The AI Daily Brief: Artificial Intelligence News and Analysis - The 5 Biggest AI Stories to Watch in December

Starting point is 00:00:00 Today on the AI Daily Brief, the five biggest AI stories to watch in December. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. All right, friends, quick announcements before we dive in. First of all, thank you to today's sponsors, Super Intelligent, Robo, Robots and Pencils, and Blitzy. To get an ad-free version of the show, go to patreon.com slash AI Daily Brief, or you can subscribe on Apple Podcasts. And if you are interested in sponsoring the show, send us a note at sponsors at aidailybrief.ai. I know that this is a busy time for lots of folks, and this is the last time to lock in 2025 prices for your ad buy.

Starting point is 00:00:41 So again, shoot us a note at sponsors at A.I.Daily Brief. And the team will get back to you with all the information you need about sponsorship. We got a big one though today. So without any further ado, let's dive in. Welcome back to the AI Daily Brief. Last month at the beginning of the month, we did a look at the biggest stories and themes and discussions that I thought were going to shape the AI conversation coming into November. That ended up being a really popular episode. A lot of you commented that you liked the combination of a little bit of a

Starting point is 00:01:10 retrospective of the month before, plus a look ahead as a way to ground yourself heading into the new month. And so I think at least for the foreseeable future, we're going to keep trying that at the beginning of each month, unless, of course, there's some huge newsday that crops things out. And I think in particular, heading into December, these themes are almost doing double duty as previewing where I think a lot of the conversation is going to be headed into 2026. This general period as one year turns into the next is, of course, a period of both reflection and planning,

Starting point is 00:01:37 and so I think makes this whole type of approach even more pertinent. Now, of course, what I think we need to do first is look back at what I had suggested were going to be the big important conversations in November and see how we did. The first one with a bullet, at least in terms of how important the conversation was going to be, was of course, the release of Gemini 3. Now, what I got right about this was that I framed it as the big remaining release of 2025 and how high stakes the release would be for both Google and for OpenAI, basically that this, if it happened, would be a seminal moment one way or another. The reason that I consider it still a bit of a miss is that I was much more skeptical than

Starting point is 00:02:15 some other commentators that it would actually happen in November. My skepticism was basically that I hadn't seen a ton of indications from the Google team that it was for sure coming in November, and I thought that the stakes were just so high that there was a non-zero chance that it would get delayed a little bit until they were sure that it would be a huge hit. Ultimately, obviously, we did get Gemini 3, and it was the biggest story of November. Although, as we'll see in a minute, something that we missed that was important was the side-long release of Nanobanana Pro. The second discussion that I suggested was going to be important when it came to November was around the AI bubble. And basically my argument was that while I thought

Starting point is 00:02:50 that people might start to have some amount of narrative fatigue around that circular investment chart that was going around, I thought we were likely to start seeing more focus on the political and jobs angle heading into the midterms. And I think that that definitely happened. Frankly, I underestimated just how sticky the market bubble narrative would be, even at one point having to do a full episode about how I wasn't going to do episodes about it anymore because it was just such an unprovable thing in the short term. However, we definitely saw an increase in the politicization of AI and specifically questions around AI-related job loss. We saw thousands and thousands of white-collar jobs disappear, such as HP announcing 4,000 to 6,000 job cuts. And even though in many cases, it seemed

Starting point is 00:03:33 fairly dubious that they were actually about AI, we began to see how AI would be used as a fall man for pretty much any types of layoffs that a company needed to do. What's more, we also saw a meaningful increase in political narratives around AI. Bernie Sanders increased his rhetoric around it, as did on the other side of the aisle, Florida Governor Ron DeSantis. We saw some bipartisan bills like the one from Josh Holly and Mark Warner to force companies to report AI-related layoffs. And I think that as we head into 2026 in the midterms, this is going to get more and more significant. On vibe coding, I said that I thought that it was the breakout use case of 2025, and the conversation that I thought we'd see more of was that it would be less about vibe coding

