Connect with us

Technology

The AI spending frenzy is just getting started

Published

on

The AI spending frenzy is just getting started

“All I know is I’m good for my $80 billion.”

Rarely does a one-liner so perfectly capture the state of the moment. Here, you have Microsoft CEO Satya Nadella saying he’s “not in the details” about Stargate, the supposedly multi-hundred-billion AI infrastructure project driven by his marquee investment, OpenAI.

Nadella not being read in on the nebulous details of Stargate says a lot about how much Microsoft and OpenAI have drifted apart. Microsoft is mentioned in the Stargate press release since OpenAI’s models are still exclusive to Azure. But the most striking aspect of Stargate is not that the money isn’t there for it yet; it’s that OpenAI’s biggest backer has decided to not participate in what Sam Altman is calling “the most important project of this era.” As Nadella made clear on CNBC this week, he’s running his own, $80 billion AI infrastructure buildout and, going forward, OpenAI can get additional compute — with his blessing — elsewhere. 

While it received fewer headlines this week, I found Nadella’s response to Elon Musk on X even more illuminating. In his response to Musk saying, “on the other hand, Satya definitely does have the money,” Nadella responded: “😂 And all this money is not about hyping AI, but is about building useful things for the real world!” 

That post can only be interpreted as a dig at Altman. Nadella could have funded Stargate for OpenAI. He didn’t. What does he know that the rest of us don’t?

Advertisement

The splashy Stargate unveiling at the White House certainly accomplished its goal, which was clearly getting everyone to talk about big numbers. The headlines it generated prompted Mark Zuckerberg to make sure everyone ended the week knowing his data center will be even bigger than Stargate.

In a Friday post on his Facebook page, Zuckerberg said that Meta’s planned 2GW data center in Louisiana “is so large it would cover a significant part of Manhattan,” with a map view of the square footage overlaid on the city to send the point home. 

From his post (my emphasis added): “We’ll bring online ~1GW of compute in ‘25 and we’ll end the year with more than 1.3 million GPUs. We’re planning to invest $60-65B in capex this year while also growing our AI teams significantly, and we have the capital to continue investing in the years ahead.”

I have no doubt that Altman, Masayoshi Son, and Larry Ellison will be able to raise the billions they need to lessen OpenAI’s dependence on Microsoft for compute. (The US government isn’t giving money to Stargate, which makes the optics of announcing it alongside Trump all the more bizarre.) Ultimately, this all points to the theme that is quickly coming to define 2025: Big Tech sees AI as the most existential technology of the coming era and will keep spending like hell to make sure OpenAI doesn’t completely run away with it.

Steve Huffman,
Illustration by William Joel / The Verge | Photo by Greg Doherty/Variety via Getty Images
Advertisement

AMA with spez

Few companies had as good of a 2024 as Reddit. Since going public last March, the company’s stock has soared 300 percent, giving the social network a valuation of $32 billion.

It’s an about-face from where Reddit was before going public, when its moderators were raging against its hurried platform changes and there was backlash to the company selling its data to Google and OpenAI.

With those controversies now seemingly in the rear-view mirror, Reddit is focused on growing its user base, staying profitable, and using AI to help people search its site more easily. I caught up with CEO Steve Huffman at CES a few weeks ago to hear his priorities for 2025, how he’s leading Reddit, his thoughts on the AI scaling debate, content moderation, and more…

The following interview has been edited for length and clarity:

Advertisement

Your IPO did very well. What have the last nine months or so been like for you personally?

We have a saying at Reddit that good numbers make good meetings. So we’ve had some good meetings.

Preparing to go public was intense. It’s telling the story over and over and over, which I enjoy doing, but it’s a lot of work. I think more than most new companies, we are in the public company rhythm already: close the quarter, do the audits, do the board meeting, earnings, and all of that. So it hasn’t been a major change for us from an operating point of view.

It’s a really exciting time for the new investors and employees. You won’t catch us complaining. What I keep telling the company is that everyone should be very proud of the work they’ve done and don’t take these moments for granted. I just tell them, look, enjoy the view. If you look at our history, there are lots of ups and downs. No doubt there are challenges in our future.

With your market cap where it is now, are you thinking of making swings you didn’t think you could make a year ago?

Advertisement

There are two classes of things that we would do. One is to execute the core strategy. We’ve got to hire. We’ve got to build. I think we’re very reasonable in terms of our investment size. The one sentence strategy for us is to grow the product and stay profitable. 

