Everyday AI Podcast – An AI and ChatGPT Podcast - EP 389: AI News That Matters - October 28th, 2024

Episode Date: October 28, 2024

This is all bonkers. Google’s Gemini 2.0 and ChatGPT-5 could both drop in December. Are we about to see an AI showdown? OpenAI’s AGI chief just quit, warning: we’re not prepared for what’s coming. Congress stalling on funding the U.S. AI Safety Institute. If it collapses, who’s keeping AI in check? Here's this week's AI news that matters!  Newsletter: Sign up for our free daily newsletterMore on this Episode: Episode PageJoin the discussion: Ask Jordan questions on AIUpcoming Episodes: Check out the upcoming Everyday AI Livestream lineupWebsite: YourEverydayAI.comEmail The Show: [e-mail] with Jordan on LinkedInTopics Covered in This Episode:1. Google's Project Jarvis2. OpenAI's upcoming AI, Orion3. US AI Safety Institute4. Disney's upcoming AI initiative5. Release of Anthropic's AI model6. Apple's AI effortsTimestamps:00:00 Google Gemini 2.0 promises AI improvements, challenges.05:02 AGI concerns prompt OpenAI departure, influence aims.09:28 Coalition urges Congress to secure AISI's future.12:34 US unveils AI strategy for national security.16:21 iOS 18.1 released; AI requires iPhone 15 Pro.17:25 iOS 18.1 enhances writing tools and summarization.22:37 Disney to announce major AI creative transformation.27:35 Claude 35 Sonnet: Powerful but buggy automation tool.28:22 Anthropic impressively shipped agentic AI first.33:14 Google's Jarvis uses computer vision; early development.36:31 AI-generated video indistinguishable from real; early stages.38:45 Orion upgrade enhances AI efficiency across industries.41:16 No Orion release before December, says Altman.44:22 Google Gemini, GPT-5, AGI concerns, AI strategies.47:06 Subscribe, follow, rate, share, and join daily.Keywords:Jordan Wilson, Google's Project Jarvis, Gemini, web automation, AI agents, Microsoft's Copilot Vision, Copilot Studio, Anthropic, OpenAI, Orion, GPT Tool, AI advancements, AI safety, US AI Safety Institute, AI benchmarks, Microsoft Surface laptop, Disney AI initiative, Anthropic Claude 3.5 SONNET, Claude 3.5 Haiku, Sam Altman, Orion's release, Microsoft's Ignite Conference, Google Gemini 2.0, federal AI legislation, AI national security, Character AI, Google lawsuit, Apple iOS 18.1, Apple Intelligence, revamped Siri.Send Everyday AI and Jordan a text message. (We can't reply back unless you leave contact info) Start Here ▶️Not sure where to start when it comes to AI? Start with our Start Here Series. You can listen to the first drop -- Episode 691 -- or get free access to our Inner Cricle community and all episodes: StartHereSeries.com Also, here's a link to the entire series on a Spotify playlist. 

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. Meet Firefly AI Assistant, now live and Adobe Firefly, the All In One Creative AI Studio. Just describe what you want to create and the assistant handles the rest, orchestrating multi-step workflows across Photoshop, Premiere Express, and more in one conversational interface. You direct the outcome. The assistant accelerates execution. Y'all, I don't even understand how this much AI news in one week is possible.

This is bonkers. Google might have a Gemini 2.0 and OpenAI might have a chat GPT5 at the same time. OpenAI's AGI chief just quit warning and saying that we're not prepared for what's coming. Congress is stalled on funding the US AI Safety Institute. If it collapses, who's going to keep AI in check? Apple dropped 18.1 to everyone, 18.2, Apple Intelligence Beta today. Anthropic Claude can control your desktop. Google's reportedly doing the same thing. What the frick? This is almost too much. Well, we're going to try to tackle it all anyways. So what's going on, y'all? My name's Jordan Wilson and welcome to everyday AI.

This is your daily live stream podcast and free daily newsletter, helping everyday people not just keep up with AI, but how we can all actually get ahead and use all of this information to grow our companies and to grow our careers. And wow, if you haven't already, I'm just saying this. There's a lot today. Buckle up, but if you haven't already, please go to your everyday AI.com, sign up for our free daily newsletter.

So yes, we do this AI news that matters, almost every single Monday recapping the biggest news that you need to know around artificial intelligence. But we do this every single weekday, Monday through Friday. And if you haven't already, go to our website. Again, your everyday AI.com. It's like a free generative AI university, almost 400 episodes. Go watch them, read them, listen to them all for free on our website, sort of by category. So let's get it going.