Starting point is 00:04:13 or not, but more about the right configuration, speed versus autonomy versus control. I thought that debate would be a template for. for other AI use cases and that there would be some amount of recalibration between agentic AI and assisted AI. Now, Gemini and ChatGPT both thought that this one was a hit. Vibe coding, for example, was named a word of the year. And there were numerous think pieces over the last month about the sort of paradox of vibe coding where AI could code faster than any human, but it creates all these new challenges

Starting point is 00:04:40 around experience and human oversight. And I think that we're going to continue to see this shift away from the first order questions of things like, is vibe coding real, to more nuanced conversation. around what the implications are and what they might be in different ways for technical versus non-technical people and software engineers versus non-software engineers and what it all suggests for agents in other areas more broadly. The fourth prediction was around the emergent 2026 discourse where I basically said that I thought that we were going to see emerging pillars around the product era of AI, in other words, a focus on applications in U.X, not just models,

Starting point is 00:05:12 a focus on context engineering and orchestration around getting the right data and structure around models and a bigger conversation around ROI happening, particularly with regard to enterprise AI. Now, this one is inherently a little bit more blurry than did Gemini 3 launch or not, but I certainly think it's the case that you started to see more of this, particularly around the ROI conversation. Now, obviously, we were a big part of that with our AI-R-OI benchmarking study running throughout the month, but we also got that much more sophisticated Wharton ROI survey, as well as reports from companies like McKinsey around the topic, and I think that's going to do nothing but accelerate in the month to come. The last area that I was

Starting point is 00:05:46 watching was to see if in the lead-up to AWS reinvent, which is actually just kicking off now, we'd get any hints or indications around how Amazon might reposition itself. I guess we did a little tiny bit in the sense that we saw a big deal between AWS and OpenAI, but overall it was pretty quiet, and so we're calling this one incomplete or neutral. Now, a couple of stories that did end up being important that we didn't really mention include some of the model drops, although I think, honestly, we could have probably assumed that if Gemini 3 came, there would be a response from chat GPT, like 5-1, although the fact that they brought it out in advance of Gemini 3, I don't think, was on too many people's radar. But I definitely think that

Starting point is 00:06:24 the most significant and impactful thing that happened, while within the orbit of Gemini 3 was the Nanobanana Pro launch, which I think completely on its own terms is extremely significant. All right, so that was November, but let's talk about what we're seeing heading into December, as key AI stories to watch. Many of these are building off of last month, and the first one has to be the Gemini narrative ascending. Coming off of this very successful release of Gemini 3 and Nanobanana Pro, Google's position in the AI race has never looked stronger. And some of that is substantive and in numbers. Over the weekend, we got this article from the Financial Times titled OpenAI's lead under pressure as rivals start to close the gap. It shared a couple of charts which you might have seen flying

Starting point is 00:07:06 around the socials, including how Gemini's app downloads are catching up to chat GPs, although their total user base remains a couple hundred million behind. You can see this huge inflection in the August-September period. And maybe more significant was this chart of the average time spent per visit, where for the first time, Gemini took the lead from ChatGPT. Now, it should be noted that this is average minutes per visit, not overall time per day, which is to me, I think, probably somewhat more significant, but it's still part and parcel of this overall story of Google's momentum. You're also seeing it just in the general user discourse. There's this post on Reddit in the OpenAI subreddit called Can Anybody Stop Google that's been shared on all the other networks

Starting point is 00:07:46 numerous times now? The post reads, I spend some time using Google products over the weekend. I have a $20 subscription to Google Plus. Honestly, other services cannot compare between Google Docs, Notebook LM, quality of Gemini 3 and App and website, Gemini plug-in for Chrome, Google CLI and VS code. Can anybody compete? The boring business account on X captured the shift writing pretty amazing just how quickly the narrative shift is happening from OpenAI to Google. Making things more challenging for OpenAI is that in addition to this being a Google good narrative, there's also some skepticism around a few moves from OpenAI more broadly, regardless of Google competition. This post from T. War Blahoe a couple days ago went viral after suggesting that OpenAI seemed on the