What can you do with a high stock price? Maybe you can look at M&A that you wouldn’t otherwise. I’d say that’s not really our orientation right now because the acquisitions we’ve done over the last two years have been these 25-to-50-million-dollar deals. It’s kind of a sweet spot for us to get tech and teams. I’d say we’re always watching the market, but we’re not pursuing anything big or crazy right now because I like the core strategy. I think we can do what we want to do within our current capabilities.

What’s the main product focus for Reddit this year? 

The first is the core of Reddit, which is community conversations. Everyone has a home on Reddit, but do you see that home in your first session? There’s a whole other dimension to our work, which is Reddit as an information source. Reddit has all of this incredible information. For the users who have a question that needs an answer, can we give them that answer? We just got into testing Reddit Answers. I’m finding that really helpful for searches about current events. A year from now, it’s a monetization product. It’s one of the few products where it kind of scratches every itch, so it’ll be a big focus. 

What do you make of this debate about whether the AI industry has run out of data?

Advertisement

I think we’d have a different answer to that question literally every month. We want to have good relationships with other people in this space. We’re open for business.

At the same time, we want to maximize the value we get out of our own data. We have not experienced conflict between the two at this point. I love the [data licensing] relationships we have — the major ones being Google and OpenAI. At this point, we don’t need to make any particular partnership. I’d say they’re all nice to have but nothing is existential for us. 

One of the challenges is that the AI companies don’t know what product they’re building. It’s not a bad thing. They are iterating themselves. ChatGPT itself, the central product in this conversation, was a demo. Then, a year later, it’s the most important piece of enterprise technology on Earth with questionable economics. That makes it very exciting. I don’t think any of these companies would be offended to hear me say that. 

You were one of the first social media CEOs I saw to be very critical of TikTok. How does a US ban affect Reddit?

If you look at Reddit’s traffic graph over the last 19 years, you will not see the rise and fall of any particular platform. I think every content type should work on Reddit. Video on Reddit is largely camera-out — what I’m looking at — as opposed to camera-in, or who am I? That’s social media. I think the ban is the right thing to do for reasons I’ve mentioned that honestly have nothing to do with competition. 

Advertisement

With Meta’s moderation changes, the broader conversation around social media feels like it’s changing right now. 

For the last 10 years, people have been talking about whether speech is the problem,  which is a crazy thought. You can’t have freedom without speech. I think that detour through questioning and relitigating core values of America, hopefully that era is coming to a close.

Are people playing politics? Of course, people always are. On the topic of moderation, we always just try to do things the right way, which, not coincidentally, are aligned with American values. It’s a Democratic platform. We believe very much in the power of people and the wisdom of crowds and voting processes. That is Reddit. So I’m glad to see a return to where we have been most of my life, which is an appreciation for free speech. 

Elsewhere

  • Competitors pounce on TikTok: With TikTok no longer available in US app stores and its in-app functionality technically constrained, everyone is doing their damndest to take advantage of the situation. Meta pre-announced its Capcut competitor and is trying to lure creators away with cash. Substack, Bluesky, and X are all making moves to encourage more video consumption. Meanwhile, President Trump says he’s OK with Elon Musk or Larry Ellison buying it. ByteDance is saying it wants to do a deal but seems increasingly backed into a corner. Ellison may have the guarantee that Oracle won’t be fined out of existence for violating the law right now, but Apple and Google have shown they are going to follow the letter of the law. With TikTok still not available to download in the US, its competitive threat to Meta, YouTube, and others decreases every day.
  • Trump gets to work for Big Tech: Why are Zuckerberg and other CEOs bending the knee? Look no farther than the comments the president made at the World Economic Forum this week, where he trashed the EU’s Digital Markets and Services Acts as a form of “taxation.” This kind of push back is exactly what Meta and other US companies have been praying for. We’ll see if it works for them.
  • More headlines: OpenAI released its AI agent called “Operator” for pro-tier subscribers… Musk told X employees that “user growth is stagnant, revenue is unimpressive, and we’re barely breaking even”…. Apple reorged again as it plays catch-up in AIGoogle is putting another $1 billion into Google Cloud via Anthropic and acquired part of HTC’s Vive team to beef up its Android XR efforts (yes, get ready for the return of Glass)… Epic Games gave an update on its push to compete with Roblox… Meta made a rare investment in Databricks.

