The AI Daily Brief: Artificial Intelligence News and Analysis - Power Ranking the Big AI Ideas for 2026

Episode Date: December 21, 2025

This Sunday long-read episode digs into a16z’s newly released Big Ideas for 2026, scoring the most interesting predictions across likelihood, real-world value, and pure X-factor. From taming multimodal data chaos and agent-native infrastructure to voice agents, multiplayer vertical AI, AI-native universities, and the industrial renaissance powered by software and automation, the episode separates what feels inevitable from what feels premature—and what’s just plain cool. The result is a highly subjective, intentionally unscientific power ranking designed to spark end-of-year reflection and a few joyful holiday arguments about where AI is actually headed next.Read the Big Ideas: https://a16z.com/newsletter/big-ideas-2026-part-1/Brought to you by:KPMG – Discover how AI is transforming possibility into reality. Tune into the new KPMG 'You Can with AI' podcast and unlock insights that will inform smarter decisions inside your enterprise. Listen now and start shaping your future with every episode. ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://www.kpmg.us/AIpodcasts⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Rovo - Unleash the potential of your team with AI-powered Search, Chat and Agents - ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://rovo.com/⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Zenflow by Zencoder - Turn raw speed into reliable, production-grade output at ⁠⁠⁠⁠https://zenflow.free/⁠⁠⁠⁠LandfallIP - AI to Navigate the Patent Process - ⁠⁠⁠⁠https://landfallip.com/⁠⁠⁠⁠Blitzy.com - Go to ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://blitzy.com/⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ to build enterprise software in days, not months Robots & Pencils - Cloud-native AI solutions that power results ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://robotsandpencils.com/⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠The Agent Readiness Audit from Superintelligent - Go to ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://besuper.ai/ ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠to request your company's agent readiness score.The AI Daily Brief helps you understand the most important news and discussions in AI. Subscribe to the podcast version of The AI Daily Brief wherever you listen: https://pod.link/1680633614Interested in sponsoring the show? [e-mail]

Today on the AI Daily Brief, a power ranking of some of 2026's big ideas in AI. 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, KPMG, Blitzy, super intelligent, and robots and pencils. 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 to find out anything you need to about the show, including sponsorship opportunities, or our latest research, go to A.

AID Daily Brief.aI. Now, today is Sunday. We are getting very, very close to the Christmas and New Year holiday. The plan as of right now is to have Monday and Tuesday be our last regular format episodes of the year. And then starting from Wednesday the 24th, we will be into our end of year pre-records. We'll have a couple of interesting interviews and a lot of look back and look forward type of content. But for this Sunday long read slash big think episode, I've been following along as

A16Z drops their big ideas for 2026. Basically, what they did is they went to their partners and asked what they thought people would be building in the new year. Now, not all of the ideas that they shared are about AI, but as you might imagine, a good chunk of them are. And so what we're going to do today is go through and review some of the ones I find most interesting and give them a power ranking. We're going to score them one to five on how likely I think it is to come to pass, one to five on how valuable I think it would be if it did come to pass. And then one to five as an X factor, which can be anything that I think is a different. interesting for basically any reason. As you can tell, this methodology is highly scientific, and of course completely subjective to myself. Mostly it's just a fun way to share these ideas,

but also give us all something to have joyful holiday arguments about. kicking off, we have an idea from partner Jennifer Lee. Startups tame the chaos of multimodal data. Jennifer writes, unstructured multimodal data has been enterprise's biggest bottleneck and their biggest untapped treasure. The limiting factor for AI companies is now data entropy, the steady decay of freshness, structured, and truth inside the unstructured universe where 80% of corporate knowledge now lives. That's why she writes, untangling unstructured data becomes a generational opportunity. Enterprises need a continuous way to clean, structure, validate, and govern their multimodal data, so downstream AI workloads actually work.