Starting point is 00:08:27 verge of integrating ads into their products. Ross Hendricks, who it should be noted, spends basically all of his time on X pushing the AI bubble narrative, still captured a lot of people's sentiment when he wrote, this will put the final nail in the coffin on Open AI. Millions of free consumers will switch to Gemini as chat GPT becomes filled with AI slop ads. Google can afford to run a purely free non-ad model as a loss leader until it fully bleeds out open AI. Now, I think that the first part is wildly overdramatic and doesn't actually comport to how user behavior works in the real world. But the point that he's making that Google has more financial might to prioritize user experience over revenue is something a lot of folks have discussed. We're also seeing the beginning

Starting point is 00:09:06 of age verification, which frankly could go both ways. Some people are sky. skeptical of the move, but other people think it's going to be good to have a more verifiable adult experience available in ChatGBT. Racer X, on the other hand, writes, absolutely thrilled about this. Age verification rolling out means we are finally getting the full adult feature set in Chatchipt, fewer interruptions, more creative freedom, richer roleplay options, and overall a much smoother experience. This is a huge step towards more capable and expressive AI.

Starting point is 00:09:33 Still, while that's a positive take, overall, Open AI is in a bit of a narrative doldrums right now, exactly, frankly, as Sam Altman predicted in the lead-up to Gemini 3's release, and that got even tougher after a recent semi-analysis report claimed, quote, OpenAI's leading researchers have not completed a successful full-scale pre-training run that was broadly for a new frontier model since GPT40 in May 2024. Now, this report generated a ton of conversation and highlights the other part of the pro-GoL narrative, which has had a major rise, which is around their TPUs. With Gemini 3 being as performant as it was, and it being trained entirely on TPUs, and there

Starting point is 00:10:10 being news that companies like meta were potentially engaging to buy Google TPUs directly. Not only did Google start to chip into the OpenAI narrative, but it also caused some consternation for Nvidia, which you'll remember last week, took the unusual step to write a post on X about how far ahead they were. All in all, heading into December, the Gemini narrative is ascending and as strong as it's ever been, and I think how this plays out over the next month is going to shape a lot of how 2027 looks at least at the beginning. Today's episode is brought to you by my company, Super Intelligent. Look, guys, buying or building agents without a plan is how you end up in pilot purgatory.

Starting point is 00:10:50 Super Intelligent is the agent planning platform that saves you from stalling out on AI. We interview teams at scale, translate real work into prioritized agent opportunities, and deliver recommendations that you can execute on, what to build, what success looks like, how fast you'll get results, and even what platforms and tools you should consider. All customized for you. Instead of shopping for hype, you get to deploy with confidence. Visit B-Super.AI and book your AI planning demo today. Meet Rovo, your AI-powered teammate.

Starting point is 00:11:20 Rovo unleashes the potential of your team with AI-powered search, chat, and agents, or build your own agent with studio. Rovo is powered by your organization's knowledge and lives on Atlassian's trusted and secure platform, so it's always working in the context of your work. Connect Rovo to your favorite SaaS app so no knowledge gets left behind. Rovo runs on the teamwork graph, Atlassian's intelligence layer that unifies data across all of your apps and delivers personalized AI insights from day one.

Starting point is 00:11:48 Robo is already built into Jira, Confluence and Jira Service Management Standard, Premium, and Enterprise subscriptions. Know the feeling when AI turns from tool to teammate? If you rovo, you know. Discover Rovo, your new AI teammate powered by Atlassian. Get started at ROV as in Victory O.com. AI isn't a one-off project. It's a partnership that has to evolve as the technology does. Robots and Pencils work side by side with clients to bring practical AI into every phase,

Starting point is 00:12:17 automation, personalization, decision support, and optimization. They prove what works through applied experimentation and build systems that amplify human potential. As an AWS-certified partner with Global Delivery Centers, robots and pencils combines reach with high-touch service, where others hand off, they stay engaged, because partnership isn't a project plan. It's a commitment. As AI advances, so will their solutions. That's long-term value. Progress starts with the right partner. Start with robots and pencils at robots and pencils.com slash AI Daily Brief.