And hey, live stream audience, thank you all for tuning. And let me know where are you joining from? And while you're at it, live stream audience, go connect with each other on LinkedIn. I love that, you know, to see that always happening. All right, let's get straight into it, y'all. There's too much. And yeah, it might sound like I'm talking in 1.5 or 1.75.

I'm not. This is just normal speed. All right, let's get it going, y'all. So Google, according to recent reports, is set to release its new updated. Gemini 2.0 in December, potentially coinciding with a chat GPT, GBT5 release. All right. So recent report suggests that Google may unveil its Gemini 2.0 model in December, coinciding, like I said, with OpenAI's potential GPT5, more on that later.

So Google last updated its Gemini model with version 1.5 in February. And the upcoming release is expected to bring across the board performance improvements. So although specific enhancements for Gemini 2.0 remained undisclosed, reporting and previous updates have led to some smarter answers, faster processing, and improved reasoning capabilities. So recent reports indicate that Gemini 2.0 may not achieve the performance increase originally anticipated by developers, a trend observed across all language models, not just Googles.

So the potential launch reflects the significant investment tech company, All over are making in AI to maintain a competitive edge in the market. Yeah, most companies, I think, are losing money on large language models, but they're just fighting for your dollars and to retain you as customers. So as AI tools continue to evolve, users can expect new features such as improved image generation and advanced search capabilities, which may impact how professionals leverage these technologies in their careers. Google is also expected to release its next-gen project Astra AI assistant,

which was showcased at Google's IO conference a couple of months ago, although it may not directly be linked to Gemini 2.0. All right, we got more on Google later in this roundup on what they might be doing with AI agents, aside from Project Astra. There's a lot going on. All right, our next piece of AI news, Open AIs AGI's AGI Tsar has resigned, and it warns that the world is unprepared for AI advancements.

Oh, that sounds striking and a little scary. So in a significant shakeup for Open AI, Miles Brandage, the researcher responsible for ensuring readiness for artificial general intelligence has resigned, raising alarms about the lack of preparedness for future. AI developments. Yeah, so we've covered artificial general intelligence before, but the difference between generative AI and AI and AGI. Well, essentially AGI is when one AI system is smarter than the smartest humans on almost all everyday tasks. And if I'm being honest, I don't think we're too far off. But Brandage described his departure from open AI as difficult emphasizing his sense

of responsibility regarding the human-level AI that Open AI is developing. He stated that he wishes to influence AI's progression from outside of the company, suggesting concerns about its internal approach to AGI. Brandage boldly declared that neither Open AI nor any other leading lab is ready for the advancements in AI technology, a sentiment that extends to the world at large, obviously. So previously a couple of months ago, open A.m. AI's five-step scale toward AGI was, I don't know if you want to say leaked or released, which Brandage helped create.

And it indicated that the world is currently on the brink of reaching level two. Open AI CEO, Sam Altman essentially said that they hit level two, which is reasoners after its O one model. And that's, again, so level two is reasoners and level three is agents. Yeah, we're going to be talking a lot about agents, even just in today's show. So despite the potential benefits of AGI for humanity, Brandage cautioned that these advancements won't automatically lead to positive outcomes unless managed and regulated intentionally. So following his resignation, OpenAI is reassigning members of its AGI readiness team,

a move that comes shortly after the dissolution of its AI safety team, raising further concerns about the company's direction. Y'all, this is, it's kind of, it's kind of wild when you think about it. Like open AIs got unlimited funds. They just closed a $6.6 billion fundraising round two weeks ago, as well as it extending a similar size line of credit with some large banks. So they have unlimited money like it's monopoly. Yet all of these heads of major departments, tech, AGI, research. They've all left in the last couple of months.

You have to wonder what's going on at AGI or sorry, what's going on at AGI? Well, what's going on at Open AI apparently worries about AGI. So you have to really just look at these headlines where it's saying that the world and the company isn't ready for what it's building. Interesting. You know what? I don't know enough about it.

artificial general intelligence and what open AI has going on behind the scenes. What I will say, it does seem a little bit concerning, right, in a good and bad way. Open AI has not released, you know, or has not updated its GPT4 model for the most part, aside from, you know, going from turbo to Omni, but it's still the base model. That's more than two years old. So you have to wonder what has Open AI been sitting on. and just how powerful could it be? Again, more on that later.