All right, so we're kicking off of the bang. Let's start with how likely I think this is. This gets at least a four on the likely scale, maybe even five. This is such a huge problem that is so universally recognized that I absolutely think there's going to be a set of startups that go after this incredible opportunity. The only reason that I'm not 100% sure it's a five is that it's such a difficult and complex problem that I wouldn't be surprised if we don't just see one overarching effort to try to get everything, but more surgical efforts around particular categories of data to start. In terms of value, I also have this as a four. I think that for the companies who all of a sudden can unleash all of their organized and structured data on the world of AI and

agents, it's going to be unfathomably valuable. But the only reason it's not a five is that how much the world cares about and needs structured corporate data is perhaps at least a little bit more of a discussion. For an X factor, I gave this just a one. It's not a glamorous build. It's just something that needs to happen. Next up, a prediction from Malika Abakarova. Agent Native infrastructure becomes table stakes. She writes, in 2026, the biggest infrastructure shock won't come from outside companies but from within. We're shifting from human speed traffic that's predictable and low currency, to agent speed workloads that are recursive, bursty, and massive. The enterprise backend of today was built for a one-to-one ratio of human action to system response. It's not

architected for a single agentic goal that trigger a recursive fan out of 5,000 subtasks, database queries, and internal API calls in under milliseconds. When an agent attempts to refactor a codebase or remediated security log, it doesn't look like a user. To a legacy database or rate limiter, it looks like a DDoS attack. Building for agents in 2026 means re-architecting the control plane. We'll see the rise of agent native infrastructure. The next generation must treat thundering herd patterns as the default state. All right, so this one is interesting to me. On the one hand for likely, I'm giving it a two, not because I don't think that she's right, but just because she's defining an incredibly broad pattern instead of something precise. Now, that's not a knock. It's just the

nature of this particular type of prediction. However, I think that she's right that there is going to be a very different set of expectations around how we get things done once we fully unleash the power of agents. I think the value of something like this comes to pass is at least a three and maybe even getting into a four or five just because of what new capabilities become unlocked when you can, in her words, trigger a recursive fan out of 5,000 subtasks. Speaking of which, for an X factor, I'm giving this one a four because you may recognize embers of the Dr. Strange theory here in this prediction. The Dr. Strange theory is, of course, the way that I've explained in the past, that I think that while right now we're kind of viewing agents as one-to-one replacements for human labor,

In the future, it will be very different and we will have legions of agents doing both different sets of tasks and then recombining as well as doing the same task over and over and over in order to see which one does best. Building the infrastructure for that is going to be a big change, and I would love to see that be something that a lot of entrepreneurs dive into in 2006. Next up, we have a prediction from prolific A16Z content creator and partner Justine Moore, who predicts creative tools will go multimodal. She writes, We now have the building blocks to tell stories with AI, generative voices, music, images, and video. But for anything beyond a one-off clip, it's often time-consuming and frustrating, if not impossible, to get the outputs you want,

especially if you want anywhere near the level of control that a traditional director would have. Why can't we feed a model a 30-second video and ask it to continue the scene with a new character created from a reference image and voice? Or reshoot a clip so that we can see a scene from a different angle, or make the motion match a reference video. 26 is the year that AI goes multimodal. Give a model whatever form of reference content you have to work with to make something new or edit an existing scene. We've already started to see some early products here like Klinga1 and Runway Aleph, but there's a lot more to be done, and we need innovation at both the model and the application layers. So I scored this one a little bit lower.

I put the likelihood at a two and the value at a two, although I might be misinterpreting exactly what Justine is saying. It seems to me like she's suggesting a user experience pattern that's about getting a lot with very little, about giving an AI tool, just a very small input, and letting it run wild and doing a ton. Now, if that's not what she's saying, I apologize for rating it too lowly. But my sense of how this plays out, at least in terms of the next sequence of things that are going to be built, is I think we're going to get much more prosumer-type tools before we get the next big iteration of general consumer versions of this. We already have a lot of these creative multimodal tools that basically come down to

give a little bit of text or a reference image and kind of let the AI do with. whatever it wants. I think where things are lacking and where we're starting to see the trend is around getting more fine-grained controls for professional or prosumer-type users. I think we're going to get AI-native Cap-cut, which, by the way, might just be Cap-cut, before we get a wild additional array of creative tools inside something like Sora. Now, the X-factor here is a three or four because I think that this sort of creation is really cool. I just think that the phase that we're going into next is a prosumer and professional phase for creative tools rather than a general consumer phase. That said, I could easily see being wrong on this one. We have not yet figured out