More links

  • What led to the DOGE falling out between Vivek Ramaswamy and Elon Musk.
  • A profile of DeepSeek, the Chinese firm that has a bunch of CEOs worried about how much they’re spending on models.
  • Dan Shipper’s hands-on experience using OpenAI’s Operator agent.
  • The “Humanity’s Last Exam” AI dataset.
  • Brian Armstrong’s takeaways from Davos.
  • A whistleblower is claiming Amazon’s $400 million deal for most of Covariant AI was a “reverse acquihire” designed to avoid antitrust scrutiny.
  • Nvidia is the top tech company in Glassdoor’s latest list of the top places to work.
  • The rise of the MAGA-bro podcast.

If you haven’t already, don’t forget to subscribe to The Verge, which includes unlimited access to Command Line, all of our reporting, and an improved ad experience on the web.

As always, I want to hear from you, especially if your data center is even bigger. Respond here, and I’ll get back to you, or ping me securely on Signal.

Advertisement
Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Technology

BEWARE SOFTWARE BRAIN

Published

on

BEWARE SOFTWARE BRAIN

Today on Decoder, I want to lay out an idea that’s been banging around my head for weeks now as we’ve been reporting on AI and having conversations here on this show. I’ve been calling it software brain, and it’s a particular way of seeing the world that fits everything into algorithms, databases and loops — software.

Software brain is powerful stuff. It’s a way of thinking that basically created our modern world. Marc Andreessen, the literal embodiment of software brain, called it in 2011 when he wrote the piece “Why software is eating the world” as an op-ed in The Wall Street Journal. But software thinking has been turbocharged by AI in a way that I think helps explain the enormous gap between how excited the tech industry is about the technology and how regular people are growing to dislike it more and more over time.

In fact, the polling on this is so strong, I think it’s fair to say that a lot of people hate AI. And Gen Z in particular seems to hate AI more and more as they encounter it. There’s that NBC News poll showing AI with worse favorability than ICE and only a little bit above the war in Iran and the Democrats generally. That’s with nearly two thirds of respondents saying they used ChatGPT or Copilot in the last month. Quinnipiac just found that over half of Americans think AI will do more harm than good, while more than 80 percent of people were either very concerned or somewhat concerned about the technology. Only 35 percent of people were excited about it.

Poll after poll shows that Gen Z uses AI the most and has the most negative feelings about it. A recent Gallup poll found that only 18 percent of Gen Z was hopeful about AI, down from an already-bad 27 percent last year. At the same time, anger is growing: 31 percent of those Gen Z respondents said they feel angry about AI, up from 22 percent last year.

Now, I obviously talk to a lot of tech executives and policy people here on Decoder, and I will tell you, they all know AI isn’t popular, and they can all see how that’s playing out in real life. Here’s Microsoft CEO Satya Nadella talking about how the tech industry needs to make the case for the investments it’s making in AI:

Advertisement

Satya Nadella: At the end of the day, I think this industry, to which I belong, needs to earn the social permission to consume energy because we’re doing good in the world.

I think it’s safe to say that the tech industry and AI have not earned any of that social permission yet. Politicians from both sides of the aisle are opposing data center buildouts. Politicians in local communities that support data centers are getting voted out of office. And in the most depressing reminder of how much political violence has become a part of everyday American life, politicians who’ve supported data centers have had their houses shot at. OpenAI CEO Sam Altman has had Molotov cocktails thrown at his house.

It’s sad that I’m going to have to say this again on the show, and it’s sad that we’re going to have commenters who disagree, but this violence is unacceptable. If you want to meaningfully oppose AI in a way that lasts, you should speak loudly with your dollars in the market and your attention online, and you should speak loudly with your votes. You should participate in a democratic regulatory and political process. Anything else will get dismissed and perpetuate the cycle. That dismissal is already happening.

I also think it’s incredibly important for our politicians and tech executives to make sure our political process makes people feel empowered, not helpless, which is a specific kind of nihilism they have all greatly contributed to. The violence is a result of that helplessness and nihilism. And the most powerful people in our society ought to reckon with that, especially as they run around saying AI will wipe out all the jobs. I’m not even exaggerating this. Here’s Anthropic CEO Dario Amodei saying he thinks AI will wipe out all the jobs:

Dario Amodei: Entry-level jobs in areas like finance, consulting, tech and many other areas like that —- entry-level white-collar work — I worry that those things are going to be first augmented, but before long replaced by AI systems. We may indeed —- it’s hard to predict the future — but we may indeed have a serious employment crisis on our hands as the pipeline for this early-stage, white-collar work starts to contract and dry up.