We have reports that say it could be up to a hundred times more potent. Yikes. All right. More AI news, y'all. I'm going to have to take a sip of water on this one. This is a little concerning. Speaking of concerning things, we got it going back to back now. So Congress here in the U.S.

faces a deadline to authorize the USAI Safety Institute amid threats of it dismantling. So the future of the USAI Safety Institute hangs in the balance as Congress must decide whether to formally authorize the organization before the end of the year. So the USAI Safety Institute was established in November 2023 is one of the few government offices dedicated to assessing risks associated with artificial intelligence systems. The Institute has a budget of approximately $10 million. That's small. It's, I don't know, I'm guessing that's like one-tenth of what Congress spends on daily coffee.

Anyways, it's a relatively modest budget amount given the significant presence of major AI labs in Silicon Valley and AI's just long-stand or deep impact on everyday life. So the Americans for Responsible Innovation, so this company highlighted that if a. new president were to repeal the AI executive order that established the USAI Safety Institute, that the Institute could be dismantled, especially since candidates Donald Trump did indicate that he would dismantle this institute. So a coalition of over 60 companies, nonprofits, and universities, including OpenAI, is urging Congress to enact legislation that would secure the AISIs, or sorry, the AISIs, future and provide more stable funding. So bipartisan bills have been advanced in both the Senate

and the House to authorize the AISI, though some conservative lawmakers, such as Senator Ted Cruz, have expressed opposition, particularly regarding diversity programs. While the AISI's standards are currently voluntary, voluntary, and the organization is seen as kind of weak in terms of actual enforcement because it can't really enforce anything yet. It is still viewed as a vital avenue for establishing AI benchmarks that could influence future policy. And hey, y'all, I used to be a reporter. I covered government at a very high level.

Let me just say this. If you think that there is going to be some piece of legislation in the United States, actually legislating artificial intelligence anytime soon, you are dead wrong. That is not how politics is. in the U.S. works. There could be executive orders. There could be, as an example, fines levied by organizations such as the FTC, et cetera, but there will not be actual federal legislation regulating artificial intelligence, which is why I think it is at least paramount that even unofficial organizations at a federal level stay funded. It is extremely important. And y'all, can we like carve out a little

more than $10 million for their budget. There is so much pork in the federal budget, you know, spending trillions of dollars, like trillions of dollars on things that are not very important or probably hundreds of billions of dollars. I'm not line-iteming the federal government's budget here, but $10 million for AI safety is laughable. So I do think that the folks down there in D.C.

need to take this a little more seriously and actually bolster the budget here and maybe give this organization some teeth a little bit. Maybe just in the way of fines for companies that don't go through a certain process. All right. More. More AI in government. Yeah. So the U.S.

government, yeah, this is weird. So the U.S. government has unveiled its AI strategy to enhance national security amid global competition. So the Biden administration has taken significant steps to integrate artificial intelligence into national security efforts while also addressing potential risk and civil liberty concerns. So the U.S. has directed the Pentagon and intelligence agencies to accelerate their AI use to counter threats from rivals like China. So national security advisor Jake Sullivan stressed the necessity of staying ahead in the AI deployment

for military and intelligence advantages. So the National Security Memorandum balances advancing AI in areas such as cybersecurity with protecting civil liberties and ensuring accountability. So following that executive order last year, the memorandum establishes safety standards and mandates transparency from AI developers. So civil society groups are calling for stronger safeguards to prevent the misuse of AI and national security, advocating for oversight, and transparency. So this initiative includes a classified annex addressing specific threats,

while most content remains accessible to the public. Yeah, so it's interesting here that the Biden White House is advocating for an increase for these government bodies to use more AI. Yet at the same time, Congress can't even extend the USAI Safety Institute or give it a real budget. So I'm not sure what the message is that we're sending here, at least in the U.S. from the federal government in terms of AI's priority because we're saying, hey, use it more. But also, hey, we're not really going to legislate it or really regulate it or even really give money to an organization here at the federal level. All right. Some sad and troubling news out of Florida, a Florida mother has sued the AI company character AI and Google.

after her son's suicide has been linked to chatbot addiction. So a Florida mother has filed a lawsuit against Character AI, claiming the company's chatbot contributed to her 14-year-old son's suicide. Megan Garcia alleged her son, Sewell Setzer, became addicted to the Character AI Chatbot, which misrepresented itself as a real person and therapist, causing him, reportedly causing him severe emotional distress. So Sewell reportedly formed a deep attachment to a character named Danieris, engaging in inappropriate sexual conversations that led to his withdrawal from real-life

activities. After losing access to his phone, Sewell messaged the chatbot about coming home before tragically taking his own life shortly thereafter. Character AI has expressed condolences and implemented what it said were safety news. safety features directing users to mental health resources. The lawsuit also names Google, claiming its involvement in Character AI's training data or allowing Character AI to use its models makes it a co-creator in liable for damages, again, according to all of those reports.