what if any native social network will arise around AI generated content. And that is such a big prize that I would not at all be surprised to see people spending a lot of cycles and entrepreneurial energy around it. Next up, Yoko Lee coming in with the philosophical, saying that 2026 is the year we step inside video. Yoko writes, in 2026, video stops behaving like something we passively watch and starts feeling like a place we can actually step into. Video models can finally understand time, remember what they've already shown us, react when we do something, and hold together with the kind of quiet consistency we expect from the physical world. Instead of producing a few seconds of disconnected imagery, these systems sustained characters, objects, and physics, long enough

for actions to matter and consequences to unfold. This shift turns video into a medium we can build on, a space where robots can practice, games can evolve, designers can prototype, and agents can learn by doing. What emerges is less like a clip and more like a living environment, one that starts to close the gap between perception and action. For the first time, it feels like we can inhabit the videos we generate. All right, so very neat, but I'm going to be a wet blanket here and give this a sad, solitary one on the likelihood. This is not about the trajectory of this prediction. It's about the timeline. I just simply do not believe that 2026 is going to be the year where we're going to see these types of capabilities. We are still just nudging into what we can do with AI video. And I don't

think we're going to be anywhere near to where Yoko is predicting here in 2020. I don't even know if we're going to be here in 2027. This feels like a 28-29 type of thing to me. I also think that because of that, any version that we get next year is going to be around a 1 on value, which is not to say that it won't be a 5 eventually and totally transform how we interact with the world

and entertainment, but in the short term I'm putting both likely and valuable at 1. However, where Yoko recovers some scores on the X factor, which I'm giving a 5 because this is hell of cool and almost the polar opposite of the multimodal data, which is so obvious, so present, and so important, but does little to stir our souls as compared to something like this, which we can really imagine a very different future to the world we inhabit today.

Next up, a prediction from Alex Imerman. Vertical AI evolves from information retrieval and reasoning to multiplayer. Alex writes, AI has driven vertical software to unprecedented growth. Healthcare, legal, and housing companies reached $100 million A.R. within a few years. Finance and accounting are close behind. The evolution was first information retrieval, finding, extracting, and summarizing the right information. 2025 brought reasoning. 2026 unlocks multiplayer mode.

Vertical software benefits from domain-specific interfaces, data, and integrations. But vertical work is inherently multi-party. If agents are going to represent labor, they need to collaborate. From buyers and sellers to tenants, advisors, and vendors, each party has distinct permissions, workflows, and compliance requirements that only vertical software understands. Today, each party uses AI in isolation, which creates handoffs without authority.

The AI analyzing purchase agreements doesn't talk to the CFO for its model adjustments. The maintenance AI does not know what the on-site staff promised the tenant. Multiplayer changes by coordinating across stakeholders, routing to functional specialists maintaining context-syncing changes. Tasks performed by AI will be completed with higher success rates, and when value increases from multi-human and multi-agent collaboration, switching costs rise. Here we'll see the network effects that have alluded AI applications. The collaboration layer becomes the moat.

So I'm putting the likelihood of this as a three, and that reflects mostly my sense that this is going to be extremely jagged and wildly unevenly distributed across different industries and horizontal applications. I think that in terms of general trajectory, it would make a lot of sense for this sort of multi-agent collaboration to start to become the norm. However, I don't think that every industry and every space is going to be at the point in 2026 where they're going to be at the stage to actually take advantage of that. However, it is likely that some leading edge spaces will be, and that's why I'm giving it a three for likelihood. The value seems high, like a three or a four. I think Galaxy's right that if these multi-agent systems work and can do these handoffs well,

it will significantly increase the value of the tasks that can be performed by AI. It loses a couple points, though, just for how difficult this is to do in practice. And as an X-factor, once again, I'm giving this a solid three. The dream of multi-agent systems remains, even if that's very much not where we were in 2025, and there's still a lot of progress to go before we get there even in 2026. Next up, a prediction from Stephanie Zhang creating for agents not humans. In 2026, people will start interfacing with the web through their agents, and what mattered for human consumption won't matter the same way for agent consumption. For years, we've optimized for predictable human behavior.

Rank high on Google, appear among the first few items on Amazon, lead with the TLDR. When I took a journalism class in high school, we were taught the 5Ws and H's for news, and to start with a hook for features. Maybe a human would miss the deeply relevant insightful statement buried on page 5, but the agent won't. This shift is about software, too. Apps were designed for humanize and clicks and optimization meant good UI and intuitive flows. As agents take over retrieval and interpretation, visual design becomes less central to comprehension. We're no longer designing for humans but for agents. The new optimization isn't for visual hierarchy but for machine legibility.

