Resources:
Join the discussion: 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
AI News Deep Dive: Google's Gemini, OpenAI's "Code Red," Perplexity Lawsuit, and Enterprise Adoption
Amid year-end headlines in the AI sector, several pivotal developments demand close attention from business leaders navigating technology strategy and enterprise growth. This analysis distills recent news covered in the Everyday AI podcast episode, translating essential insights into practical, actionable takeaways.
Google Gemini Deep Think: Advanced Reasoning for Professionals
Google's release of Gemini 3 Deep Think Mode sets a new benchmark for AI reasoning capabilities, specifically engineered for professionals and researchers. Available to Ultra-tier subscribers, Deep Think leverages parallel reasoning to solve multifaceted math, science, and logic problems, delivering scores of 93.8 on GPT QA Diamond and 45.1 on Arc AGI-2 with code execution.
For businesses seeking the strongest available large language model (LLM), Deep Think rivals any competitor—surpassing previous Gemini iterations and even OpenAI’s top pro model on complex benchmarks. Although high pricing ($250/month for Ultra users) and initial stability concerns limit broad access, this model is positioned as a premium tool for organizations requiring nuanced code, advanced data visualizations, and robust prototyping.
Perplexity Lawsuit: Copyright Risk for AI Content
The New York Times and Chicago Tribune have initiated lawsuits against Perplexity, signaling escalating legal risks for AI firms leveraging news content without authorization. Both publishers allege that Perplexity trained its models on licensed articles and reproduced original content verbatim—potentially fabricating information and crossing paywalls.
These legal cases join the ongoing New York Times vs. OpenAI and Microsoft suit, which remains a major precedent-setter for future copyright policy in AI development. Should settlements or court decisions favor publishers, a cascade effect may force stricter licensing agreements, significant financial settlements, and operational shifts for all AI businesses built on third-party data.
Amazon’s Nova 2 and NovaForge: Enterprise Custom AI Models
At AWS re:Invent, Amazon debuted four Nova 2 AI models (Lite, Pro, Sonic, Omni) designed for specialized enterprise needs, from everyday reasoning to speech-to-speech conversational AI and multimodal processing. Unique among major providers, AWS launched NovaForge, a cloud-based service allowing businesses to train custom Nova-based models ("Novellas") with proprietary data (starting at $100,000/year).
Though the Nova models are not projected to lead the market in raw AI capability, NovaForge introduces a flexible framework for enterprises to tailor models without full open-source exposure. This approach is noteworthy for organizations deeply integrated into AWS and Bedrock, seeking to differentiate with in-house data and workflows.
Accenture Deploys ChatGPT Enterprise: Consulting at Scale
Accenture's formal deployment of ChatGPT Enterprise to tens of thousands of consultants marks one of the largest organizational adoptions of generative AI tools. With a commitment to outcomes-based consulting billing—shifting away from legacy hourly models—clients will benefit from greater transparency in AI-driven productivity and clear business results.
This move reflects a broad trend: major consulting and financial firms that previously resisted off-the-shelf LLMs are now adopting proven models from OpenAI, Google, and Claude. Enterprise certification programs and cross-functional AI integration are becoming standardized, impacting billing structures, competitive positioning, and client expectations.
Anthropic Eyes IPO: Financial Implications for AI Startups
Anthropic's preparation for an IPO, potentially valuing the company above $300B, signals increasing maturity—and profitability—among AI startups. With revenue projected to reach $9B in 2025 and strong enterprise sales, Anthropic appears primed to break even by 2028, outpacing OpenAI’s longer break-even horizon.
For investors and strategic planners, Anthropic’s trajectory suggests that tighter cost controls and enterprise focus may drive earlier public offerings and more stable long-term financials. The race to IPO between Anthropic and OpenAI may reshape capital flows and valuation standards in the sector.
Workspace Studio & Replit: No-Code AI Agents and Vibecoding Expansion
Google’s Workspace Studio now empowers employees to build, manage, and share AI agents across Gmail, Drive, Chat, and other apps using no-code or low-code interfaces. With 20M automations executed in 30 days among early adopters, businesses are increasingly able to delegate and automate routine workflows without software engineering expertise.
