is actually a good time to talk about it because it kinda went viral because a lot of people with Fable, when Fable was released, they found out, wait. This is actually something Fable is really, really good at using this new connector. So people were showing off all these really cool, videos that were just edited with, well, Claude and the new hyper frames, connector for Claude. So like I said, this is now an official Claude connector just released hours ago that converts and essentially harnesses Claude's reasoning power into video content. Right? So hey, Jane. Hey, Jen.
Jordan Wilson [00:23:01]:
Obviously, a very popular AI video editing platform. So now they're bringing their hyper frames technology, which I think is really cool into Claude. So you could use Claude's brain, to create videos without really knowing anything about video editing. So Hyper Frames is an open source agent tool that turns prompts, specs in documents into code first compostable scenes that renders shareable m p fours that are then usable in Claude and Claude code. So this is pretty cool. Like I said, it runs with the claud connector via MCP, but it lets you generate I mean, anything you can think of. So if you don't have a video editing team, but you can sit in front of a camera like me, and maybe that's why I'm excited about this one. And I know that you this is has been available, in codecs, but the downside with codecs is it only has five five x high.
Jordan Wilson [00:23:56]:
Right? And the best model, I will still argue, okay. Probably one a one b would be, the new kind of Mythos, Siri, and Fable with Five five Pro. One thing maybe I should talk about this more. People don't understand when you look at all these benchmarks. No one benchmarks Five five Pro. Maybe number one because it's too expensive, but most people just don't use it. I still think five five pro for a lot of my multi step, you you know, prompts that it re require calling multiple tools, all these things. I still think it's better than Fable five, but you can't have that in codec.
Jordan Wilson [00:24:32]:
So that's why I don't think I've really used this new HeyGen, plug in inside codecs because it could really benefit from, like, five five pro, which is why I'm excited for this clawed connector because you can use it with the new Fable five model. So I'll I'll probably be testing that out, or I'll actually have codex. Yeah. I've I've been having codex run my Fable, like, around the clock. It just checks every hour how much uses do I have left on my $200 clod plan. And if I still have anything left, codex controls clogged code and it just builds me stuff and I just look at what it built. So this is one I'm gonna add, to the list now to see if it can generate any cool videos with this new, with this new feature. So, I mean, who's gonna find it valuable? I mean, number one, if you're already a video editor and you just need to do some more cuts and you don't have time, I think that's it.
Jordan Wilson [00:25:20]:
But anyone who's making trail, training videos, sales video, customer supports, content creators, marketers, or if you're just a developer, that feels good working with agents, but you're not necessarily a video editing creative, I think this is gonna be a big one as well. Alright. Two more quick ones for you next. A new one that really slipped under the radar. And I think this is because it's a Microsoft Copilot, but I only think this is available right now Copilot online. So not Microsoft three sixty five Copilot that most people use. So if you have a paid Copilot online subscription like I do, this is available. So I actually didn't even see an announcement post from this.
Jordan Wilson [00:26:05]:
I was just having some agents write every to plan for the show, I have a bunch of agents that go and do work. And it actually saw this new feature on its own. I didn't see it announced literally anywhere. But when I checked last week, it wasn't there, and it's available this week. So this was rumored or leaked back in February. Haven't seen anything since until this week. So here's what's new at least if you have a Copilot online, and I think this is pretty cool. It's the new Microsoft Copilot team of advisers.
Jordan Wilson [00:26:34]:
So this is a round table of AI generated adviser voices that appear on screen when you ask a question. And each one has a distinct perspective, area of expertise, and way of reasoning. So the advisers challenge build on and rethink, off each other in real time, which is really cool. Working toward a fuller conclusion rather than handing back one answer. So this is cool. You know, it's this multi voice panel. So not just like the right. Right.
Jordan Wilson [00:27:04]:
So obviously, you might be thinking of something like the audio overviews, from notebook l m, which are amazing. And they do I actually have a live version of that that you can jump in and talk with the two hosts, on the deep dive. But this is a little different. So let me read what Microsoft sets. So they say team of advisers brings together a round table of AI generated voices to help you explore ideas from different angles. Ask a question in a panel of AI advisers appear on a screen, each with a different perspective, each ready to challenge, build on, or rethink the others in real time. Every adviser brings a distinct way of reasoning, a different area of expertise, and a unique point of view. This isn't an ordinary chat.
Jordan Wilson [00:27:45]:
It's a panel of AI advisers thinking together, questioning one another, and working toward a fuller, more grounded conclusion. Whether you're considering a decision, generating ideas, organizing complex information, or trying to see a problem from a new angle, team of advisers help you explore more possibilities with clarity and confidence. These advisors are designed to spark insight, surface alternatives, and support your thinking, not to replace human expertise. The advisors you see on screen are generated by AI. Their perspectives are meant to help you reflect, compare interpretations, and consider possibilities. So this is really cool. And, I mean, who has access? So, yes, this is, an experiment right now, but I do believe anyone that has Copilot online, this is available now if you look under the experiences, tab, and if you have Copilot labs enabled. So you don't have to be on a wait list as long as you have a Copilot online account, not Microsoft three sixty five Copilot.