What I see when I encounter clips like this is the true gap between the tech industry and regular people when it comes to AI — and also the limit of software brain. Like I said, everyone in tech understands how much regular people dislike AI. What I think they’re missing is why. They think this is a marketing problem. OpenAI just spent $200 million on the TBPN podcast because the company thinks it will help make people like AI more. Sam Altman has said so explicitly:

Sam Altman: Oh, they are genius marketers and I would love to have better marketing. Somebody said to me recently that if AI were a political candidate, it would be the least popular political candidate in history. And given the amazing things AI can do, I think there’s got to be better marketing for AI.

It feels like someone just needs to say this clearly, so I’m just going to do it. AI doesn’t have a marketing problem. People experience these tools every single day. ChatGPT has 900 million weekly users, trending to a billion, and everyone has seen AI Overviews in Google Search and massive amounts of slop on their feeds. You can’t advertise people out of reacting to their own experiences. This is a fundamental disconnect between how tech people with software brains see the world and how regular people are living their lives.

Advertisement

Image: The Verge

So what is software brain? The simplest definition I’ve come up with is that it’s when you see the whole world as a series of databases that can be controlled with structured language and software code. Like I said, this is a powerful way of seeing things. So much of our lives run through databases, and a bunch of important companies have been built around maintaining those databases and providing access to them.

Zillow is a database of houses. Uber is a database of cars and riders. YouTube is a database of videos. The Verge’s website is a database of stories. You can go on and on and on. Once you start seeing the world as a bunch of databases, it’s a small jump to feeling like you can control everything if you can just control the data.

But that doesn’t always work. Here’s an example: Elon Musk and DOGE showed up in the government, and the first thing they did was take control of a bunch of databases. And they ran into the undeniable fact that the databases aren’t reality, and DOGE ended in hilarious failure. It turns out software brain has a limit, and the government isn’t software. People aren’t computers, and they don’t live in automatable loops that can be neatly captured in databases.

Anyone who’s actually ever run a database knows this. At some point, the database stops matching reality. And at that point, we usually end up tweaking the database, not the world. The AI industry has fully lost sight of this. AI thrives on data. It’s just software. And so the ask is for more and more of us to conform our lives to the database, not the other way around.

Advertisement

Let me offer you another example that I think about all the time, especially as AI finds real fit as a business tool. It’s the idea that AI is coming for lawyers and the legal system. The AI industry loves to talk about not needing lawyers anymore, which is already getting all kinds of people into all kinds of trouble. But I get it. I’ve spent a lot of time with lawyers. I used to be a lawyer. My wife is still a lawyer. Some of my best friends are lawyers.

Verge subscribers, don’t forget you get exclusive access to ad-free Decoder wherever you get your podcasts. Head here. Not a subscriber? You can sign up here.

I also spend all of my time at work talking to tech people. And so over time, I’ve learned that the overlap between software brain and lawyer brain is very, very deep. Alluringly deep. If the heart of software brain is the idea that thinking in the structured language of code can make things happen in the real world, well, the heart of lawyer brain is that thinking in the structured legal language of statutes and citations can also make things happen. Hell, it can give you power over society.

There are other commonalities. Both software development and the law depend heavily on precedent. We have a body of case law in this country, and we use it over and over again to help us resolve disputes. Much like software engineers have libraries of code that they turn to repeatedly to build the foundations of their products. I can go on.

At the end of the day, both lawyers and engineers do their best to use formal, structured language to guide the behavior of complicated systems in predictable and potentially profitable ways. I am far from the first person with this idea. Larry Lessig wrote a book called Code and Other Laws of Cyberspace in 2000. It’s just as relevant today as it was a quarter century ago.

And so you have this intoxicating similarity between law and code, and it trips people up all the time. People are constantly trying to issue commands to society at large like it’s a computer that will obey instructions. There are examples of this big and small. My favorite are those Facebook forwards insisting Mark Zuckerberg does not have the right to publish people’s photos. Honestly, I look at these, and I think it would be great if the law was actually code. Maybe things would be more predictable. Maybe we’d feel more in control.