Concurrent partnerships, such as Google Cloud with Replit, extend advanced Gemini AI coding capabilities to enterprise developers. The industry-wide “vibecoding” trend—tools like Anti Gravity, Opal, Jules, and Gemini CLI—enables nontechnical staff to assemble bespoke business software, lowering development barriers and accelerating operational agility.
OpenAI "Code Red": Accelerated Model Releases in Competitive Response
Amid Google's breakthrough with Gemini 3, OpenAI leadership reportedly declared an internal "code red," fast-tracking the possible release of GPT-5.2. This pivot deprioritizes consumer-facing features in favor of raw model improvements in speed, reasoning, and reliability. Fast-tracked model releases reflect intensifying competition, as Google surpassed OpenAI’s benchmarks in several technical domains.
Businesses relying on OpenAI technologies should monitor this product cycle closely, as near-term shifts may enhance performance but delay or deprioritize supplementary features such as GPTs, Atlas, and project automation.
Summary: Strategic AI Outlook for Enterprises
The evolving AI landscape demands adaptive strategies focused on model performance, legal compliance, enterprise customization, and operational integration. Businesses should:
Consider premium models for advanced research, prototyping, and automation needs (e.g., Gemini Deep Think);
Scrutinize copyright risks in AI content and prepare for tighter regulation or licensing regimes;
Evaluate custom AI options from leading cloud providers for internal data-driven innovation;
Align consulting and service engagements to transparent, outcome-based AI productivity;
Monitor financial signals from leading AI startups for future investment and competitive planning;
Enable cross-functional teams to rapidly automate and innovate with no-code AI development tools;
Stay agile as providers accelerate product release cycles amid escalating competition.
Responsive, informed adoption will define operational advantages as these AI sector developments continue to unfold.
Topics Covered in This Episode:
- Google Gemini 3 DeepThink Model Launch
- Perplexity Lawsuit by New York Times, Tribune
- Amazon AWS Nova 2 AI Model Family
- Accenture Deploys ChatGPT Enterprise Partnership
- Anthropic Prepares for $300B IPO
- Google Workspace Studio No-Code AI Agents
- Google and Replit AI Coding Partnership
- OpenAI Code Red: GPT-5.2 Model Release Rumors
Episode Transcript
Midroll [00:00:00]:
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.
Jordan Wilson [00:00:17]:
...and Google very quietly released a new version of Gemini three that's actually more powerful than the version of Gemini three pro they just released, like, two weeks ago. Yet there's just two words that people are focusing on. Not those stories. It's code red. That is because according to reports, OpenAI is kind of shaking in their AI generated boots as they've kind of slowly lost their stranglehold on the AI world mainly due to Google's just extremely impressive growth this year.
Jordan Wilson [00:01:12]:
What's going on y'all? Welcome. My name is Jordan Moulson, and this thing's for you. Everyday AI is a daily livestream podcast and free daily newsletter helping everyday business leaders like you and me not just keep up with the AI news, but how we can make sense of it. Cut the BS, the marketing, and just really extract the good insights that we need to grow our companies and our careers. So if that's what you're trying to do, great. Starts here with the unedited, unscripted livestream podcast. Make sure you go check out the free newsletter though on your everydayai.com. Each day, we recap that day's highlights from the podcast.
Jordan Wilson [00:01:44]:
So if you ever miss an episode, don't worry. And, hey, just so you know, on Mondays, we go for the AI news that matters. Most Tuesdays do kind of like a hot take Tuesdays. On Wednesdays, we do AI at work. On Wednesdays, more of a practical hands on look at AI. And usually on Thursdays and Fridays, we do some interviews. Alright. So enough chitchat.
Jordan Wilson [00:02:02]:
...Google rolled out a more powerful version of Gemini three as if that was possible. So Google has launched its new deep think mode for Gemini three, now available to AI Ultra subscribers. So deep think is Google's most advanced reasoning model designed to tackle complex math, science, and logic problems using what they call, parallel reasoning to explore multiple hypotheses at once. So the feature builds on Gemini 2.5, which also had a deep think version with marked improvements and key benchmarks, including a 93.8 score on GPT QA Diamond and a 45.1 score on Arc AGI two with code execution.
Jordan Wilson [00:02:56]:
...Unfortunately, at least for the first couple of days, it wasn't super usable. Right? There was a lot of, kind of downages or outages, whatever you wanna call it.
