Everyday AI Podcast – An AI and ChatGPT Podcast - EP 528: OpenAI rolls out Codex coder, Google goes full AI multimedia & more AI News That Matters

Episode Date: May 19, 2025

OpenAI made a coding splash. Anthropic is in legal trouble for .... using its own Claude tool? Google went full multimedia. And that's only the half of it. Don't spend hours a day trying to keep up with AI. That's what we do. Join us (most) Mondays as we bring you the AI News That Matters. Newsletter: Sign up for our free daily newsletterMore on this Episode: Episode PageJoin the discussion: Have a question? Join the convo here.Upcoming Episodes: Check out the upcoming Everyday AI Livestream lineupWebsite: YourEverydayAI.comEmail The Show: [e-mail] with Jordan on LinkedInTopics Covered in This Episode:Salesforce Acquires AI Startup ConvergenceGoogle AI Studio's Generative Media PlatformMajor AI Conferences: Microsoft, Google, AnthropicAnthropic's Legal Citation Error with AIDeepMind's Alpha Evolve Optimization BreakthroughUAE Stargate: US and UAE AI CollaborationOpenAI's GPT 4.1 Model ReleaseOpenAI's Codex Platform for DevelopersTimestamps:00:00 Busy week in AI03:39 Salesforce Expands AI Ambitions with Acquisition10:31 "Google AI Studio Integrates New Tools"13:57 Microsoft Build Focuses on AI Innovations16:27 AI Model and Tech Updates22:54 "Alpha Evolve: Breakthrough AI Model"26:05 Google Unveils AI Tools for Developers28:58 UAE's Tech Expansion & Global Collaboration30:57 OpenAI Releases GPT-4.1 Models34:06 OpenAI Codex Rollout Update37:11 "Codex: Geared for Enterprise Developers"41:41 Generative AI Updates ComingKeywords:OpenAI Codex, Codex Platform, Salesforce, Convergence AI, Autonomous AI agents, Large Language Models, Google AI Studio, generative media, Imagine 3 model, AI video generator, Anthropic, Legal citation error, AI conference week, Microsoft Build, Claude Code, Google IO, agentic AI, Alpha Evolve, Google DeepMind, AI driven arts, Gemini AI, UAE Stargate, US tech giants, NVIDIA, Blackwell GB 300 chips, Wind Surf, AI coding assistant, codex one model, coding tasks, Google Gemini, Semantic search, Copilot enhancements, XR headset, project Astra, MCP protocol, ChatGPT updates, API access, AI safety evaluations, AI software agents, AI studio sandbox, GPT o series, AI infrastructure, data center computing, Tech collaboration, international AI expansion.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 in 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. It's kind of AI week.

Some of the biggest companies in the world are going to be unveiling what they've been working on in the AI in large language model space this upcoming week. But what's happened in the past seven days has really laid the groundwork for what's to come. So I don't want you to spend hours every. single day trying to keep up with all of this stuff that's going on. It's so hard. That's what we do here on Everyday AI with our pretty much every single week Monday AI News That Matter segment. So that's what we're going to be doing today, giving you the AI News That Matters. What's going on, y'all? My name's Jordan Wilson, and I'm the host of Everyday AI. And this is for you. This is your daily live stream podcast and free daily newsletter helping us all not just keep up with HAT,

with what's happening in the world of AI, but how we can actually use this information to get ahead, grow our companies and our careers. So if that's what you're trying to do, welcome. This is your new home. It starts here live, unedited, unscripted with a live stream and podcast. But how you actually leverage all this and grow your company and career is by going to our website at your everyday AI.com.

So there you will find a spot that you can go sign up for our free email daily newsletter where we recap the live stream slash podcast each and every day. But there's also like 530 back episodes that you can go listen to, watch the videos, listen to the podcast, read, summaries, all on our website, all for free, all categorized by whatever you're trying to learn. So make sure you go check that out. All right. Enough chit-chat. Let's get into what matters. This in AI News. And hey, live stream audience, love to see, love to see y'all here beating me to the, to the punch. So good morning, everyone on YouTube, Big Bogey and Michael and Dr. Scott here on LinkedIn, Cecilia, everyone else. Thanks for dropping by. Love doing this with you all live. So

let's get into it right away. All right. An acquisition. Pretty big one actually just happened. this week. So Salesforce has acquired convergence AI, a London-based AI startup specializing in agents, specifically autonomous AI agents that can perform complex tasks independently by perceiving, reasoning, and adopting to changing environments. So this technology directly competed with leading AI operators in the space, including open AIs operator. So, Agenic AI represents a pretty big shift from simple AI tools that follow more scripted instructions to now intelligent systems that can manage workflows, making decisions and solve problems with minimal human input.