Starting point is 00:12:49 This episode is brought to you by Blitzy, the Enterprise Autonomous Software Development Platform with infinite code context. Blitzy uses thousands of specialized AI agents that think for hours to understand Enterprise-scale code bases with millions of lines of code. Enterprise engineering leaders start every development sprint with the Blitzy platform, bringing in their development requirements. The Blitzie platform provides a plan, then generates and pre-compiles code for each task. Blitzy delivers 80% plus of the development work autonomously, while providing a guide for the final 20% of human development work required to complete the sprint. Public companies are achieving a 5x engineering velocity increase when incorporating Blitzy as their pre-IDE development tool, pairing it with their coding pilot of choice to bring an AI-native

Starting point is 00:13:29 SDLC into their org. Visit blitzy.com and press get a demo to learn how Blitzie transforms your SDLC from AI-assisted to AI native. Now, of course, maybe the biggest thing to watch for, which could kind of be a perennial contender on this list, is whether we actually get any new models this month. Now, part one of this conversation, of course, has to be from the big guys. And specifically, I think the thing that everyone is wondering is, will we get some new model from Open AI to try to reclaim some of that momentum? I polled for that last night on Twitter and found a little under two-thirds of people saying no, with just a little over a third of people.

Starting point is 00:14:11 saying yes, but others are more optimistic. Part of that optimism is sort of generic, as we're starting to see an upt in OpenAI vague posting, Adam GPT from OpenAI, for example, tweeted this weekend, it's the bottom of the second inning, as well as a basketball still, which had people wondering if this is the first public glimpse of a new image generation model from OpenAI. Now, I will say, and this is the prediction that the AIs will judge me on at the beginning of next month when we're looking back, I think if OpenAI has an updated image generation model, they will release it this month. And if I was forced to make a prediction on whether we will actually see one, my answer would be yes. I think at this point, it's not just about stealing momentum back from Gemini. I think it's that the

Starting point is 00:14:55 raw functionality and increased value of the native multimodality that is now embedded in Gemini kind of forces OpenAI to catch up from just a utility perspective, hold aside whatever the narrative is. Right now, the biggest reason that I find myself on any given task switching from ChatCTPT to Gemini 3 is less likely to be about perceived model quality differences and more likely to be about the integration of image generation as part of an overall goal set. So if I have one big prediction from a new model perspective, I do think that we will see an image generation model from OpenAI this month. Interestingly, though, while people might be wondering what OpenAI specifically is going to do

Starting point is 00:15:35 and whether we'll get any more from the big labs, just hours into December 1st, and already more new models have been released this month. One big one is DeepSeek V3.2, which DeepSeek calls a reasoning first model built for agents. Now, we're going to dig into this, I'm sure, more tomorrow when we have a little bit more space, as it just came out as I was preparing this episode,

Starting point is 00:15:57 but initial impressions nudged towards the impressed. Chubby on X writes, The Whale is back. Holy moly, look at those e-vals. state-of-the-art, even outperforming Gemini 3.0 Pro and GPD5 high on several benchmarks. Now, Chubby said, of course, it needs to be tested if it's just benchmark maxed, but so far very impressive. Jimmy Apples points out that they released an IMO gold medal model before Google or OpenAI did. Remember, while both Google and OpenAI have claimed gold metal performance in

Starting point is 00:16:23 the IMO, they haven't released a specialized version of that model, and Deepseek now claims to have that same performance. Clem DeLang from Hugging Face wrote, As far as I know, there isn't any chatbot or API that gives you access to an IMO 2025 gold medalist model. Not only does this change today, but you get to download the weights with the Apache 2.0 open source release of this new deep seek model. When Dan Mack argued that the benchmarks weren't that impressive and asked as deep seek still relevant, some pointed out that it was important to take into consideration cost, which people are putting at roughly 30 times cheaper than Gemini 3.0 Pro. Now, it wasn't just deepseek. Another model that people have been wondering

Starting point is 00:16:59 about is the video generation model that I've been floating around in testing with the codename whisper thunder. Turns out whisper thunder is the new runway gen 4.5, which the company claims sets a new state of the art for video. Rites runway, Gen 4.5 is state of the art and sets a new standard for video generation motion quality, prompt adherence, and visual fidelity. It certainly outperforms on the text of video leaderboard from an ELO score perspective, and it seems like a lot of the advancement would fit in the category of what I would call the unlock score, basically improvements that unlock use cases that would have been difficult if not impossible before.