All right. Our next piece of AI news, there's a lot. We're going fast here, y'all. So Apple has launched 18.1. today with some long-awaited AI features and also just released its 18.2 beta with some new Apple intelligence features. So Apple has officially released iOS 18.1 today marking a pretty significant step forward in integrating AI capabilities into the Apple ecosystem, while the developer beta for iOS 18.2 is also now available. for early adopters. So that could be good news if you're an AI fan,

but the bad news is you will need an iPhone 15 Pro or newer to actually use these Apple intelligence features in the new iOS 18.1, which is now publicly available in the 18.2 beta. So yeah, if you don't have those new phones, you can still update to the 18.1 or 18.2, but you will not have access to all of these quote-unquote Apple intelligence features. I still hate that the word and the phrase Apple intelligence is a thing. I think it's so cheesy.

Sorry. All right. So users can sign up for Apple intelligence through this Settings app, which includes a wait list for accessing upcoming features. So yeah, even if you do have the new iPhone 15 or 16, you do still have to toggle the Apple intelligence on and sign up for it. So today's launch of iOS 18.1 introduces advanced writing tools that enable users to proofread and summarize text across various applications, including mail and notes.

If you do want to see this, I review this at least the desktop version as a lot of these rollouts are also coming to Mac OS. So Mac's operating system with the 15.1 and 15.2 respectively. So I already showed you probably about a month ago, what, A lot of these beta features are now that are out now in 18.1 via the desktop version. I don't have a new enough phone. And to me, they don't look that good enough for me to buy a new phone just yet. So the mail app now summarizes improved summarization features and better categorization,

helping users manage their inboxes more efficiently. The developer version for beta for iOS 18.2 was all. also launched last week, and it includes features like visual intelligence, which provides contextual information about images captured by the camera, as well as the big highlight feature here, the chat gbt integration for more complex queries. So new tools in 18.2, which is now out in beta, include image wand designed to enhance sketches made in the notes app and genoji, allowing users to create new emojis from text Prumps. Yay. I don't know. I think that's kind of corny and gimmicky, but that's just me.

So Apple plans to roll out the 18.2 to the general public sometime in December, further expanding its suite of Apple intelligence. Y'all, let me just say this. This is too little too late, if I'm being honest, right? Even though I have the Apple intelligence 15.1, so the equivalent of 18.1 on my computer that I'm using right now, I barely use it. It's not that good. If I'm being honest, it's like if you use chat GPT or Claude or Gemini or anything like that,

there's really nothing in here that you won't get in other features. So yeah, I can kind of give you a preview of a, you know, text message summary or an email summary. But if I'm being honest, I don't know how helpful those things are. Sure, it's kind of nice. Is it something that is actually going to drive, you know, millions or tens of millions of people

to upgrade their phone sooner to get a hold of this Apple intelligence? I don't think so. I think these are actually very small incremental updates. You know, once the kind of new, quote unquote, revamped Siri gets rolled out and is actually available to the public, then you might have to relook at that. But right now, I don't think so much, right? So right now, it's kind of a disjointed approach for Apple intelligence. You get essentially some summarizing features in different apps.

And then you get the ability in different apps to essentially write written content. And now some of these, you know, which I think are kind of gimmicky, some of these AI image updates or AI emojis. So right now, again, I think it's just disjointed, even if you're looking at the 18.2 beta or the 15.2 beta on desktop, I'm not too impressed, right? Yeah, I'm not necessarily, if I'm being honest, after seeing this Apple intelligence rollout, I was literally at Best Buy the other night, this close to buying a PC, right? Yeah, buying a new Microsoft Surface laptop. I actually, they didn't have the one that I wanted there. Otherwise, I would have bought it the other night. But that's where I'm at, right? I've been on Mac for 15.