So Salesforce's agent force platforms leverages this approach to offer businesses scalable, unlimited workforce solutions powered by AI agents. So Salesforce CEO Mark Benoit forecasted that customers will deploy one billion AI agents by 2025, underscoring Salesforce's aggressive push into the AI agentic space. And the acquisition of convergence really strengthens Salesforce position as a key player, competing against Open AI and other leaders. in the AI agenetic space. So convergence, crazy enough, was founded just over a year ago by former Shopify and

cohered engineers bringing deep expertise in autonomous task execution and AI agent design. So yeah, we did not see an exact total on how much this acquisition was for, but it does represent Salesforce's entry into the U.S. UK seen a little bit more as well, establishing a new AI research and development hub in London with Convergence now as its core team. So Ballerton Capital, which led convergence, convergences, a 12 million dollar pre-seed funding round alongside Salesforce ventures and Shopifyed ventures, acknowledged the startup's exceptional AI talent, but noted the acquisition was somewhat unexpected. So the move signals a strategic challenge to open AI's leadership in AI agents by Salesforce, aiming to offer businesses more adaptable in autonomous AI workforce solutions across industries.

So yeah, I think that Salesforce tried to really make its mark with its new agent force offering that was announced last year and has been slowly rolling out to customers, where essentially their AI agents will make use of all of your Salesforce data and essentially follow up on leads for you. Based on the context of those conversations, the details of the opportunity, etc. But it looks like maybe Salesforce needed a little bit more help with being able to work with data outside of just its own Salesforce platform to make Agent Force a little bit better. So pretty big acquisition here. So I don't know if anyone else in the live stream audience used convergence at all. I did test it out.

For me, I know a lot of people thought it was better than Open AI's operator. I didn't think so. I still thought operator was a little more consistent. Convergence was pretty capable for sure. But it felt like much more of a beta project than did Open AI's operator. So, you know, I would say that there was, you know, over the last couple of months, there's been dozens of these, you know, agentic AI systems. But I would say convergence was probably in the top five, at least for tools that were

publicly available, you know, including Open AI's operator, Manus out of China, you know, in a few other very promising ones. So pretty telling that already with only a year since its formation, that convergence is already getting acquired. So I would assume that a lot of these other agentic AI startups, even ones that have launched within the last quarter, I would assume many of them will be acquired by the end of 2025 by larger companies that are playing hard in the space.

All right. Our next piece of AI news, Google and their AI studio has unveiled what looks like a small update, I think a pretty big one inside of its AI studio with its new generative media tab. So Google AI Studio has launched a new integrated generative media experience combining its V-O-2, Gemini 2.0 native image generation and editing in the latest Imagine 3 model, all accessible for free trial inside Google's AI studio. So Google announced the availability of these three models coming together inside AI Studio,

emphasizing that they are all ready for developers to build applications through the API. So so far, users expressed feedback on some missing history features for VO2, the AI video generator, prompting Google to commit to enabling automatic saving to prevent loss of creations. The platform currently supports turning images into videos, addressing user interests and expanding creative outputs beyond static images. Some users reported that simple prompts on Imagine 3 are frequently flagged despite being non-offensive, with Google also committing to working to reduce false positive blocks to improve the user experience.

So the integration of these models offers new opportunities for AI-driven art and media production, potentially impacting creators and businesses seeking to automate or enhance content generation. The announcement also hints at future model additions, indicating ongoing development and expansions of Google's generative AI capabilities. So yeah, this was actually a low-key announcement. I think this was just put out there by Google's Logan Kilpatrick on Twitter, but it's pretty big.