Jordan Wilson [00:03:57]:
...new Deepthink mode in Gemini three, absolutely bonkers. Hopefully, this will roll out to all paid users. You know, it's it's it's kind of a bummer that more people can't, you know, use this mode because it is really good. So I'll say this. If you've ever used ChatGPT 5 pro or ChatGPT 5 one pro, it's very similar.
Jordan Wilson [00:04:45]:
Right? It's a slow model. Right? And sometimes, unfortunately, it's gonna kick out errors. But when it does work, right, and if you do have access, this is the most powerful, AI large language model that there is available, bar none. Right? So, I said this on last week's show and the week before this that I still thought Gemini five or or sorry, OpenAI's ChatGPT 5.1 pro was the best model even after, Gemini three pro was released. But now I need a little bit more time. Google might have, turned turned me on this one.
Jordan Wilson [00:05:31]:
...that's because the New York Times and the Chicago Tribune have filed lawsuits against Perplexity, marking a major escalation in the battle over news content and AI. So The New York Times alleges that Perplexity used its articles without permission to train AI models and power their chatbots in a very similar, you know, lawsuit that they had with OpenAI. And Perplexity's AI products reportedly, according to the Times, reproduced Times articles verbatim and have falsely attributed fabricated information to the newspaper.
Jordan Wilson [00:07:00]:
...I don't think that case will go to trial, but whatever does happen between the New York Times and OpenAI, I think will set the precedent for the future of AI and copyright. Right? If for some reason there's a settlement in OpenAI or Microsoft, has to pay, that's gonna trickle down to literally every single, you know, AI company.
Jordan Wilson [00:07:50]:
And, conversely, if OpenAI just beats the lawsuit with no slap on the wrist, yeah, a lot of these lawsuits are gonna fall. And just the kind of the whole concept of copyright, at least here in The US, is gonna come into, question. Alright. Next piece of AI news, Amazon making headlines with the launch of four advanced AI models and a new service that lets company build their own custom versions as the company announced at their AWS re:Invent conference. So AWS introduced the Nova two family featuring four new AI models designed to handle complex tasks across text, images, video, and speech. So the Nova two Lite model is focused on cost efficient reasoning for everyday tasks, while Nova two Pro tackles highly complex jobs like coding and advanced analytics. Then you have Nova two Sonic, which brings speech to speech conversational AI, and then Nova two Omni, which is their multimodal reasoning model able to process and generate across text, images, video, and speech. So aside from the four new models that Amazon announced, right, nothing new there.
Jordan Wilson [00:09:04]:
Everyone kind of assumed that the second version of these would come out a year later, but AWS also rolled out NovaForge, which is essentially an AI inferencing, cloud service, but it's a service that allows enterprise customers to create customized versions of Nova models called Novellas, for a $100,000 a year. So NovaForge gives companies access to pre trained, mid trained, or post trained models, enabling them to further train these models using their own proprietary data. So interesting, approach here, from Amazon, almost like an in between, not truly open source, but I haven't seen this a lot. Right? Companies giving, two Frontier models, and we'll see where the new, Nova two models land if they're, you know, really on the cutting edge or the Frontier. I'd say probably not. I don't really expect them, ultimately to be top 10 models, but even if they are, you don't really see that a lot. Companies, you know, having the option to use either pre trained, mid trained, or post trained versions of models. So, pretty novel approach, from Amazon with this to see if they can compete in the, on the AI enterprise side.
Jordan Wilson [00:10:24]:
Because aside from the cloud, take out the cloud. Right? Take out AWS. I don't personally know any company that uses the Nova models. Again, unless they are, really just handcuffed, into AWS for any reasons or if they have their entire, company, you know, running in bedrock or something like that. But I don't know any company that just chooses to use, Amazon's Nova model. So maybe this could change the conversation. I guess we'll see in 2026. Well, guess who's gonna be using at least one model? Accenture.