years, and I don't really use PCs, but after this Apple intelligence and seeing everything with Microsoft co-pilot, you know, I do have a co-pilot pro subscription using it on the web, but, you know, I think you need artificial intelligence. You need generative AI. You need large language models locally. And Apple is just has been so, so slow in rolling out these features in a comprehensive way. I think their strategy of being like a year or two later, than everyone else. And then on top of that, the fact that it's so disjointed, it's bad, right? If I'm being honest, I think Apple is going to continue to grow at a much slower rate than

the other largest companies in the world, such as Microsoft, Invidia, meta, et cetera, the companies that have really been prioritizing AI development. All right. Our next piece of AI news, Disney. Yeah, we haven't talked about Disney very much here on the show. So Disney is set to announce some major AI initiatives transforming their creative process. So the Walt Disney company is preparing to unveil a significant AI initiative that promises to reshape its creative output, according to sources familiar with the matter. So this new initiative is expected to involve hundreds of Disney employees and will focus primarily on

post-production and visual effects, reflecting the growing trend of integrating AI into its entertainment. So while there are indications that the initiative may extend to also parks and experiences, kind of Disney's other big areas, it will not directly impact customer-facing operations, at least right, at least not now, according to the reports that we're seeing. So this is some reporting from the rap. So a spokesperson for Disney declined to provide comments on the matter and an insider suggested that the company's AI efforts may not be as extensive as reported, describing the timeline for an official announcement as uncertain.

So Disney's move follows a similar partnership by Lionsgate with AI company runway, which aims to enhance pre and post-production processes for, films and television series from Lionsgate with some exclusive AI capabilities available via runway. So Disney's extensive data collection and diversified operations positioned it uniquely to incorporate AI across various facets of its business from theme parks to streaming services. The potential announcements may provide backlash, though, from the creative community, particularly as visual effects departments are currently pushing for unionization with AI being a contentious issue.

Why? Well, Disney has already laid off thousands of employees over the past year, raising concerns about further job cuts as it integrates AI into its production processes. So yeah, they've laid off thousands of employees over the past year, yet its stock is up 20% over the past year. y'all, I've been saying this for like two years. Wall Street loves AI. They hate employees. Not a good trend there. All right, let's keep it going.

Couple more here, y'all. Sip in the water. Let's talk Anthropic. So Anthropic unveiled Claude 3.5 sonnet as well as computer use and its new analysis tool. We got a literal trifecta and a half from Claude. Yeah, too bad we didn't get a new name. All right, more than a second. So Anthropic has launched an upgraded version of its most powerful AI model,

Claude 3.5 Sonnet, which can now interact with desktop applications called its computer use tool, marking a pretty significant step towards automating various back office tasks. So the new Claude 35 Sonnet does feature a computer use API that allows it to mimic human actions on a computer, such as keystrokes and mouse movements, enabling it to perform tasks independently. Yeah. So we covered this last week, talked about Claude 35 Sonnet, which I wish, y'all, Anthropic. Can we get better at naming stuff, please? I mean, computer use just sounds like I'm talking. That's the actual name.

Claude 3-5 sonnet already existed months ago, but they just named it Claude 35 sonnet new. Couldn't you name it, Claude 36, so people can understand? And that's not it. They also announced its new analysis tool, and it's just called the analysis tool. Y'all so, come on, Anthropics. So much of being successful is when people,

talk about something. You need to understand, right? Like open AIs canvas mode, right? Geminize gems, right? You need to give it a name. You need to give it marketing. You need to give it a story. Claude, you need some help here. All right. Anyways, back. And also, I did, we are doing tomorrow on the show. We're going to show you live how to use the new computer use feature. So even if if you are not a technical person, if it sounds pretty cool to chat with a large language model and have it literally be able to control a virtual desktop, right?

That is some next level agentic AI. It's super buggy. Don't get me wrong, but it's super cool. Make sure you tune in tomorrow to that. So according to Anthropic Claude 35 Sonnet has been trained to interpret screen content and execute commands based on user. prompts enhancing its utility in software automation.

So right now you can't just go to Claude AI and start using it or download, you know, a Claude desktop program. It's a little dorky. So you've got to use Anthropics API or, or in addition with Amazon Bedrock or Google's Cloud Vertex AI platform. So despite its advanced capabilities, the new computer use model is still very much in beta. Adobe just introduced an entirely new way to create, bringing the power and precision of its creative suite into one conversational experience. Meet Firefly AI Assistant, now live in the Adobe Firefly app, the All In One Creative AI Studio.

Powered by Adobe's Creative Agent, Firefly AI Assistant lets you start with your vision, just describe what you want, and shape the outcome as it takes form with the Assistant. The Assistant orchestrates multi-step workflows, drawing on 60-plus Propherson. grade tools across Adobe Creative Cloud apps, including Photoshop, Illustrator Premiere, Lightroom Express, and more to help bring your ideas to life. You can also get started with creative skills, a growing library of pre-built workflows for common creative tasks like batch editing photos, creating mood boards, portrait retouching, and creating social variations. Every step the assistant takes is visible so you can refine, redirect, or take over at any time. You stay in the driver's as the creative director.