Advertisement

But law isn’t actually code, and society and courts aren’t computers. I have to remind our fairly technical audience on Decoder and at The Verge all the time that the law is not deterministic. You simply cannot take the facts of a case, the law as written, and predict the outcome of that case with any real certainty, even though the formality of the legal system makes people think it works like a computer, that it’s predictable.

Because at the end of the day, it’s actually ambiguity that’s at the very heart of our legal system. It’s ambiguity that makes lawyers lawyers. Honestly, it’s ambiguity that makes people hate lawyers because it’s always possible to argue the other side, and it’s always possible to find the gray area in the law. That’s why prosecutors end up working as defense attorneys and why our regulators tend to end up working for big corporations.

So you can see the obvious collision between software brain and lawyer brain. This thing that looks like a computer isn’t actually anything at all like a computer. A lot of people even argue that the law should be more like a computer, that the system should be verifiable and consistent, and that merely issuing the right commands at the right times should lead to objectively correct outcomes.

Bridget McCormack, who used to be the chief justice of the Michigan Supreme Court, was on Decoder a few months ago pitching a fully automated AI arbitration system. Her argument to me was that people perceive the traditional legal system to be so unfair, they will accept a worse outcome from an automated system as more fair as long as they feel heard. And if there’s one thing AI can do, it’s sit there and listen all day and night. I don’t know if any of that is correct or even workable, but I do know software brain, and that is pure software brain. The idea that we can force the real world to act like a computer and then have AI issue that computer instructions.

You can see the same thing happening in every other kind of industry. You don’t hire a big consulting firm to actually come in and study your business and make it more efficient. You hire them to make slide decks that justify layoffs to your board and shareholders. Big consulting firms are great at this, and now they’re just going to generate those decks with AI. They are already doing this and the layoffs have already begun.

Advertisement

Any business process that looks like code talking to a database in a repetitive way is up for grabs. That’s why Anthropic has been so relentlessly focused on enterprise customers, and it’s why OpenAI is now pivoting to business use. There’s real value in introducing AI to business because so much of modern business is already software, collecting data, analyzing it, and taking action on it over and over again in a loop. Businesses also control their data, and they can demand that all their databases work together. In this way, software brain has ruled the business world for a long time. And AI has made it easier than ever for more people to make more software than ever before, for every kind of business to automate big chunks of itself with software. The absolute cutting edge of advertising and marketing is automation with AI. It’s not being in creative.

But not everything is a business, not everything is a loop, and the entire human experience cannot be captured in a database. That’s the limit of software brain. That’s why people hate AI. It flattens them. Regular people don’t see the opportunity to write code as an opportunity at all. The people do not yearn for automation. I’m a full-on smart home sicko; the lights and shades and climate controls of this house are automated in dozens of ways. But huge companies like Apple, Google and Amazon have struggled for over a decade now to make regular people care about smart home automation at all. And they just don’t.

AI isn’t going to fix that. Most people are not collecting data about every single thing that they do. And if they’re collecting any at all, it’s stored across lots of different systems — your email in Gmail, your messages in iMessage, your work schedule in Outlook, your workouts in Peloton. Those systems don’t talk to each other and maybe they never will, because there’s no reason for them to. And asking people to connect them all freaks them out.

Even taking the time to consider how much of your life is captured in databases makes people unhappy. No one wants to be surveilled constantly, and especially not in a way that makes tech companies even more powerful. But getting everything in a database so software can see it is a preoccupation of the AI industry. It’s why all the meeting systems have AI note takers in them now. It’s why Canva, which is design software, now connects to corporate email systems. My friend Ezra Klein just went to Silicon Valley, and he described the people that are actively trying to flatten themselves into a database:

Ezra Klein: You might think that A.I. types in Silicon Valley, flush with cash, are on top of the world right now. I found them notably insecure. They think the A.I. age has arrived and its winners and losers will be determined, in part, by speed of adoption. The argument is simple enough: The advantages of working atop an army of A.I. assistants and coders will compound over time, and to begin that process now is to launch yourself far ahead of your competition later. And so they are racing one another to fully integrate A.I. into their lives and into their companies. But that doesn’t just mean using A.I. It means making themselves legible to the A.I.

You can give it access to everything that’s there: your files, your email, your calendar, your messages. It operates continuously in the background, building a persistent memory of your preferences and patterns so it can better act on your behalf. The cybersecurity risks are glaring, but there’s a reason millions of people are using it: The more of your life you open to A.I., the more valuable the A.I. becomes.

