Thoughts on Tech News of Note - 03-13-2026
- Meta Acquihires Moltbook
- Death of the Status Upgrade
- The World Models are Coming
Meta Acquihires Moltbook
There is a paragraph in the TechCrunch article on this story that caught my attention:
"For an agentic web where businesses’ agents and consumers’ agents can work together, though, the agents first need to be able to find each other, connect, and coordinate their activities. As Facebook once built the “friend graph” — a network defined by social connections between people, where every individual is a node — an agentic web could benefit from an “agent graph,” a system that maps out how various agents are connected and what actions they can take on each other’s behalf. This could span areas like travel, online shopping, media and research, productivity tools, and more."
Now, I just watched a video on the infamous book "If Anyone Builds It, Everyone Will Die: The Case Against Superintelligent AI", so my thinking is perhaps a bit poisoned right now. Nevertheless, I'm not sure of the intrinsic value of helping agents find each other and coordinate their activities. If I've installed OpenClaw or other similar agentic AI tool, I want that tool to help me with my tasks. I do not want it gallivanting around the internet with other agents without my knowledge. While there may very well be scenarios where interacting with agents from other companies might benefit me and help me get tasks done more efficiently, I'm not sure my agent needs to hook up with another person's agent to accomplish most of the things I want to get done. As agentic tools become more prevalent, I can imagine a personal agent connecting with the agent that operates on behalf of a company like Shopify to help me order items and track their progress until they reach my door. I can conceive a scenario where my personal agent connects with my bank's agent to provide me with real-time information on transactions processed or the remaining balance in my account and use that information to keep my budget tracker up to date. But unless someone's personal agent is connected to business I need to conduct or I know that person and connecting with their agent helps keep our calendars or communications in sync, I don't see a need to just go seek out other agents without some parameters or guidelines for when it might make sense to do so. It seems more like a recipe for disaster than an enabler for getting stuff done.
I do believe that as companies begin to build out their agentic tools, it will make sense to have a framework or platform that establishes how personal agents connect with corporate agents. Yet Moltbook does not seem to be the tool that would get us to that place. Perhaps Meta is looking to take the idea of Moltbook and scale it up to be a type of Yellow Pages or a Grand Central Station of sorts for all agents, but this would seem to require companies to partner with Meta to make this work and I'm not sure Meta is the company most other companies would voluntarily choose to build out a platform like that. I'm also not sure that a social network paradigm (unless maybe we're talking something like China's WeChat/Weixin) is right for this kind of functionality. I do think that eventually there will need to be some kind of directory and connection system for agentic tools that will help people and corporations conduct business outside of their internal networks. Perhaps Meta will be able to capitalize on this opportunity and get out in front of it, but I suspect an AI company that has invested in creating agentic tools will be better positioned and more trusted to create the infrastructure that will be needed to take these tools to the next level.
Death of the Status Upgrade
Ziff Davis (parent company of CNET and PCMag) released a research study that outlines changes in consumer behavior toward technology upgrades. In the early days of the iPhone, it wasn't uncommon for people to upgrade their smartphones every year. T-Mobile even had a rate plan called "JUMP" that stood for Just Upgrade My Phone. I was enrolled in that plan, and you could actually upgrade every 6 months if you were crazy enough. I took advantage of that opportunity to try out phones I probably wouldn't have otherwise tried because back in those days, we had more options than just Samsung and Apple. I happily rocked the legendary HTC One for at least one of the terms on that rate plan. (BTW, I sincerely miss HTC.) Anyway, phones were changing rapidly enough that the most financially capable and fiscally irresponsible could upgrade every year and see noticeable changes. Hardware designs would change. Cameras would get serious improvements. Screens would get sharper and brighter. The user interfaces would be overhauled, sometimes for the better and sometimes not. But as the years have passed and competitors have died off, our choices in smartphones have narrowed and the companies that have come out as winners in the United States don't need to work as hard to get our attention, so they've slowed down the innovations. Customers have noticed and have slowed down the pace with which they upgrade their devices. The results of the study don't really address why people have slowed down, but it does offer some obvious possibilities such as the current state of the economy, the increasing cost of new devices, and the lack of significant improvement from year to year. It also seems that companies' determination to shove as many features as possible into our devices and label them with "AI" is turning some customers off rather than drawing them in. It may be generational, but a significant percentage of consumers is somewhat resistant to new AI features in devices and hasn't yet grown or learned to see the value in them. Younger buyers may appreciate flashy new AI features more, but even they are seeking more than just AI when they opt to upgrade. Many people value battery life above almost all else and will often hang onto an aging phone until the battery performance has degraded enough that using the phone becomes more of an exercise in finding a charger than anything else. And even then, some companies offer battery replacement programs that can help people stretch a bit more life out of an elderly phone.