Jordan Wilson [00:10:56]:
Thousands of them are gonna be using ChatGPT. That's because consulting giant Accenture is teaming up with OpenAI to roll out ChatGPT enterprise to tens of thousands of its consultants, marking one of the largest enterprise adoptions of GneAI to date. So the partnership is expected to accelerate how Accenture develops consulting, operations, and tech services. And according to CEO Julie Sweet, this move will help, quote unquote, accelerate enterprise reinvention in business outcomes and quote for clients. So pretty big shift here, right, that Accenture is making in this official partnership. And OpenAI did put out some, some, words on this as well, and they did confirm that Accenture will have the largest number of professionals certified in OpenAI technologies. So if you turn back the clock, right, like two and a half years. So, you know, we are now into year four of ChatGPT, ChatGPT is more than three years old, but if you go back two years, right.
Jordan Wilson [00:12:05]:
I said this at the time. I'm like, all these, you know, consulting companies were writing off large language models, which I found to be funny. I'm like, there's no choice. Right? I said, all of these large companies that early on, we're writing them off while we're not using them. Oh, no. We're just gonna use our own. I'm like, no. You're not.
Jordan Wilson [00:12:24]:
Right? You're either gonna go out of business. Eventually, you're gonna get leapfrogged by smaller competitors or you're going to end up using models from Google, Claude or OpenAI. You don't have a choice. Right? So it's foolish looking back at it now. You know, some of the big, financial institutions, consulting firms, you know, back in late twenty twenty two, early twenty twenty three that said, yeah. We're not touching this chat GBT thing. Never. Never.
Jordan Wilson [00:12:50]:
It's like, no. Yeah. You are. There's no getting around it. That's like saying you aren't gonna use the Internet. They become one now. So, regardless, it should be interesting to see how the consulting space is shaping up. I do think that there is, well, gonna be a lot of things that I talked about this on my year end predictions, last year.
Jordan Wilson [00:13:10]:
But I do think that you're gonna see kind of hourly based, consulting eventually got the window. And we're gonna look back at it and be like, I can't believe consulting companies charge hourly. Right? Mainly, it's getting moved to outcomes based. It has to. There's already been a couple of, stories, showing a pretty big percentage of consulting projects that have for decades just been billed hourly are now being billed by results because companies are getting smart. They're like, hey. We're not gonna pay Accenture or or KPMG or anyone else, hourly anymore because we don't necessarily know how they're using AI. So we're gonna pay for the outcome or we're gonna use AI ourselves and not pay these big consulting companies for this.
Jordan Wilson [00:13:51]:
Alright. Anthropic could be going public. Yeah. So, Anthropic is reportedly making early moves toward a potential public offering that could just reshape the complete tech landscape. So according to reports from the Financial Times and others, Anthropic is reportedly seeking a new funding round that could value the company at over $300,000,000,000, and the company has hired a law firm this month to begin laying the legal groundwork for an IPO, a major step for any tech firm considering going public. So despite these preparations, Anthropic has still essentially said that there's no immediate plans for an IPO, but that's usually what a company that's, laying the groundwork for an IPO says about their upcoming IPO. They says, hey. There's no current plans, but, well, it's gonna be rolling out, I would assume.
Jordan Wilson [00:14:46]:
So industry speculation suggests that Anthropic could list shares as early as the 2026, potentially racing against rival OpenAI for a first spot on the public markets. So Anthropic has reportedly held informal talks with major investment banks about a possible IPO signaling serious interest from Wall Street. So the company's valuation has soared from 183,000,000,000 back in September to now more than 300,000,000,000, driven by major recent investments from Microsoft and NVIDIA last month. So Anthropic's annualized revenue runway is projected to reach $9,000,000,000 by the 2025 with enterprise customers making up about 80% of sales. Also, reportedly, their new Cloud Code product alone is approaching $1,000,000,000 in annualized revenue. So, it is gonna be interesting to see, if Anthropic or, OpenAI is the first to go public. It seems like it might be Anthropic. We've seen reports that said OpenAI might be late twenty twenty six as soon as early twenty twenty seven.
Jordan Wilson [00:15:59]:
Now we're seeing some of this code red talk from OpenAI, which we're gonna talk about here in a couple of minutes. And I will say this, and I've said this before. I think OpenAI has a much higher ceiling than Anthropic. However, I would assume that Anthropic is much more profitable per user right now than OpenAI, which probably makes them a little more primed for an earlier IPO. Right? Because when you look at their finances, and their CEO did say that they're on track to break even by 2028. Whereas in, you know, OpenAI, a lot of the estimates have been 2029 or 2030. So, it seems like OpenAI may be burning billions of dollars for more years longer than Anthropic, and Anthropic could be profitable at a much sooner rate. So that might make a little bit of a more attractive offering for going public sooner, but we'll see.