Jordan Wilson [00:28:46]:
Copilot online, so like a personal account, it should be available for you now. So I think this is great. Right? To and and what's cool is they have these prebuilt, kind of these prebuilt experiences I'm looking through. Like, they have historical figures roundtable, which is cool to talk with people from the past, a product discussion roundtable, brand naming workshop, lifestyle exploration. So So it's not like you have to go give personas to the handful of advisers that are gonna be on there. They have these ones prebuilt, which I think is a smart move. You know, and it looks like they have about a dozen different ones, You know? So I I named some of them, but they have ones on relationship topics, startup idea clinics, business watch room, general decision making. So, you know, they essentially have some good templates set up and ready to go.
Jordan Wilson [00:29:36]:
I might do an episode on this. I don't know. I'm gonna have to test a little bit more, but I think it could be fun. Alright. And last, this one, again, might seem small, but my gosh, I like this one. And why didn't we have this, like, six to nine months ago? Alright. This one, sending emails in chat g p t. Yes.
Jordan Wilson [00:29:59]:
We've had chat g p t connectors. You've been able to draft an email, and there used to be a button that you could click. Right? Like, Claude has this that you can click, and it'll open said draft in Gmail. Now you can literally send emails inside chat g b t. And that may sound like a nothing burger to a lot of people, but if you're someone that like me, I get spammed so much. Right? And then I miss out on emails I actually need to be reading and responding to. And the cool thing is is you can go through and edit this as well. So this is a great way, especially when I have these long threads.
Jordan Wilson [00:30:38]:
Right? I have these long threads with, you know, 30, you know, 30 back and forth, and I forget because I dozens of these and I suck the email. And I use Chad GPT and Claude a lot to go through my email. So this is actually cool. So I can, the cool thing is is it has that writing blocks, so you can actually go in there and edit. So you can have Chad GPT go through, be like, hey. You you know, what are my emails waiting on the response? It's gonna tell you. Be like, hey. You know, give it your three bullet points that you want it.
Jordan Wilson [00:31:07]:
It's gonna type up an email. You can click edit to make sure that it's correct how you want it, and then just click send. So, if you're connected to Gmail or Outlook, you can now ask chat g p t to draft and send emails directly from the same conversation. So the drafting and inline editing happen in the new writing blocks feature. So like I said, that's the thing that kind of replaced the canvas dedicated toggle. So now writing blocks are something if you're asking about emails, usually, they're always gonna automatically open in a writing block. If not, just ask for a writing block. And then that new step is then being able to hit send.
Jordan Wilson [00:31:42]:
So there's literally, you know, if you're watching the live stream version here, you see it on this tweet that I have up. But it literally has a send button and then it sends. So you obviously have to already authorize, this in your email. So, you know, if previously, if you only gave it read only access inside the Gmail or Outlook connector, you would have to go in and enable right access. And if you're on a team's plan, right, your administrator would have to enable, that kind of step or that tool call as well. So, right now, it's only available for paid subscribers. Alright? And it's on the web only. So it's web only right now, not on the mobile app yet.
Jordan Wilson [00:32:29]:
But I do think that they're gonna be, rolling that out pretty soon. So, I mean, why is this useful? Well, if you're someone that uses ChatGPT to go through your emails to help you triage, prioritize, go through long threads like me, I think you're gonna find it immediately useful. And then you can obviously highlight any line, request targeted changes, and then accept or reject edits without copying anything. Right? So no more. I've I've gone through this before. Right? Sometimes when you copy and paste things, you know, it's weird formatting. It's just background. Right? The it it has all these crazy indents.
Jordan Wilson [00:33:03]:
You know, got to the point where I'd be doing these things, and I'd have to paste them into a text edit, you know, stripped all the formatting. And then at that point, I'm like, okay. I could have just written this myself. So this is actually really nice. You know, again, small quality of life feature that if you're just spending all week playing with, Fable, you'd probably miss some of these smaller things in chat gbt that I think might make a big difference. Alright. That is a wrap for this week y'all, but I'm telling you, it might have seemed a little lighter on features, but I think the AI world knew and expected that Anthropic's Fable five was coming out. So presumably so, it was a lighter week on other features.
Jordan Wilson [00:33:42]:
Right? There were some codex features that, I decided not to even go over. I thought they were pretty cool for developers, but not for everyone else. That'll be in the newsletter today. But I I do expect next week is going to be a big one. Next week and the week after because, presumably, we're either gonna be seeing the super app announcement from OpenAI, maybe GPT 5.6, which I expect to be at Fable five, levels, or we might see Gemini 3.5, pro, again, at Fable levels. So hot AI summer might have taken a brief pause at least when it came to releases not named fable five this week. But I guarantee you next week on this show, there's gonna be a ton of big ones. So thank thank you for tuning in.
Jordan Wilson [00:34:27]:
If you haven't already, please make sure to subscribe to the podcast. If you're listening on Spotify or Apple Podcasts, it not only helps me do this for you, but it also helps other peep other, nontechnical business leaders, find the podcast and hopefully help them. I wanna continue to do this as long as possible. Got some, thunder alerts going off. So thank you for tuning in. Hope to see you back, tomorrow and every day for more everyday AI. Thanks, y'all.