I think, you know, Google's AI Studio, it's ultimately a sandbox for developers. But with this push right here, If I'm being honest, I find Google's AI studio more useful than their, then their front end Google Gemini chatbot. One of the reasons is, is, well, I mean, you see it right here with this new create generative media experience, right? So bringing in three huge multimedia features available inside Google Gemini under one roof with even more controls than you would get inside of their front end AI chatbot,

Google Gemini. So the big ones, I mean, first and foremost, their V-O-2 AI video tool. It's by far the best and the most flexible AI video generator out there. It is better than OpenAI's SORA. It is better than Runways Gen 4. Although I do think a lot of these other companies, including a lot of the Chinese companies, are closing the gap. I would say two months ago, V-O-2 was in a wide lead, but now I would say they're, you know, more like 1A, 1B with some of the other competitors.

But, you know, to be able to bring in their industry-leading V-O-2 model inside Google's AI Studio, their Imagine 3 text to photo or AI photo generator, and then their multimodal native image and image generation in editing tool, the Gemini Image Generation, So, you know, if you haven't used that, that's very much, well, not very much, but it's fairly similar to the ultra viral, open AI, GPT40 ImageGen, right? The one that continues to go viral, nonstop, right? That's actually available inside Google's Gemini as well. And specifically, a lot of great features and functionality here now inside Google's AI Studio, which is available for free. So if you log into Google's AI Studio, look on the left hand side under the Generate Media tab, and you will see these three new, kind of not new, but new for Google's AI Studio and all under one roof. And like I said, the thing that I really like about it is it brings a clean and consistent interface to these tools where, you know, if you're using as an example, V-O-2 inside Google, front-end AI, you know, Gemini chatbot, you don't get these, you know, nice kind of drag-and-drop

options or slider options for the number of results, the aspect ratio, you know, changing the video duration, the frame rate, the resolution, negative prompt, etc. So now all of these things are more easily available. So, you know, like I said, this is kind of built for developers, but even for myself, I find myself spending much more time inside Google's AI Studio. Then I do in their front end, Google Gemini chatbot. And I think this is a pretty big push forward. 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 pro-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 variation. Every step the assistant takes is visible So you can refine, redirect, or take over at any time You stay in the driver's seat as the creative director Adobe Firefly AI assistant now in public beta See it today at firefly.adobie.com

Pretty good, pretty good comment here from Michael Saying to be honest, I hate that my news feed has become filled with AI slopped People just abuse these tools, absolutely. And I agree. But here's the thing. Whether you're following this on social media, whether you're seeing more and more updates from companies, right? Companies that maybe only used to, you know, send out a, you know, monthly newsletter, a couple posts on social. Now all of a sudden companies are posting every day or they're posting, you know,

whereas before they had a monthly newsletter. Now it's twice a week because of kind of the democratization of AI, right? And yes, what that does mean is a lot of slop. So if you are a business leader out there, you still need to focus on quality over quantity, focus on education, right? As weird as that sounds, education in knowing how these large language model and knowing how these AI tools work can really separate you from just being AI slop and being you know, just another company putting out mediocre AI powered content versus putting out

content that resonates with your current customers or potential clients. So a good call out there from Michael and just a word of warning to people. Stop falling, you know, stop just putting out AI slot. All right. Our next piece of AI news, it's essentially AI week. All right. So starting today and going throughout the rest of this week, we have some of the biggest companies with their conferences. So kicking off today, we have Microsoft build. So we're expected out of, so this is happening in Seattle kicking off, I believe in an hour or two. So by the time this is on podcast, you know, this already, a lot of this news might already be out what's being announced. But so far, what Microsoft is expected to announce on the AI side at their build conference

includes enhancements to code pilots, including semantic search in settings, file explorer, and Windows search, the introduction of co-pilot agents, AI avatars designed to assist users. Updates to Windows 11 featuring finally, right, we've seen the slow rollout to the recall feature on copilot plus PCs and we'll probably see some more and deeper integration from Microsoft with the MCP, the Model Context Protocol. Also, Microsoft may potentially be unveiling a new proprietary AI model to compete with existed offerings, right?