Jordan Wilson [00:16:56]:
Alright. Next piece of AI news, one I'm excited about and didn't really see this one coming. So Google has officially launched Workspace Studio, a new platform that lets employees build, manage, and share AI agents to automate routine tasks. So, yeah, this is pretty cool. It's a no code, low code automation builder. So if you've ever used something like Zapier, like Make, you know, or even their other product, which is kind of similar, Google Flows, this is a more robust version for workspace customers. So the system uses Gemini's multimodal reasoning, allowing users with no coding experience to create custom automation and agents in minutes. So Workspace Studio integrates directly into Gmail, Drive, Chat, and other workspace apps.
Jordan Wilson [00:17:46]:
And according to Google, early adopters have already used Workspace Studio agents for more than 20,000,000 tasks in the past thirty days alone ranging from reminders to handling legal notices in travel requests. So here's how it works. Users design automations just by describing their goals in natural language. So shifting away from more rules based systems and lowering the barrier for nontechnical employees to automate workflows. So agents can be shared across teams, supporting, collaborative work in faster project cycles. Workspace also can connect to popular third party tools such as Asana, Jira, Salesforce, Mailchimp, and others, and also allows more technical users to add custom logic via script, apps script, from Google or Vertex AI. Google does plan to roll out Workspace Studio to business customers in the coming weeks and is inviting organizations to join its Gemini alpha program for early access. I did get access, to this already.
Jordan Wilson [00:18:51]:
I've been playing around with it a little bit. Not enough yet that I am ready to jump in and do a full episode. If I'm being honest, y'all, the last, I don't know, month or so, it's been hard even as someone that does this every single day. Right? I still wanna go I still wanna go use, Google's workspace, flows. Right? A little different than workspace, Studio. And I barely had time to do that even though it's baked in now, to the sidebar of my Gmail. Right? So, just it's been, the last, like, five or six weeks. It's been just so hard, even as someone that does this daily to keep up, but this is one I'm excited for, inside works, workplace studio, launch.
Jordan Wilson [00:19:36]:
So, hey, live stream audience. Let me know. Have you played with this yet? What are you looking forward to? Should we do a show on it? You let me know. I work for you. Alright. Speaking of Google news, big partnership with Replit. So Replit and Google Cloud have announced a major expansion of their strategic partnership, aiming to bring more advanced AI powered coding tools to enterprise developers. Well, in shorter words, Google and Replit shook hands on bringing five coding to more people.
Jordan Wilson [00:20:09]:
So under this new multiyear agreement, Replit will rely even more on Google Cloud's infrastructure and services, including Google Cloud Run, BigQuery, and others to support future scaling. So Replit will integrate Google's latest AI models, including Gemini 3, Gemini 2.5 Flash, Gemini 2.5 Flashlight, and Imagine 4 powering new coding in multimodal features for Replit developers. So Google 3 was recently added to Replit's design mode, offering advanced capabilities for cogeneration and design workflows. As part of this agreement, Google Cloud will remain Replit's primary cloud provider, and both companies will pursue joint go to market efforts through the Google Cloud marketplace and co selling programs. So yeah. Another one that was kinda like, Didn't see this one coming. You know, really, getting the enterprise kind of handshake on vibe coding. And if you haven't noticed, Google has gone all in and then some on vibe coding.
Jordan Wilson [00:21:17]:
I mean, we've seen multiple new kind of vibe coding tools, from Google including anti gravity, Opal, Jules, Gemini CLI. Right? So there's four, kind of vibe coding tools that didn't exist last year. Then you have, the build feature in AI Studio, and now you have this, official partnership with Replit, one of the other leaders in the vibe coding era. So, Google is not shying away, from vibe coding. Right? If you thought that this was a a trend, if you thought it was gonna, you know, be here today, gone tomorrow, absolutely not. It is here to stay. And, again, hate to keep going back to my old, you know, 20, 24 predictions that I made about 2025. I said everyday people are gonna be building their own software, and here you have it.