Starting point is 00:17:35 Runway writes, the model excels at understanding and executing complex sequence instructions. You can specify detailed camera choreography, intricate scene compositions, precise timing of events, and subtle atmospheric changes all within a single prompt. Runway's CEO, Cristobal Valenzuela, writes, I haven't had this much fun with a model ever. Whisper Thunder delivers every time. One of the parts I like most is how steerable it is with movements, cameras, and super-specific prompts. Now, once again, I need more than 45 minutes to really understand how much has changed

Starting point is 00:18:05 with this model release, but it is extremely notable if we are talking about the conversations that will shape AI in December that right out of the gate, we have a couple major model releases from the contenders in the App Labs, which brings us to our next conversation to watch, which I'm calling vertical agent lab positioning. And really what I'm referring to here is the idea that it is not just the foundation model companies that are jockeying for position. There are entire other categories of companies that are arguing that specific vertical focus is ultimately going to be what wins, especially when it comes to things like business usage. Now, Sean, aka Swix, has articulated this in his agent labs thesis,

Starting point is 00:18:44 which you can find at latent. space. Remember, Sean joined cognition and has talked a lot about this idea of agent labs recently. The information recently picked up that investors are pouring a ton of money into this next generation of labs. And I think that heading into 2026, a lot of these companies are going to take advantage of the end of the year to try to get themselves positioned to be the bearers of ROI and impact, especially for business users heading into next year. An example of what I'm expecting to see a lot more of is the recent Sierra announcement that the company had reached $100 million in ARR as part of a broader press push to really try to claim some of the territory around enterprise customer service AI. Now, if I think that the vertical and

Starting point is 00:19:25 Agent Labs are going to be positioning heading into the new year. I also think that their counterparties and the enterprises are going to be doing a bit of turf claiming as well. Now, admittedly, what I mean here is a little bit more vague. But in short, I think that broadly speaking, the story of 2025 when it came to Enterprise AI is that this stuff, by which I mean AI and agents, is extremely valuable, but not as simple as flipping on some switch and hoping it all works. It involves real serious reorganization, reimagining of the enterprise.

Starting point is 00:19:57 It involves data readiness, training, capacity building, leadership, new approaches to risk, new thinking about workflows. And increasingly, we are seeing a big gap between the leaders and the laggers. And in that space, I think we're going to see enterprises start to claim some turf as firmly in that leader camp. So what might that look like? Well, I did notice literally as I was recording, that OpenAI and Accenture, had made some announcement around some much deeper integration, and I think it's going to be things

Starting point is 00:20:26 like this. Accenture to equip tens of thousands of its professionals with ChatGBTBTE Enterprise and have the largest number of professionals upskilled through OpenAI certifications. Flagship AI client program launched to help organizations bring AI into every part of their business. Deloitte had their massive deal rolling out Anthropic to 470,000 employees at the beginning of October, and I don't think it's just going to be the consulting firms. I think you're going to see enterprises across every category, start to claim some of this AI. turf as well. Next up, as much as I would love for it not to be the case, I do think that the AI bubble conversation continues headlong right on into 2026. There's just too much momentum right now

Starting point is 00:21:04 around the OpenAI bear case. This new HSBC report is getting a bunch of traction, articulating why they think there's going to be a big hole for Open AI, even if they make a ton of money. I think you're going to see some amount of focus shift away from the surface layer conversations of boom versus bubble to more specific focus around things like we've been talking. talking about, like the financing structure of a lot of these deals. And counter to that, I think that you will start to see more and more people point out similar to we have on this show that even if we're in a bubble, it's not one that really has any foreseeable short-term catalyst for popping. And so at some point, we're going to get at least a little bored of the