So, yes, OpenAI and Microsoft have been partnering, and the GPT and the O series models have been powering a big portion of Microsoft co-pilot's offerings, but we might finally see Microsoft put out its own large language model products behind its co-pilot, or at least give users the option to use its own models or others outside of just open AI models. All right. Next, a newer conference happening later this week. We have Code with Claude happening on May 22nd, which is Thursday. So we're expected to see some hands-on workshops exploring real-world implementations using the

Anthropic API and the MCP, so Anthropic. obviously made way and created the model context protocol, which all other AI labs have thankfully gotten on board with. We'll also see some interactive labs and sessions with Anthropics, executives, and product teams. And we may see a couple of smaller announcements from Anthropic, right? We've seen some rumors in the last couple of weeks about Claude, or sorry, Anthropic, maybe launching new Claude models.

I'm not sure if we'll see that. I don't think so, but a lot of people have been speculating as such. So not only do we have Microsoft and Anthropic, we obviously have Google as well with their Google I.O. Conference kicking off tomorrow in California. So we're expected to see the official official introduction of their Gemini 2.5 Pro I.O. model, which was introduced about two weeks ago and is the most powerful AI model in the world.

We're also expected to see, according to reports, a new and maybe more availability to their project Astra, which is kind of their agent that can see and interact with humans in real time that can combine their AI assistant with voice input in visual perception, as well as maybe seeing a potential unveiling of an XR headset developed in collaboration with Samsung and Qualcomm. Also more on this story later, but we'll probably be seeing a lot of new Gemini model enhancements and updates, as well as some agentic offerings. So not only do we have those three official conferences, but we're also seeing rumors that Open AI may even release even more new updates, even after they released their Codex platform.

So more on that here in a minute. Speaking of Anthropic, our next piece of AI news is well Anthropic, the AI company behind the Claude Chatbot has acknowledged a quote unquote honest citation mistake in a legal filing related to a copyright dispute with music publishers Universal Music Group, A, BKCO, and Concord. So the error involved an incorrect citation generated by Claude that included wrong titles and authors through the publication title year. Sorry, though the publication title year in link were accurate according to Anthropics defense attorney. So the citation mistake was discovered after the filing was submitted in late April by Anthropic as part of the company's defense,

claims that copyrighted lyrics were used to train clawed. So Universal Music Group's attorney called the source a quote unquote complete fabrication, implying it was hallucinated by the AI, but Anthropic insists the source itself was genuine and the error was just an internal fabrication. Anthropic has apologized for the confusion, describing the incident as a quote unquote embarrassing and unintentional mistake, highlighting the risk of relinquency. lying on AI tools for legal citations without thorough manual checks. The case adds to a growing trend of AI generated errors in legal settings,

including a recent California court's reprimand of law firms for submitting AI-created briefs containing bogus materials that did not exist. When are we going to get over this? And to see this from one of the AI labs is absolutely. bonkers. This is, I mean, this is straight face palm. Okay. Anytime. One of the big AI labs, right, whether it's Anthropic, Microsoft, Open AI, Google, etc. They're going to find themselves in the courtroom very often over the next few years and probably beyond that as the law of copyright gets challenged, right? So please, this is a personal plea. Legal teams don't rely on your own AI

tools alone when submitting legal briefs and exhibits to the courtroom. I can't believe this has to be said. But we talk about human in the loop, which I think is actually sometimes a bad thing, right? I like to say we need expertise in the loop. The fact that this was not double, triple checked by a company as big as anthropic with pockets as deep as anthropic, I think is indicative of a larger issue of just being lazy humans, right, and slapping the big red easy button on AI. Yes, AI can do research. Yes, it can pull in factual information.

But literacy is clearly low. If you literally have Anthropics Legal Counsel making elementary mistakes, when the stakes are so hot, literacy, we need it. AI literacy, when you are even seeing the frontier labs make elementary mistakes. It is a sad day for just humanity that we are now getting so lazy that we can't do the basic work of checking citations, right? It's not good. What is good is our next piece of AI news. This is extremely exciting. It's a little dorky.