And that will change the way we create and the tools we use to do it. So on this, I'm giving it a five for likely. I think 100% this is happening. You're already starting to see this in areas like e-commerce, and I think that that's going to be propagate across the web. In terms of value, I'm not sure. I don't really feel like I have a sense yet of which of the use cases that we distribute to agents

of stuff we used to do on the web is going to be most valuable. And so it's hard for me to gauge how these new interfaces that enable agents better change our lived experience and the value we get. The X factor is similarly weird. This could be a three. It could unlock new possibilities. But it could also be a negative three.

It could be terrible. It could start to tear down things about the web that we know and trust and atrophy other parts of the web that we interact with all the time. kudos to Stephanie for hitting on one that I think is one of the most wildly hard to wrap our heads around before it actually happens. Next up, moving a little bit more practical, Santiago Rodriguez talks about the end of screen time KPI and AI applications.

Santiago writes, for the last 15 years, screen time has been the best indicator of value delivery in both consumer and business applications. We've been living in a paradigm focused on hours of Netflix streaming, mouse clicks in a healthcare EHRUX, or even time spent on chat GPT as the key performance indicator. as we move to a future based on outcome-based pricing that perfectly aligns incentives between vendors and users, we'll first move away from screen time reporting. Now, Santiago points out that this also creates new challenges. How much an application can charge per user requires a more complex method of measuring ROI.

And indeed, then, while I think in some ways this seems pretty obvious, those challenges are not to be underestimated. And so I'm giving the likelihood of this a three. I think that he's probably right that we start to see nudges away from this, but I tend to think that our belief in how fast we're going to shift to totally new methods of measuring value and then pricing has been pretty overstated. I think that we're going to continue to see lots of experiments in outcome-based pricing, but I'm not sure that it's going to become ubiquitous at any time in the next couple of years. I certainly think, though, that there is a

lot of value in moving away from screen time KPI as a metric. Too many things suck and consume our attention. And so even if these are business applications, not entertainment applications, I would still certainly like to see us more focused on measures of outputs than on measures of inputs. As an X-factor, I'm giving it a one. Again, nothing wrong with it. It's just not something that's going to get my particular juices flowing. Hello, friends. If you've been enjoying what we've been discussing on the show,

you'll want to check out another podcast that I have had the privilege to host, which is called You Can With AI from KPMG. Season one was designed to be a set of real stories from real leaders, making AI work in their organizations. And now season two is coming and we're back. with even bigger conversations. This show is entirely focused on what it's like to actually drive AI change inside your enterprise and as case studies, expert panels, and a lot more practical goodness that I hope will be extremely valuable for you as the listener. Search you can with AI on

Apple, Spotify, or YouTube, and subscribe today. 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 Blitzy 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 Blitzie as their pre-IDE development tool, pairing it with their coding pilot of choice to bring an AI-native SDLC into their org. Visit blitzy.com and press get a demo to learn how Blitzy transforms your SDLC from AI assisted to AI native. Today's episode is brought to you by my company Super Intelligent. Superintelligent is an AI planning platform. And right now as we head into 2026, the big theme that we're seeing among the enterprises that we work with is a real determination to make 2026 a year of scaled AI deployments, not just more pilots and experiments. However, many of our partners are stuck on some AI plateau. It might be issues of governance. It might be issues of data readiness. It might be issues of process mapping. Whatever the case, we're launching

a new type of assessment called Plateau breaker that, as you probably guess from that name, is about breaking through AI plateaus. We'll deploy voice agents to collect information and diagnose what the real bottlenecks are that are keeping you on that plateau. From there, we put together a blueprint and an action plan that helps you move right through that plateau

into full-scale deployment and real ROI. If you're interested in learning more about Plateaubreaker, shoot us a note, contact at B-super.a.i with plateau in the subject line. Small, nimble teams beat bloated consulting every time. Robots and pencils partners with organizations, on intelligent, cloud-native systems powered by AI. They cover human needs, design AI solutions, and cut-through complexity to deliver meaningful impact without the layers of bureaucracy. As an AWS-certified

partner, robots and pencils combines the reach of a large firm with the focus of a trusted partner. With teams across the U.S., Canada, Europe, and Latin America, clients gain local expertise and global scale. As AI evolves, they ensure you keep peace with change. And that means faster results, measurable outcomes, and a partnership built to last. the right partner makes progress inevitable. Partner with Robots and Pencils at Robots and Pencils.com slash AI Daily Brief. Next up, another big zoom out from Jonathan Lye.