Advertisement

I’ve reviewed a lot of tech products over the past decade and a half, and all I can tell you is that it is a failure when you ask people to adapt to computers. Computers should adapt to people. And asking people to make themselves more legible to software, to turn themselves into a database, is a doomed idea. It’s an ask so big, I can’t imagine a reward that would make it worth it for anyone, even if the tech industry wasn’t constantly talking about how AI will eliminate all the jobs, require a wholesale rethinking of the social contract and — oops — also the latest models might cause catastrophic cybersecurity problems that might lead to the end of the world.

Does this sound like a good deal to you? Can you market your way out of this? This only makes sense if you have software brain, if your operative framework is to flatten everything into databases that you can control with structured language. The people paying thousands of dollars a month to set up swarms of OpenClaw agents and write thousands of lines of code, they’re people who look at the world and see opportunities for automation, to repeat tasks, to collect data, to build software. AI is great for them. It’s even exciting in ways that I think are important and will probably change our relationship to computers forever.

For everyone else, AI is just a demanding slop monster. It’s a threat. I’m not saying regular people don’t use Excel or Airtable to plan their weddings or have fun throwing PowerPoint parties, or even that AI won’t be useful to regular people over time. I think a lot of people enjoy data and tracking different parts of their lives. There’s my WHOOP band. I’m just saying these things aren’t everything. Not everything about our lives can be measured and automated and optimized. It shouldn’t be.

And so the tech industry is rushing forward to put AI everywhere at enormous cost — energy, emissions, manufacturing capacity, the ability to buy RAM — and locked into the narrow framework of software brain without realizing they are also asking people to be fundamentally less human. They then sit around wondering why everyone hates them. I don’t think a couple haircuts are going to fix it.

Questions or comments about this episode? Hit us up at decoder@theverge.com. We really do read every email!

Advertisement

Decoder with Nilay Patel

A podcast from The Verge about big ideas and other problems.

SUBSCRIBE NOW!

Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

Continue Reading

Technology

Toyota’s CUE7 robot shoots hoops using AI

Published

on

Toyota’s CUE7 robot shoots hoops using AI

NEWYou can now listen to Fox News articles!

Most people think of Toyota and picture a Camry, a Tacoma, maybe a Prius. A 7-foot-2 robot shooting free throws at halftime of a professional basketball game? That’s a harder image to conjure. But recently, that’s exactly what happened at Toyota Arena Tokyo, and around 8,400 fans watched it go down live.

The robot is called the CUE7. It smoothly stood up from a seated position, dribbled a basketball and sank a free throw without any human input. The crowd applauded. The engineers probably exhaled. Toyota had officially debuted its most advanced AI-powered humanoid robot, and it chose basketball as the venue.

So why is a car company building basketball robots? And what does any of this have to do with you? More than you might think.

Sign up for my FREE CyberGuy Report

Advertisement
  • Get my best tech tips, urgent security alerts and exclusive deals delivered straight to your inbox.
  • For simple, real-world ways to spot scams early and stay protected, visit CyberGuy.com – trusted by millions who watch CyberGuy on TV daily.
  • Plus, you’ll get instant access to my Ultimate Scam Survival Guide free when you join.

AI-POWERED ROBOT SINKS SEEMINGLY IMPOSSIBLE BASKETBALL HOOPS

Toyota’s CUE7 robot handles the ball with precision, showing how AI can learn complex physical movement. (Toyota Motor Corporation)

The CUE7 started from scratch, on purpose

Here’s the thing that makes the CUE7 genuinely different from its predecessors: Toyota’s team discarded everything they had built and started over.

“We made full use of AI, and we discarded everything we had built up and started again from scratch,” said Tomohiro Nomi, research leader for humanoid robots at Toyota’s Frontier Research Center.

That’s not a small statement. The CUE series goes back to 2017, when a group of Toyota employees launched it as a voluntary side project on their own time. It eventually became an official research program, and over nearly a decade, the team stacked up some genuinely impressive hardware. The CUE3 earned a Guinness World Record in 2019 for most consecutive basketball free throws by a humanoid robot (assisted), sinking 2,020 in a row. Then the CUE6 earned the record for the farthest basketball shot by a robot, connecting from about 80 feet 6 inches) away.

So the legacy was already there. What changed with CUE7 was the philosophy behind how it learns.