Jordan Wilson [00:22:07]:
I mean, we've seen multiple new kind of vibe coding tools, from Google including anti gravity, Opal, Jules, Gemini CLI. Right? So there's four, kind of vibe coding tools that didn't exist last year. Then you have the build feature in AI Studio, and now you have this official partnership with Replit, one of the other leaders in the vibe coding era. So, Google is not shying away from vibe coding. Right? If you thought that this was a trend, if you thought it was gonna be here today, gone tomorrow, absolutely not. It is here to stay. And, again, hate to keep going back to my old, you know, 2024 predictions that I made about 2025. I said everyday people are gonna be building their own software, and here you have it.
Jordan Wilson [00:23:18]:
So the alert reportedly called for an immediate shift in priorities across OpenAI with teams directed to focus more on improving the speed, reliability, and reasoning power of ChatGPT rather than focusing on developing new consumer features. So staff were reportedly instructed to accelerate the development and deployment of the next major model update reportedly called GPT-5-2 with a target release date, which now could, according to The Verge, be tomorrow. So, yeah, first, we heard from the information and others that it might be early January, and then it was like, oh, mid December, late December, and, nope, now it's just could be tomorrow. So the irony for those that are keeping score at home is it was, well, the tables have been flipped. Right? OpenAI here calling a code red on Google because, hey, two years ago, OpenAI had no real competitors. Now Google has been running away, with the at least on the model side, in 2025, but it was actually flipped. Because if you go back to December 2022, reportedly, it was Google that had its own emergency code red meeting in December 2022 after OpenAI released ChatGPT. Because most people don't know this, but Google actually had AI products that they were testing in the market before ChatGPT landed, but their approach was just too conservative.
Jordan Wilson [00:24:47]:
And I'd say that at least in 2025, Google has been winning the model race, just on the model race, while OpenAI has been winning kind of the modes and features and growing the consumer base. But, apparently, Sam Altman wants to be winning the model race and is apparently worried enough that he's calling a code red on this one. So the reports are saying that OpenAI is now gonna be shifting back a little bit more of its focus toward models. What I really hope is I hope that this doesn't mean, you know, a lot of these other, you know, features that were maybe more consumer based get abandoned. Right? I hope not. Like, GPTs and projects and Atlas and, agent mode at some of these fantastic, just literally, like, mind bogglingly good features. I hope they don't all, get kind of left out to die in lieu of newer models. We'll see what OpenAI does.
Jordan Wilson [00:25:39]:
But a little bit more on this GPT-5.2 update that could come as early as tomorrow is designed to close the performance gap with Gemini 3, focusing on, like I said, improvements in speed, reasoning, and reliability rather than just new features. So this was originally planned for later in December, but the launch was kind of moved up because of this pressure from Google and a little bit from Anthropic as well. Internal sources say that OpenAI's own evals show that GPT-5.2 could match or exceed Gemini 3's capabilities. So, regardless, it's not even ship miss yet. Right? So if you're wondering, like, what the heck is Shipmas? Well, last year, OpenAI and Google went literally toe for toe. Right? OpenAI did more of, like, a big announcement, some splashy marketing. Right? They did this twelve days of Shipmas. So right before the Christmas holiday here in the U.S. OpenAI had 12 different releases.
Jordan Wilson [00:26:44]:
It was a pretty big deal and they were good, really good. Google didn't come out with any marketing and they, in the midst of it came out with some huge releases. And a lot of people said Google probably won last December. So a lot of people were saying, like, okay. Is this thing still gonna happen? There were some chatter online, at least from OpenAI that said, yeah. We're still gonna be bringing a lot of new updates in December. Obviously, we got Google's maybe big updates in November. Is there gonna be more? I don't know.
Jordan Wilson [00:27:15]:
But, apparently, there may be a new OpenAI model, their best one as soon as tomorrow. Alright. I know that was a lot, but there's technically a lot more. So at the end of our AI news recaps, kind of, episodes, we do a what's new and what's next. So these are some rumors, some small pieces of news, some big pieces of news that maybe just didn't make our main news stories because there was a lot happening. So here we go. Our quick list here of what's new and what's next. Alright.