World Models take the spotlight and storytelling. In 2026, he says, AI-powered world models will revolutionize storytelling through interactive virtual worlds and digital economies. Technologies like Marble from World Labs and Genie 3 from DeepMind already generate full 3D environments from text prompts, allowing users to explore them as if in a game.

As creators adopt these tools, entirely new storytelling formats will emerge, potentially culminating in a generative Minecraft where players co-create vast evolving universes. These worlds could blend game mechanics with natural language programming, such as commanding, create a paintbrush that changes the color of anything I touch to pink. The rise of world models signals not just the new genre of play, but an entirely new creative medium and economic frontier.

This is very similar to the year we step inside video for me, in that I think that the X factor is a four or five. I think this is super fascinating. I think it is a whole new area of exploration of human experience, but I do not believe that it is a lot of, a 2026 thing. I have the likelihood and the value both at a two. And honestly, maybe to remain consistent with Step Into Video, I should have it as a one. The early demos and experiments we've seen from World Labs and DeepMind are incredibly impressive. They show that there is going to be something

phenomenal here. But at this point, we don't even really have like the GPT1 version of this stuff. We certainly don't have anything close to the GPT 3.5 moment for these things. I think that we are going to make a lot of progress in this area next year. I think we're likely to see more. more startups working around it. And we may see even some of the first really niche experiences that capture people's attention. But when it comes to this actual revolution in storytelling, I think you're looking way farther out towards the end of the decade. Next up, a prediction from Emily Bennett, that is something that I've certainly spent a lot of time on based on my background. She predicts the first AI native university, or as she puts it, an institution

built from the ground up around intelligent systems. Over the past several years, she writes universities have dabbled in AI-enabled grading, tutoring, and skisks. scheduling, but what's emerging now is deeper, an adaptive academic organism that learns and optimizes itself in real time. Picture an institution where courses advising, research collaboration, and even building operations continuously adapt based on data feedback loops. Schedules optimize themselves, reading lists evolve nightly and rewrite themselves as new research appears. Learning paths shift in real time to meet each student's pace and context. We're already seeing precursors. ASU's campus-wide partnership with OpenAI produced hundreds of AI-driven projects across teaching and

administration. In the AI Native University, professors become architects of learning, curating data, tuning models, and teaching students how to interrogate machine reasoning. Assessment shifts too. Detection tools and plagiarism bans give way to AI-aware evaluation, grading students on how they use AI, not whether they used it. The AI Native University will become the talent engine for a new economy. Now, once again, I have a fairly high X-factor here, a three or four. I think this is a super interesting and important topic area. In addition to everything else, I'm a venture partner at Learn Capital and first worked with those guys back in 2011, 2012, so I've been thinking about these issues for some time. I built a program at

Northwestern right out of school, and I got kids who are four and seven, who are just at the beginning of their educational journey. So why am I not ranking this one of five across the board? I actually have the likelihood down closer to a two. I have the value also kind of low. And the reason for that might seem weird. It's not that I disagree with the value of anything that Emily is predicting here if I could snap my fingers in it all existed right now. What I wonder about is the longevity of this and how intermediary this step is. The issue for me when it comes to predictions of how education is going to evolve

is that we don't yet really have a good picture on what the skills needs will be on the other side of AI development to the extent that you can even think about it as the other side of AI development. A lot of what Emily is proposing is just a better AI-ified version of the same education we have now.

But I tend to think that what we need to train people for is going to look so wildly different than it did in the past that this might end up feeling like a very intermediary step, and I wonder if there will be enough momentum, given how intermediary we are, to make a lot of these big changes knowing they might just change again a couple years down the line.

I don't know how good this analogy is, but I see it in somewhat of a similar light to current AI process mapping tools, where an AI sits and watches how a human does something so that an agent can do that same thing. On the one hand, this is intuitive and makes sense. On the other hand, in what universe do we think that an agent is going to just do things the same way that a human did just a little bit faster.