Advertisement

From human programming to AI that figures it out alone

Earlier versions of the CUE relied on something called model predictive control. Basically, human engineers programmed exactly how the robot should move, step by step. It worked well enough to break world records. But it also had a ceiling. Every new motion required new programming by a human being.

The CUE7 instead uses reinforcement learning powered by artificial intelligence. It learns to shoot the ball based on its own experience and trial and error rather than pre-programmed instructions. The AI acts as an autonomous agent: it tries something, observes the result, adjusts and tries again. Over enough repetitions, it gets good. Really good.

The hybrid control system merges reinforcement learning with model predictive control, creating a robot that adapts to unexpected situations rather than just following a fixed script. Think of it as the difference between a player who memorized every play in the book and one who reads the game in real time. CUE7 is learning to read the game.

What’s actually inside the CUE7 robot

The CUE7 stands about 7 feet 2 inches tall and weighs roughly 163 pounds, making it about 40% lighter than the previous version, which came in around 265 pounds. Toyota pulled that off by simplifying the structure and reducing the number of axles.

It also switched from four wheels to two, which makes its movement faster and more fluid. One moment that really stood out was how smoothly it can rise from a seated position. That kind of motion, especially at this size, takes serious engineering and drew a reaction from a crowd of more than 8,000 people.

Advertisement

For sensing and aiming, the robot uses lidar sensors in its torso to detect its surroundings, along with a stereo camera in its head to calculate distance and angle. It is powered by high-performance batteries adapted from Toyota’s racing tech.

Here’s where it gets interesting. The robot measures the distance to the hoop, calculates the angle, determines the right trajectory and then releases the shot with controlled force. If it misses, it learns from that attempt and adjusts on the next one.

ROBOT PLAYS TENNIS WITH HUMANS IN REAL TIME

During a live game demo, the robot lines up a shot, highlighting how machines can adapt in real-world environments. (Toyota Motor Corporation)

The AI that actually makes this work

Toyota trained the system using human motion data, which is what gives CUE7 its surprisingly natural movement. Rather than looking mechanical, its actions mirror how a person actually moves, and that’s by design.

Advertisement

That same combination of real-time calculation and learned experience is what lets it handle something like dribbling (fluid, continuous) alongside shooting (precise, calculated) without the two working against each other.

Toyota says testing that kind of learning in a live environment is a key part of the project.

“We believe it is an exceptionally valuable opportunity to validate a reinforcement-learning-based robot in the inherently uncertain environment of a basketball arena,” Tomohiro Nomi, Head of Humanoid Robotics Research Unit, Frontier Research Center, Toyota Motor Corporation, told CyberGuy. “Moving forward, we will continue developing robots that inspire and bring joy to people.”

What this means to you

You’re probably not buying a robot basketball player anytime soon. But here’s the part worth paying attention to: the same AI that helps CUE7 sink free throws is the technology Toyota is actively developing for manufacturing, automotive systems and real-world robotics.

Basketball demands everything that manufacturing robots struggle with: target identification, distance gauging, trajectory computation, coordinated movement and precise force control, all in sequence and under pressure. Toyota chose basketball specifically because it tests all those capabilities at once, in an environment where success and failure are completely obvious.

Advertisement

The reinforcement learning powering CUE7 could eventually show up in factory robots that adapt mid-shift when production requirements change, in vehicles that handle unexpected road conditions more fluidly, or in home and care robots that need to navigate unpredictable environments. Toyota treats CUE7 as a testbed for vision systems, motion control and coordinated movement, with capabilities that reach well beyond halftime demonstrations into broader real-world applications.

When Toyota teaches a robot to play basketball, it’s really teaching machines how to learn. And that skill transfers. In other words, this is less about basketball and more about teaching machines how to learn physical skills in unpredictable environments. That is where the real impact starts to show up.

THE NEW ROBOT THAT COULD MAKE CHORES A THING OF THE PAST

CUE7 sinks a free throw, a simple moment that reflects a bigger shift toward AI that learns through experience. (Toyota Motor Corporation)

Take my quiz: How safe is your online security?

Advertisement

Think your devices and data are truly protected? Take this quick quiz to see where your digital habits stand. From passwords to Wi-Fi settings, you’ll get a personalized breakdown of what you’re doing right and what needs improvement. Take my Quiz here: Cyberguy.com.