I'm going to try to break it down. But Google's deep mine has discovered and created and released details on alpha. Alpha evolve, an AI tool that can automatically find better ways to write programs and solve complex problems. So using AlphaEvolve, Google improved how jobs are scheduled in its data centers, freeing up 0.7% of the world's computing power. Pretty big deal given how much energy and money data centers use. So instead of just understanding human language, Alpha Evolve works by testing and improving solutions step by step, like how living things evolve over time.

So one impressive success was Alpha Evolves finding a new faster method to multiply certain math matrix matrices, beating a solution that has stood for more than 50 years. Yes. So this is a new AI. a tool or a new AI model from Google DeepMine that has solved a decades-old math problem and has made it much more optimized. So Alpha Evolve is essentially creating new algorithms on the fly that are solving real-world problems, at least for Google right now. And you might say, oh, okay, freeing up 0.7% of the world's computing power,

big deal. For one advancement to potentially save just one percent of compute, it is fascinating. And that is huge news because this is the first iteration of Alpha Evolve. I would assume that as Google's researchers and the DeepMind team continue to work on AlphaEvolve, we will probably in the coming years see that, you know, you know, 0.7% multiply many times over. Okay. So the AI was able to rediscover known math solutions 75% of the time and even found improvements in 20% of those cases showing that it can both learn and innovate.

So yes, for all those people that say, oh, AI is just a, you know, very advanced auto complete. No. It's literally developing new algorithms on the. fly that are solving decades-old math problems way faster and with more accuracy than humans can do it. So experts see great promise in Alpha Evolve, but stress it must be used to carefully avoid mistakes or security problems from automated coding. So for businesses and professionals, this means new tools could soon make computing faster

and cheaper, helping companies grow and innovate more. quickly. Sorosch here on YouTube says can't wait for all the Google updates tomorrow. Not sure what else they have up their sleeves. Yeah. Let's just say you should tune it. All right. Yeah. We'll be covering all week, both from Microsoft, Google, Claude, and the rumored Open. You know, if Open AI does also release something. We could see new updates from the four biggest players in the game this week. So yeah, we'll be covering it all this week on the everyday AI show. All right. Speaking of Google and what they may be showcasing, well, according to reports, Google is demonstrating a new AI agent designed to assist software engineers throughout the entire development process,

from task management to code documentation, according to a report by the information. So this reported new AI agent aims to streamline software creation, potentially making coding more efficient and accessible for developers by handling routine tasks and improving workflow. Google may also reveal how its Gemini AI chatbot integration with its Android XR glasses and headset highlighting advances in voice-controlled AI interactions and wearable technologies. So these developments come as Google Facebook. This is growing pressure from investors to prove returns on a significant investment in artificial intelligence amidst riding competition in the AI sector. So Google's annual conference, like I said, is scheduled for tomorrow in California.

So like I said, we'll be covering those, both the rumored new AI software agents and improvements to Gemini. We're probably going to be seeing a lot of multimodal offerings out of Google I.O. As well. So very exciting week. It's like Christmas in May. All right. Our next piece of AI news, the United Emirates and U.S. tech giants are uniting to build a massive AI data center, now referred to as the UAE Stargate in Abu Tabi.

So the UAE Stargates project will be a huge AI data center covering 26 square kilometers with an energy capacity of 5 gigawatts, highlighting the UAE's major ambitions in the AI space. So the data center, which was just announced last week with a convoy of U.S. political and AI leaders, visiting the region, but the data center will integrate cutting edge American tech. technology, including Nvidia's Blackwell GB300 chips, and the first phase alone will feature a one gigawatts computing cluster, signaling a massive leap in AI computing power. So top U.S. leaders, such as Jensen Wong from Nvidia, Sam Altman from OpenAI, and G2 Patel from Cisco personally attended the signing in the UAE, showing strong U.S. corporate commitment to this international partnership. So the, uh, emirate, the UAE company G 42 will lead the construction and integration efforts,

reinforcing the UAE's strategy to become a global tech hub and diversify its economy beyond oil. So the UAE has committed to investing 1.4 trillion over 10 years into the US economy, focusing on AI and semiconductors made here in the U.S. underscoring a new level of economic collaboration between the new nations. So this project complements OpenAI's own U.S.-based Stargate campuses planned across 16 states, starting with Texas and aiming for completion by mid-20206, indicating a coordinated global expansion of AI infrastructure. So beyond the,

the Gulf, the UAE is investing $50 billion in France for an AI campus and is active in crypto innovation recently launching a central bank regulated stable coin showing a broader push into emerging tech sectors. So pretty huge news here, right? The leaders of the biggest companies in the U.S. heading over to the UAE. as well as political leaders as well, really signaling a joint collaboration and international expansion on what has already powered the U.S. artificial intelligence seen here in the U.S., namely, right, the Stargate Project collaboration between NVIDIA and OpenAI.