There are going to be totally new agent-native processes that don't frankly make sense to humans but still get the outputs that we're looking for. I kind of think that this AI Native University feels like a similar intermediary step. But then again, maybe this is an intermediary journey that's 10 to 20 years, and we just need to make a bunch of these changes now, even before we can know what the full future holds. Next up, let's talk about ChatGBTBT becoming the AI App Store, a prediction from Anish Ashchariah. Anish writes,

Consumer product cycles require three things to work. New technology, new consumer behavior, and a new distribution channel. Until recently, the AI wave had fulfilled the first two conditions, but had no new native distribution channel. Most products grew off the back of existing networks like X or by word of mouth. With the recent release of OpenAI's apps SDK, Apple's support for mini-apps and ChatGPT's rollout of group messaging, though, consumer developers can now tap ChatchipT's 900 million user audience directly

and also grow with new networks of mini-apps like Wabi. As the final piece of the consumer product cycle, this new distribution channel is set to kick off a once-at-a-decade goldbrush in consumer tech in 2026. Ignore at your own peril. Boy, fortunes are going to be made and lost on what venture capitalists think about this. I, for my part, despite not really being a VC, am in the more skeptical category.

I'm putting likelihood of two, because while I think that there will be some value to the distribution that Chatsubit can provide for apps, my instinct is that it looks a lot closer to a new version of SEO and or a new channel for ads as opposed to an app store where people are actively looking for things. It is powerful

that people go to chat CBT with extremely high intent and are looking for answers to their problems or particular types of information that does create an opportunity to serve them highly targeted recommendations which for the moment of course chat chabit is not calling ads that point people to apps that might be useful for them. Basically I think that it is a good distribution channel

but I don't agree that it is a once-in-a-decade channel. However, like I said, I could be very wrong about this one, and the cost-benefit analysis for app developers on at least trying to use this new channel is probably pretty high. Next up, a prediction from Olivia Moore called Voice Agents Take Up Space. Olivia writes, In the last 18 months, the idea of AI voice agents managing real interactions for businesses has gone from science fiction to reality.

Thousands of companies from SMBs to enterprises are using Voice AI to schedule appointments, complete bookings, run surveys, do intakes, and much more. These agents save costs for businesses, generate additional revenue, and free up human employees to do higher leverage and more enjoyable tasks. But because the space is so nascent, many companies are still in the voice as a wedge phase, offering one or several types of calls as a point solution. I'm excited to see voice agents expand into handling entire workflows,

and even into managing full customer relationship cycles. I'm almost going to take this one under direction. That's a little bit different than Olivia, or more expansive. I think the likelihood of this is extremely high, I give it a four, and I think that the value and my X factor are also pretty high. I think voice as the modality of interaction is still wildly undertapped. And you can tell it's wildly undertapped because we're still doing workarounds like using whisperflow instead of the native voice-the-text solutions that our devices offer right now, which are still unbelievably bad. I think that there is so much rich territory to explore here,

and people are going to get really, really used to talking to their phones and their computers in a way that they don't currently. Next up, back to the enterprise from Seema Amble, AI creates a new orchestration layer and new roles in the Fortune 500. In 26, enterprises will shift further from isolated AI tools to multi-agent systems that will need to behave like coordinated digital teams. As agents start to manage complex interdependent workflows, organizations will need to rethink how work is structured and how context flows across systems. The Fortune 500 will feel this shift most acutely. The transition will force leaders to reimagine roles in software. The rise of multi-agent systems isn't just another step in automation. It represents a restructuring of how,

enterprises operate, how decisions are made, and ultimately where value is created. Look, this one is pretty easy for me. This is a yep, yep, and a yep. I think it's very likely. I'll give it a three because it's hard, even though if we expand out over the next two to three years, it's definitely a five. The value I'll also give a three, but also again, only because it's going to be hard. And the X factor, I'll give a three on this one, because even though it's a corporate thing, some of the biggest opportunities for an increase in the satisfaction that we have at work come from these new roles in redesigning how we interact in a big way. For the sake of being a little bit more contentious, let's look at Mark Andrusco's

prompt-free and proactive applications arrive. Mark says, 2026 marks the death of the prompt box for mainstream users. The next wave of AI apps will have zero visible prompting. They'll observe what you are doing and intervene proactively with actions for you to review. Your IDE suggests to refactor before you ask. Your CRM drafts the follow-up email when you finish a call. Your design tool generates variations as you work. The chat interface was training wheels.