Starting point is 00:21:40 narrative. I also think that bulls want to be in control on Wall Street, and especially if we get a rate cut that now sits at 90% odds according to markets. We're going to finish the year, not strong necessarily. I think too many people are taking profits and closing it out for it to be a real big run on the market, but certainly not in the doldrums. If I have to make one specific prediction, I think before this particular narrative cycle ends, Google flippins Nvidia to become the most valuable company in the world. I don't necessarily think that happens in December, but I think at some point it happens. It sort of strikes me as the inevitable implication of all of this, and I don't think it would take more than a couple of TPU deals announced for Google to jump up ahead of

Starting point is 00:22:20 Apple and then Invidia. Rounding us out just a couple more, the next is an extension and frankly an acceleration of the trend that I was calling out last month, which is the acceleration in anti-AI politics as part of a larger positioning exercise heading into the midterms. Now, I think you're going to see this happen both behind and in front of the scenes. We've got the emergence of pro-AI political action committees. We've also got increased media scrutiny of the White House's relationship with AI. We've got the growing battle between the federal government and state governments around approaches to AI regulation. You'll remember that Trump had to back off his executive order that would have initiated a committee to go basically sue states that tried to put into place their own

Starting point is 00:22:59 AI legislation. And of course, he had to back down because of his own party. In fact, if I had to make a very specific prediction for what we're going to start to see in December, I think you're going to start to see a more full-throated and clearly articulated anti-AI position from the right. That doesn't mean that I think that all of the right will become anti-AI, but I think that those who are will get louder, and I think they'll start to put it in political soundbite terms that can be easily repeated by their base. I think Bologi Shrinivasa is directionally correct with his tweet, 2020, blue and tech against red, 24, red and tech against blue, 28, blue and red against tech. Now overall, I think that December is going to be prelude to where this really picks up,

Starting point is 00:23:41 which is next year with the midterm election cycle, but I still think it is going to be noticeable even this month. Lastly, More of a question than anything else, will we see signs of life from the big tech giants that have lagged a little bit behind when it comes to AI, Amazon, or Apple? Let's do Apple first. If forced to predict my answer is probably not. I think that the dye has been cast, and I think that every week that goes on, Apple gets farther and farther away from trying to compete in any meaningful way on AI, and will at some

Starting point is 00:24:11 point just use its balance sheet to buy their way in and make sure their products and devices don't suffer for lack of AI, even if they don't control the whole. experience. Maybe we do get official confirmation of the Apple-Google Gemini relationship, but that's about as much as I can see. Amazon, I think there's a little bit bigger of a question. AWS's big reinvent conference kicks off today on Monday, December 1st, and runs throughout the week. This has historically been a time that Amazon has tried to reposition for the coming year when it comes to their AI strategy, and so while I don't know exactly what to expect, it is something that I'm going to be watching closely. If I had to make a prediction, I don't think that this event is

Starting point is 00:24:48 going to be about model releases like the Amazon Nova family that was announced last year at this time. Instead, I kind of think AWS might really try to focus down on their cloud leadership and try to pick back up ground that has been lost a little bit to Google Cloud and Microsoft Azure in the AI era. It may be that instead of trying to compete on all the fronts, including the model front, they really hone in on being the infrastructure provider for those enterprise AI leaders that we talked about before, and position themselves as the cloud provider that actually understands what the enterprise needs and can provide that without any conflicts of interest with regard to their own models. Now, I'm not sure if that's what we'll see, but if I had to make a prediction,

Starting point is 00:25:29 that's what it would be. In any case, I do think that it will be relevant to watch, and once completed will help us understand really what the state of the AI race is across all the big tech companies heading into 2026. So those are the biggest AI stories to watch in December, some a little bit new, some a continuation of what we had before, and a lot positioning for the year to come. I'm looking forward to covering it here with you guys. Despite the holidays, most of the month will be a pretty normal month here until we get towards the very end. But for now, that's going to do it for today's AI Daily Brief. Hope you all had a great long Thanksgiving weekend.

Starting point is 00:26:03 And until next time, peace.

The AI Daily Brief: Artificial Intelligence News and Analysis - The 5 Biggest AI Stories to Watch in December

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.