So pretty big news, both for those companies, as well as the U.S.'s larger rollout across the world. All right. Our next piece of AI news, two last ones for you, both Open AI. So Open AI has rolled out its GPT 4.1 and GPT 4.1 mini models to chat GPT users. So Open AI has launched and rolled out its GPT 41 and GPT 41 mini within chat GPT, aiming to improve coding assistance and instruction following while offering faster response times compared to previous GPT40 models.

So previously, this new GPT41 was only available in the API. Yet last week, OpenAI said due to essentially popular demand that they were rolling out GPT 4.1 to paid users. So right now, Chad GPT Plus, Chad GPT Pro, and chat GPT team subscribers will have access to GPT 4.1, whereas GPT 4.1 Mini is accessible to both free and paying users as well. So the OpenAI clarified that GPT 4.1 is not a frontier model, and therefore does not require the same level of safety reporting as did previous more advanced models. So to address transparency issues, OpenAI also announced the launch of a new safety

evaluations hub, where it will publish internal safety evaluation results more regularly, signaling a commitment to clearer communication about AI risks. So this new GPT 4.1 release, you know, one of the things that it's really improved on is coding in prompt adherence as well as speed, but it comes about the same time that Open AI has repeat. importantly closed on a $3 billion acquisition of Winsurf, a leading AI coding assistant. So that is pretty timely because our last piece of AI news is a coding assistant that OpenAI rolled out. So Open AI has announced and rolled out its codex platform to Chad GPT Pro, Enterprise and team users with a focus on safer and more transparent coding.

So this move highlights OpenAI's ongoing efforts to make AI generated code more reliable while emphasizing the need for human oversight. So Codex is designed to tackle more frequent issues flagged by developers with earlier AI coding tools such as producing code that doesn't meet standards might be hard to debug or has security vulnerabilities. So the fine tuning behind its new model. So this new model is called Codex 1, and it is a fine tune of OpenAI's most powerful O3 model.

So essentially, this is a new model and mode inside ChatGPT. So if you are on a ChatGPT Pro enterprise or team account for now, if you look on the left-hand side, you will see the new Codex kind of interface. kind of interface. Also, it did get rolled out to mobile over the weekend as well. So the fine tuning behind Codex 1 helps the system show its reasoning and intermediate steps during coding tasks, which can't take anywhere from 1 to 30 minutes. So essentially all these different coding tasks get their own sandbox and they can go work on all of these tasks kind of while you wait. So OpenAI stresses that despite these improvements, users must

still manually review and validates all AI generated code before integrating or running it to avoid potential errors or security risks. So the rollout currently, like I said, includes chat GPT Pro, Enterprise and team users with plus and education users expected to gain access in the coming weeks. So for the next few weeks, OpenAI is also offering a generous no extra cost access to code so users via the API, so users can explore its capabilities, but plans to implement rate limits and a new pricing model later. So this release reflects ongoing challenges in balancing AI automation

with the need for human control, especially for developers relying on AI to write or assist with complex coding projects. So a lot of people myself included, you know, so Open AI hinted at something that they were going to release on Friday. So a lot of people thought this might have been an official windsurf product, but it was not. So Open AI had already been working on a coding specific model. Okay. So this is more than a mode within ChadGBT. This is a fine tuned new large language model that Chad GPT has released. So now as an example, right, you have GPT4O, which is kind of the workhorse

model that's used for everything. including their very popular image gen. Then you have some of these O series models, right, such as O3, which is the most powerful model. Then you have O4 Mini-4 Mini-Hi. So these are models that kind of think, reason, plan ahead. It can access the Internet, can access your code base as well. You have the deep research kind of mode, which is a fine-tuned version of the O3 model.