Jordan Wilson [00:27:52]:
Meta. Sorry. Still overcoming a sickness here, so losing my voice. But Meta has signed licensing deals with new, news publishers for Meta AI. Apple got a new AI chief after their old AI chief has retired. Anthropic acquired BUN, JavaScript runtime, marking the company's first ever major acquisition. Clang AI released their b two six of their video product. Pretty good, but I would say not really on v o three one, or Sora level just yet.
Jordan Wilson [00:28:22]:
Speaking of Google, Google is reportedly working on a Nano Banana 2 Flash. So might be a little confusing because I think it's technically, Nano Nano Banana Pro, and then Nano Banana 2 Flash. So we'll see if that's the naming that they stick with. Mistral AL released the Mistral 3 family in 10 open models. Google may be rolling out Gemini 3 to NotebookLM soon. So currently, NotebookLM runs on Gemini 2.5, so it may be getting the Gemini 3 update soon. OpenAI announced a $4,600,000,000 Australian AI campus. OpenAI also announced yeah.
Jordan Wilson [00:29:02]:
This one, I think, threw a lot of people for a loop. They announced an ownership stake in Thrive Holdings. Right? Yeah. So OpenAI getting into the capital game. Google, according to Leaks, is testing project, projects and scheduled actions manager. Runway announced their gen four five AI video model, and it actually topped Google, Google's v o three one and OpenAI's Sora, two on a lot of the benchmarks. However, it seems like it hasn't really rolled out yet. So they said, like, hey.
Jordan Wilson [00:29:38]:
It's out. But, most people, including myself, I've had a Runway subscription since I don't know. It came out. I don't have access to that yet. I don't see a lot of people do. So I don't know. Maybe they had a problem. They couldn't, maybe they didn't have enough compute to roll it out.
Jordan Wilson [00:29:52]:
Maybe they ran in some issues after they announced it. So a pretty interesting one there from runway. Notebook l m launched, like, over the weekend, I think, launched some major mobile updates including image support. That's one that I love. I just wish you could do more than 10 images at once because I I have so many screenshots on my phone, and just being able to upload them into notebook l m as kind of just my my my brain, to be able to do that quickly, great. But, unfortunately, only 10 at a time. Meta acquired AI wearables startup limitless, formerly called Rewind. And apparently, they just might be squashing their development and just having them work on Meta Ray Bans.
Jordan Wilson [00:30:29]:
Apple displays executive, in charge of the Vision Pro joined OpenAI's hardware team. Pretty big news. We've been talking about that a lot. OpenAI making a big play on hardware, specifically coaching a lot of people from Apple. xAI's GROC 4.2 O wins a real money trading league, outperforming other frontier models. OpenAI's People First AI fund awarded $50,000,000 to nonprofits. OpenAI tested a confessions training method so that models could self-report hallucinations and shortcuts. That was a fun one to read.
Jordan Wilson [00:31:06]:
Anthropic landed a $200,000,000 Snowflake deal, integrating Claude 4.5 into enterprises. OpenAI acquired Neptune AI, boosting their internal training analytics infrastructure, and NVIDIA invested $2,000,000,000 in Synopsys for AI chip design. Alright. That's it for this week, but AI doesn't sleep and we won't either. It is the end of the year, which means a lot of us are looking back on 2025, taking an inventory, seeing what worked, what didn't, and planning ahead towards 2026. That's exactly what we're gonna be doing here as well, over the rest of December and in January. So if you are making a concerted effort to grow with AI, to use AI more in your company, to get more of your team on board, that's exactly what we're gonna be doing. Hopefully, some fun announcements here, in the next week or two, so make sure you stay tuned for that.
Jordan Wilson [00:32:05]:
So I hope this is helpful. We spent a lot of time trying to make sure you are the smartest person in AI at your company. So if this is helpful, if if you're on the podcast, please make sure to like and subscribe to the show. I'd appreciate it if you could also leave a rating. Takes, like, ten seconds. That'd be great. Also, if you're listening on the livestream, please take a minute to repost this. Don't be greedy.
Jordan Wilson [00:32:25]:
Share this free AI information with your friends, your neighbors, your babysitters, your dog walkers, your friends' neighbors, babysitters, dog walkers. Make sure they all know the realists in AI. That's what we would bring you. Thanks for tuning in. Hope to see you back tomorrow and every day for more everyday AI. Thanks, y'all.