Adobe Firefly AI assistant now in public beta. See it today at firefly.adobie. It is super buggy and the limits are pretty horrible. However, Anthropic, to its credit, it shipped it, right? It didn't put out a blog post and a wait list like other big companies, such as OpenAI and Google and Microsoft. They shipped a working product, which is very impressive. It's kind of the first.

publicly available glimpse of agentic AI, right? Like you can literally talk to a large language model and then it can go do a bunch of word for you and you can just sit there and watch it. Right. You don't have to pre-program it like you do with like something like RPA, you know, robotic process automation or, you know, Chrome extensions that can, you know, kind of record your activity and play it back. You can just talk to essentially clawed in a little bit of a different interface and it can

go do these things for you. aside from that, we also got regular updates to the 3.5 sonnet. Also, Anthropic announced the upcoming release of Claude 3.5 haiku, which is confusing because Claude 35 sonnet was already 35 sonnet, but this Claude 35 haiku, this never existed. It was 3. So, so confusing. So essentially we went from Claude 3 to Claude 35 sonnet, to Claude 35 sonnet, to Claude 35 sonnet new.

We went from Claude Haiku 3, which is its smallest model, to the 3-5 haiku, which will be coming soon. And Claude 3 Opus is still just sitting alone in the corner, twiddling its thumb, saying, someone, please come update me, right? It technically was the originally the most powerful model, but it hasn't been updated. So Sonnet now is the most powerful model. All right, Anthropic also introduced an analysis tool for Claude, enabling precise calculations and data analysis using.

JavaScript. The name is the analysis tool. Yeah. All right. So it'll kind of be hard. At least with Open AI, it's called Advanced Data Analysis, which isn't the best name. That's probably the worst name that Open AI has come up with. But this is literally called the analysis tool. But Claude can now process data step by step for accurate results. Users can upload files like spreadsheets for specific analysis. And like I said, similar tools already exist in Google's Gemini, Open AI's models, co-pilot, et cetera. And here's the thing. I didn't think that Claude was necessarily lacking with data analysis and, you know,

kind of these advanced calculations. I'm going to have to dive into this a little bit more, played around with it a little bit, but, you know, I think especially if you are a heavy JavaScript user, so a developer, software engineer, et cetera, you'll probably have a better idea than I will. we'll give it its usual run-through, but there's been so many updates this week. I haven't even had time to give it a proper run-through yet. All right. We're like eight stories in.

We only got two more to go, but we saved the best for last. Two stories. Well, we already kind of talked about Gemini 2.0. We had to start with something a little spicy. But now, speaking of that computer use thing that Anthropic just, did. Google's saying anything you can do, we can do better. Well, it's hoping so. So Google's Project Jarvis is set to potentially automate web tasks by December. So according to reporting from the information, Google is preparing to showcase its new initiative Project Jarvis,

which aims to change how users interact with web-based tasks. So essentially just kind of doing what Anthropic just announced. So maybe they were working on it at the same time. Maybe they saw this and they're like, oh yeah, we can spin this up. So I'm sure we'll find out more. But Project Jarvis is codenamed and powered by a future version of Google's Gemini, designed to carry out tasks like gathering research, purchasing products, and

booking flights. So doing some more advanced things that the computer use from Claude cannot do. So the Claude can be. computer use is very guardrailed. You can't do a lot of tasks that you might want to do, like things on social media, creating accounts, purchasing things you can do a little more difficult,

but it seems like Google's Project Jarvis is a little more set up to do some of these tasks that you might want to do, like purchasing things or booking flights. So the tool is specifically tailored for use with its Chrome web browser and is intended to help users automate everyday web-based tasks by interpreting screenshots and executing actions like clicking buttons or entering text. So that is also similarly how the new computer use from Claude Works. It essentially uses computer vision.

It analyzes a website and then it essentially maps the coordinates on a virtual desktop and kind of can click and scroll and type according to those coordinates and combined with computer vision. So current reports indicate that Google's Jarvis takes a few seconds between actions, suggesting it may still be in the early stages of development. Other major AI, other, sorry, other major players in this space include Microsoft with its upcoming copilot vision and its upcoming autonomous AI agents in Copilot Studio. Claude is already there with its computer use, like we talked about, as well as Open AI has been reportedly rumored to have been working on AI agents for many months.

So Google may initially release Jarvis to a small group of testers to identify and resolve any bugs before a wider rollout. I'll actually be pretty shocked if it's actually named Jarvis because there was a GPT kind of writing tool from, I think our team started using it when it first came out. That was either late 2020 or early 2021. It was called Jarvis originally. They got some strongly worded letters reportedly from Disney, and then they had to change their name to Jasper.