So this is a brand new model. Codex 1. It is based on a. fine tune of their most powerful model 03, and it is specifically for coding. So what I'll say, and this is with my very little experience, because this thing's only been out for like not even 72 hours, it's been out for like 60 hours. So I haven't had a ton of time to use it just yet, but I will in the coming weeks. So this is not one of those vibe coding assistants, okay?

you know, where you kind of with natural language, you know, create an app and launch it and iterate on it. This is really, I would say, more geared towards software engineers. It's more geared toward enterprise operations, I would say, than these more consumer tools like windsurf, like cursor, right? Like bolt, like lovable. I think some of these, you know, more vibe coding tools that are more for consumers, entrepreneurs, entrepreneurs, people just looking to kind of code from scratch

where I think it looks like Codex 1 in OpenAI's codex is more geared at least right now toward more sophisticated and complex coding projects. So we might not be covering it as much but this could also be laying the groundwork

for windsurf, right? So maybe this will be their first version of codex and maybe codex will be updated essentially with a more consumer or vibe coding model via windsurf or maybe, you know, Open AI will just keep windsurf as is. So a lot of people were expecting this because it was teased and previewed to be for developers. People were expecting a windsurf integration, but we didn't get it. We got something a little bit more for software engineers, for more hardcore developers,

and less for kind of consumer vibe coding. which is maybe what a bigger chunk of our audience might be focused on. However, like I talked about, there are some reports and rumors that open AI, even though they did just release this on Friday afternoon. There are some reports that they may be getting in on the kind of AI week this week with either some new features that they'll be announcing or even potentially, right? We've heard rumblings of an O3 Pro. Maybe we will get something officially on windsurf or maybe we'll get nothing at all as Anthropic, Microsoft, and Google dominate the week with their new AI announcements.

All right. That is a wrap. Let's very quickly run down the AI news that matters for the week of May 19th. So first, Salesforce has acquired AI agent startup convergence, which I think was a top five agent. tool and a direct competitor to open AI's operator. So pretty big move there from Salesforce as they look to expand on their agent force capabilities. Next, Google AI Studio quietly unveiled a unified generative media platform inside Google's AI studio that brought together its V-O-2, Gemini 2.0 image generation in Imagine 3 models.

Next, it is literally AI conference week with Microsoft Build, Claude, Code, and Google AI, or sorry, Google's I-O that will be bringing a lot of AI all happening this week, starting with Microsoft Build today. Next, Anthropic admitted that it used its own AI chatbot and made a citation error in a legal filing. Huge face palm. Not a face palm. The next piece of AI news, Google. DeepMine unveiled Alpha Evolve an extremely groundbreaking advancement in artificial intelligence.

Google also reportedly is going to showcase some new AI agentic software development tools at their IO conference this week. The UAE and US Tech Giants united to build a massive UAE Stargate in Abu Dhabi, a pretty significant joint venture between some big U.S. companies and UAE. Next, OpenAI rolled out GBT 4.1 into chat GBT, which was previously only available via the API. And then last but not least, we saw OpenAI announce and roll out its new Codex, AI coding platform for pro enterprise and team users. All right, that was a lot to cover. I hope this was helpful.

And yeah, there's a lot more coming in the coming hours and the rest of the week. So you better believe we'll be keeping you up to date here on everyday AI. Huge week for generative AI, large language models, AI advancements. We're going to be keeping you in the loop. Don't worry. So for now, if you haven't already, please share this if this was helpful, right? If this, you know, if you like tuning in on Monday to get the AI news and so you don't have to spend hours every single day scratching your head, trying to separate, you know, separate the meaningful from the BS. Please tell someone about it.

Share this if you're listening on Twitter or LinkedIn. If you are listening on the podcast, specifically Spotify, thank you. We appreciate if you could subscribe to the show as well as. leave us a review. That would be huge. We'd really appreciate it. So then go to Your EverydayaI.com. If you haven't already, sign up for the free daily newsletter. We're to be recapping today's show as well as a whole lot more as a very exciting and fast-paced week in AI kicks off. Thank you for tuning in. Hope to see you 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.adop.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.