The results of this study seem to hold true with the experiences I've had in real life. People who know I'm into tech used to ask me about new devices and whether they are worth it. Now even if people ask, they seem to have already made up their minds in advance that nothing could convince them to upgrade. I do think that right now the state of the economy in the United States is as much of an aggravating factor as the slowing of innovation. Phones and tablets have gotten more expensive with time and even though it used to be that most Americans worried less about the price of a phone because they were financing it over 24-36 months via their mobile carrier plan, people do now seem to be more price sensitive. Life feels a little less secure and the idea of adding $30-50 to a phone plan for an extended period of time may cause more angst than it used to. We are at a place in society where most people believe they need a smartphone, but that doesn't necessarily mean we need a new one. If the old one isn't broken, we aren't often looking for a fix.
I am interested to see how well this pattern holds up when Apple introduces their foldable iPhone. I don't think people line up for new Apple releases anymore like they did many years ago, but I wouldn't be at all surprised if lines form for this new device, depending on how it is priced. I think there is pent-up demand for something new and interesting from the Apple camp and a foldable could unleash some real excitement and get people to open up their pocketbooks even if those pocketbooks are just about empty. BTW, I'm going to predict right now that you should wait until at least version two before you buy an Apple foldable. I'll revisit that in the future when it finally becomes real.
Source: https://www.pcmag.com/articles/the-status-upgrade-is-dead-73-of-people-just-want-tech-that-works
The World Models Are Coming
There is a growing movement that believes that large language models (LLMs) will not be the vehicle by which humanity establishes super intelligent digital beings. The foundational idea is that human intelligence isn't based entirely on language; we are able to reason and make decisions before we even have language to express to others what is going on in our minds. It is our interaction with the world in front of us that helps us form opinions and gather facts. Therefore, anything that will be at least as intelligent as us in the fullest sense will also need to have knowledge of the world in which we live. A different model will be needed to get us to the next level and at the moment it seems we've landed on the term "world model". AI expert Yann LeCun, who once worked as the chief AI scientist at Meta, has started his own company, AMI, to pursue this goal. He has raised over $1B at a $3.5B valuation. The funding round was led by Cathay Innovation (early backer of Chinese e-commerce company Pinduoduo, now listed on the NASDAQ), Greycroft (backer of Venmo, Character.ai, Bumble), Hiro Capital (backer of fitness platform Zwift), HV Capital (backer of HelloFresh), and Jeff Bezos-backed Bezos Expeditions.
The first question I had and to be honest, the question I still have is what is a world model, and perhaps more importantly, how would one build one? Most of the definitions and descriptions I've seen explain that world models map out states and actions. This could be boiled down to an f if/then scenario - if I do X, then what happens next? The data that could be used to help shape the model could include sensor data, global positioning information, video, LIDAR, etc. The idea is for the model to predict the next state like an LLM predicts the next word. Companies that are moving along this path are starting to introduce video into their LLM training data. Companies like Google and Elon Musk's xAI could begin to merge data from their self-driving ventures into their LLM training data to enhance the scope of those models. It would seem that companies that are farther behind in video production and/or don't have easy access to self-driving data might find themselves behind the curve not too far down the line if this is an approach that yields real results and improves the intelligence and capabilities of today's LLMs. It would also seem that establishing world models will be instrumental, if not absolutely necessary to enable the robotics revolution. Companies that have already made strides in robotics may find themselves pairing up with AI companies that don't have other good options for expanding the abilities of their LLMs. On paper, world modeling, whatever it ends up being, seems like the logical next step and even if it doesn't get these companies to where they believe they want to be, it will undoubtedly take them many steps closer. It will be interesting to see what new companies spring up and what mergers and acquisitions are produced by this convergence.
Keep an eye on EV manufacturers, robotics companies, the big chip makers like NVIDIA and AMD, and the prominent AI companies to see where they direct their energies and monies moving forward. BTW, that statement also applies to building out the infrastructure for connecting personal and corporate AI agents as referenced in the first story. All of this is going to start to meld into beautiful convergence or a big mess very soon.
Source: https://techcrunch.com/2026/03/09/yann-lecuns-ami-labs-raises-1-03-billion-to-build-world-models/