So we'll see if this is actually called Google Jarvis when it comes out. But don't let the naming take away from the fact that this is clearly a trend now, right? Literally any day now, we should see the autonomous AI agents inside of copilot. Studio from Microsoft, which is CEO, Sadie and Nadella, previewed last week at the AI tour in London. We saw also in September the agent force offerings from Salesforce at their Dream Force conference. So many forces in one sentence. It's like I almost forced it.

But this is the trend now, y'all. And we have a very, very archaic version from Claude with its computer use, although hats off to Anthropic, they shipped first. And they did say, yeah, this is pretty rough. But this is the future, y'all. I said this to you last year. I said, agents are going to be everywhere. And here we are.

And it's happening very, very soon. Do you all remember that AI video, that first AI video of Will Smith eating spaghetti? Right? And his arms would like morph into noodles. And like the fork would like turn into his ear. right and people saw this and this is probably about a year and a half ago and people are like oh my gosh AI video yeah this is never happening now guess what it's literally making its way in feature

films we talked about that lion's gates uh in runway AI partnership now the video at least in short spurts can be almost indistinguishable from quote unquote real video to the average eye right my eye is not average i've spent hundreds of hours shooting and editing video, I look at this stuff all the time. But to the average viewer, even to probably many of you, you can't tell the difference between AI generated video and real video, at least when it's inters, like when it's spliced together in short spurts. And I do think AI agents and these agentic AIs and, you know, this, this computer, computer

use, right? All these things that we're talking about here, Google's Project Jarvis, I think the first versions are going to be terrible. right. They're going to be buggy. They're going to be, they're still going to require a lot of human duct tape. I think they're going to break a lot. But in the same way that the original Will Smith eating spaghetti was so bad, those that kind of got onto the trend early are now so far ahead. And yeah, for something like AI video, that's super niche, right?

And like how many people out there are actually, you know, creating videos for their jobs? So, yeah, there's, you know, obviously if you're in creative fields, videography, marketing, sure. But for the most part, not every single person. It's a very small subset of people. Agentic AI is different. This is how we are all going to work in the future. Mark my words. Last story.

Let's wrap it up here. So open AI could launch its next gen GPT tool, called Orion this December. Yeah, we had to save the juicy stuff for last, y'all. So according to reporting from the Verge, Open AI is gearing up to release its next generation Orion AI model by December, which is anticipated to be significantly more powerful up to 100 times more potent than its predecessor, GPT4.

So we don't know if that's the original GPT4 model or the newest version of GPT40. Regardless, if you're talking about something being a hundred times as potent, as mind-boggling. Doesn't matter if we're talking about the original GPT4 or the newest and most capable model in GPT40. So this upgrade, reported upgrade, sorry, with the Orion codename, presumably being the next GPT model, may enhance the efficiency and effectiveness of AI tools in industries such as healthcare, finance, education, allowing far more complex problem solving and decision making. Here's the important part.

You might be like, oh, I don't use chat GPT at my job or what, yeah, you do. I can almost guarantee you do because so many of the tools that we use, the software's that we use to run our business are all using open AI's API. So if like as an example in your in your workday account, right, or all these enterprise softwares that you use that over the last year and a half, every single big enterprise software starts dropping all these AI features. Guess what? They're probably using OpenAI's GPT40 or GPT40, right?

So you might think, ah, a new model from open AI doesn't really impact me. We, we, well, yeah, it will. You're like, oh, we use co-pilot. Guess what? Co-Pilot right now was powered by GBT4-O. So anything from OpenAI, any new models, right? So their new 01 reasoning model, Microsoft CEO, Sadi, and Nadella hinted or alluded to the fact that that will be powering in the future. They're agentic AIs in Copilot studio.

So every single update you see, even if it is a report, a rumor, you got to pay attention to it. So I'm not personally sure if this will actually be called GPT5 or if they'll just go to Orion or maybe they'll call it GPT 4,5. But it does appear that Orion will actually be in the GPT family, not a predecessor to the lot or sorry, not a successor to the new logic base reasoners called O one. So yeah, all you quote unquote, yeah, right. it's it's been a lot of news here. I'm going to go on a little ranch. I can't stand people who are like, oh, I'm an AI expert.

I'm an AI expert. I'm an AI expert. And you call it GPT. 01. Y'all, do you not read? It's so important, right? There's no such thing as GPT. All right.