Kurt’s key takeaways 

The CUE7 is a fascinating piece of technology, but the real story isn’t about basketball. It’s about a fundamental shift in how robots are trained, moving away from rigid human programming toward AI systems that learn through experience and adapt on the fly. What started as a voluntary employee side project in 2017 has grown into a genuine proving ground for Toyota’s embodied AI research. Nearly a decade in, the results are landing in front of thousands of live spectators and stacking up Guinness World Records along the way. The CUE7 made a free throw at halftime in front of a packed arena. More importantly, it demonstrated that AI-powered machines can now acquire complex physical skills through trial and error, the same basic way humans do. That’s a shift with implications that reach far beyond the basketball court.

If a robot can teach itself to make free throws better than most humans ever will, purely through AI-driven trial and error, what physical skill do you still believe machines will never be able to learn on their own? Let us know by writing to us at Cyberguy.com.

CLICK HERE TO DOWNLOAD THE FOX NEWS APP

Sign up for my FREE CyberGuy Report

Advertisement
  • Get my best tech tips, urgent security alerts and exclusive deals delivered straight to your inbox.
  • For simple, real-world ways to spot scams early and stay protected, visit CyberGuy.com – trusted by millions who watch CyberGuy on TV daily.
  • Plus, you’ll get instant access to my Ultimate Scam Survival Guide free when you join.

Copyright 2026 CyberGuy.com. All rights reserved.

Continue Reading

Technology

The Iranian women Trump ‘saved’ from execution are simultaneously real and AI-manipulated

Published

on

The Iranian women Trump ‘saved’ from execution are simultaneously real and AI-manipulated

Only the night before, he had posted on Truth Social about the imminent executions of these women, quoting a screenshot that included a collage of eight glamorously backlit, soft-focus portraits. The photos of the women were immediately accused of being AI-generated. “Trump is begging Iranian leaders to not execute 8 AI-generated women. This is the funniest thing I’ve ever seen,” said one viral X post.

On top of that, almost immediately after Trump’s announcement, Mizan, an Iranian state news agency, called the president a liar. “Last night, Donald Trump, citing a completely false news story, called on Iran to overturn the death sentences of eight women.” Mizan said that some of the women had already been released and others were facing prison time but not execution, and furthermore said that Tehran had made no concessions — presumably, the status of the women has not changed.

The X account for the Iranian embassy in South Africa, perhaps the most relentless shitposter among Iran’s state-affiliated accounts, was quick to pile on by generating its own set of eight women:

The collage that Trump posted is, at the very least, AI-modified, Mahsa Alimardani, the associate director of the Technology Threats & Opportunities program at WITNESS, told The Verge. But the women themselves are real. The woman in the top right corner of the collage is Bita Hemmati, whose photograph appeared in several news stories in various right-leaning news outlets last week. Hemmati is confirmed to have received a death sentence issued by Branch 26 of the Tehran Revolutionary Court for “operational action for the hostile government of the United States and hostile groups.”

Alimardani named six of the women (Bita Hemmati, Mahboubeh Shabani, Venus Hossein-Nejad, Golnaz Naraghi, Diana Taherabadi, Ghazal Ghalandri), and said that the identities of the final two (said to be Panah Movahedi and Ensieh Nejati) were still unverified. The six verified women participated in protests against the government in January. Aside from Hemmati, none of the other women are reported to have received death sentences.

Advertisement

It’s not surprising that Trump has a careless disregard for the truth; it’s not surprising, either, for the Iranian regime to fudge the details to suit its own narrative, or to make light of real political prisoners in order to dunk on the United States.

The additional wrinkle is that the account mocking Trump for coming to the rescue of “8 AI-generated women” is the very same one that landed South Korean president Lee Jae-myung in hot water when he quoted a misleading labeled video posted by that account. Israeli officials have accused the account of being “well-known for spreading disinformation.” The case of the sketchy Lee Jae-myung quote-post is a story of mingled truth and misinformation, where the post got facts very wrong, but the video — of Israeli Defense Forces soldiers shoving a limp body off a rooftop in Gaza — was real, documenting an event that possibly implicates Israeli forces in a violation of international law.

The case of the eight Iranian protesters also features that same mingling of fact and fiction into a fuzzy distortion that fuels an endless disputation of real human rights violations. Their lives have been reduced to glossy pixels and quote-dunks, the stuff of propaganda and parody. While known liars fight with each other on the internet about who these women are and what will happen to them, they — verifiably six of them, at least — remain real people who exist beyond the Iranian internet blackout.

Continue Reading
Advertisement

Trending