Now AI becomes invisible scaffolding woven throughout every workflow. activated by intent rather than instruction. I'm given this a one, one, and a one. I don't think it's likely, I don't think it's valuable, and I just don't really like it in practice. Now, I understand why so many people think that there are big limitations with the prompt box is the only way that we interface with AI.

And certainly I think that there are going to be interface evolutions and changes, but we keep having this discussion and people keep trying things that are different only to come back to the chat interface is a really good default option. Also, at the risk of being biased on early versions, I don't know that I've ever disliked a few,

feature as much as I dislike Gemini's interaction with Gmail right now, where it by default puts in a response to an email, and I have to click around to get out of that. Basically, this might be me being a boomer curmudgeon, but I think that a lot of these suggestions, at least at this stage, are a hell of a lot closer to products trying to convince you that they're valuable than they are actually being useful. And honestly, my suspicion is that that's not because of some big capabilities gap. I think that this idea of being activated by intent that Mark is talking about is incredibly, incredibly difficult to do well. Intent is such a subtle and multivarius thing. And when it comes to proactive suggestions,

even if the intent that it's guessing at is a little bit off, that makes the whole thing completely useless. Now, I could be wrong and I probably need to soften a little bit on this. For example, I am noticing that the vibe coding tools are getting a lot better at suggesting the next thing to do. And so maybe there's more discrete space for this than I'm giving it credit. But yeah, in general, I think that the death of the prompt box is wildly exaggerated. Lastly, two that I'm going to combine that get my most rah-rah, hell yeah rating. The first is building the AI-native industrial base. David Ulovich writes,

America is rebuilding the parts of the economy that create real strength. Energy manufacturing, logistics, and infrastructure are back in focus, and the most important shift is the rise of an industrial base that is truly AI-native and software first. This is opening major opportunities in advanced energy systems, robotics-heavy manufacturing, next-generation mining, and much more. AI can design cleaner reactors, optimize extraction, engineer better enzymes, and coordinated fleets of autonomous vehicles with a level of insight no legacy operator can match. The same shift is reshaping the world outside the factory.

Autonomous sensors, drones, and modern AI models can now give continuous visibility into ports, rail, power lines, pipelines, military bases, data centers, and other critical systems that were once too large to manage comprehensively. Aaron Price Wright adds to this with her prediction of the Renaissance of the American factory. Aaron writes, America's first great century was built on industrial strength, but it's no secret that we've lost much of that muscle. Some of it due to offshoring, some of it due to an intentional society-wide failure to build. But the rusty wheels are starting to creak into motion again, and we're witnessing the rebirth of the American factory with software N.A.I. By applying techniques that Henry Ford developed a century ago, planning for scale and repeatability on day zero, and layering in the latest advances in AI,

will soon be mass-producing nuclear reactors, building housing that meets our nation's demand, constructing data centers at breakneck speed, entering a new golden age of industrial strength. Like I said, these get a rah-rah, hell yeah for me, five's unlikely, five's on valuable, five's on X-factor. I don't think it's that hard to understand why I would think this is super valuable if it happens, but I want to talk about likelihood. Here's why I give this a five, maybe even a six on likelihood. Not only is there immense financial incentive and immense demand and need for this, it is also, I believe, when done well, the best answer to the political acrimony that AI is going to face in 2026. I've said before, and I'm quite sure that 26 is going to be a rough year when it comes to AI politics.

A lot of politicians who are trying to get elected or reelected in the midterms are going to use anti-AI positions as a populist cudgel. It is insane to me that we are and need to build all of this crazy infrastructure and that that is turning into a liability rather than an asset. Why the companies who are building that infrastructure are not doing more to get the communities where it's happening excited and involved and upskilled and more prosperous because of, of that building is just mind-blowing to me.

But I think, optimistically, that we'll stop screwing that up in 2026, which is why I think all of this is so likely. So, friends, there are a lot more big ideas that I didn't even get to, but those are a sampling, those are my power rankings. Hope you enjoyed this big think episode of the AI Daily Brief. Like I said, we've got a little bit more in terms of normal episodes before we fully settle into end-of-the-year content. For now, that is going to do it for today's episode.

Appreciate you listening or watching as always. Until next time, peace.