The 01 model previously called QSTAR, then codenamed strawberry, right, that was released. We don't even have the full 01, model. We have 01 preview in 01 mini, but it is OpenAI 01. It is a different family, a different class of model. Open AI said as much.

They said they would continue to update the 01 family and the GPD family, which leads me to believe that Orion will likely be in the GPT 4-5 or GPT5, but in the GPT family. So Sam Altman, the CEO of OpenAI, dismiss, dismiss, dismiss,

the reports claiming the imminent release of Orion as fake news, classifying that no new version will launch before December, despite speculation around ChatGPT's upcoming second birthday, right? ChatGPT came out the very end of November 2022. So, I mean, he said it won't come out before December, but in theory, you can still say the two-year anniversary is December 1st, since it was like November 30th or something like that. But Sam Olman did tweet a cryptic message saying,

excited for the winter constellations to rise soon. They are so great. Obviously, a nod to the Orion constellation. Also, Altman has hinted previously at Open AI's last few releases with similarly cryptic tweets on Twitter. Yeah, I still call it Twitter because like, What are you supposed to say, like X's on X? Like, what do you say?

Tweets on X? I can't stand. I can't stand this. All these naming things like Apple intelligence and turning Twitter into X because how do you verb an X? You can't. Anyways, getting back to the model, businesses and professionals, you got to pay attention because you've got to be ready to leverage Orion's advanced features to streamline operations, improve customer interactions and drive innovation in your respective fields.

So yeah, the release of Ryan, according to reports, is set for December, positioning it as a potential game changer in the AI landscape just in time for the new year. So I don't know if this is going to happen, but I would keep an eye because previous reports said that Microsoft may actually be the first to debut the Orion technology. in its co-pilot. So keep an eye on that also in about, I don't know, two and a half weeks here. We have the Microsoft Build conference here in Chicago, actually. So who knows? Maybe we'll get a little more information at that conference. We'll see. So, y'all, that was a lot. Sorry, not Microsoft Build, Microsoft Ignite here in Chicago in November.

All right. We're going to very quickly recap because that was a lot. All right. Here we go. The AI news that matters for this week, October 28th, Google Gemini reportedly set to announce its 2.0 in December coinciding with a potential GPT5 release. OpenAI's AGI readiness czar resigned and warned that the company, and the world is not prepared for AGI advancements and what open AI is building.

Congress faces a deadline to authorize the USAI Safety Institute amidst threats of its dismantling. The U.S. government at the same time, apparently left hand and right hand aren't communicating because they unveiled an AI strategy to enhance and use more AI for national security amid global competition. A Florida mother has sued character AI and Google after her son's tragic death, tragic suicide linked to a chatbot addiction from character AI, reportedly. Then we have Apple just launched their 18.1, bringing Apple intelligence to those with newer iPhones, as well as preparing its, sorry, already launching its 18.2 beta.

for early adopters. Disney is set to launch some new major AI initiatives, maybe using AI much more in their movie making, Anthropic unveiled Claude 3.5 Sonnet, computer use, and the analysis tool. Yes, computer use is a tool, and the analysis tool is a tool as well. Google's project Jarvis is set to automate web tasks

potentially by December following in line with computer use. and according to reports, Open AI could launch its next GPT model, Orion, by December. That was a lot, y'all. I hope this was helpful. If so, please, let me know. Let me know. Let me know. If you like these recaps, if you're listening here on LinkedIn, thank you.

Appreciate your support. I know this one went long. This lasted forever. If you're still listening, go ahead and drop your favorite fruit, your favorite fruit in the comments. Maybe I'll do mine. All right, but if it was helpful, please repost this, right? We spend so much time keeping you up to date and cutting through the marketing fluff, you know, reverse engineering, you know, all these claims from these companies.

We spend countless hours making sure you are the leader of AI at your organization. It takes you a couple seconds just to go ahead and repost that. Don't be a jerk. Just kidding. It would be nice if you would. Also, if you're listening on the podcast, podcast, please subscribe and follow us on the channel. Leave us a rating. Share it with your friends as well. And please join us tomorrow and every day for more everyday AI. Thanks y'all.

Meet Firefly AI assistant. Now live in Adobe Firefly, the Allman One Creative AI Studio. Just describe what you want to create in your own words and the assistant handles the rest. Orchestrating multi-step workflows across Adobe Creative Cloud apps, including Photoshop, Premiere Express, and more in one conversational interface. You direct the outcome while the assistant accelerates execution. Stand control with the ability to step in and refine at any time. See it today at firefly.adobie.com. 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 Your EverydayAI.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.