Technology

Inside Mark Zuckerberg’s AI hiring spree

Published

7 months ago

June 14, 2025

Inside Mark Zuckerberg’s AI hiring spree

AI researchers have recently been asking themselves a version of the question, “Is that really Zuck?”

As first reported by Bloomberg, the Meta CEO has been personally asking top AI talent to join his new “superintelligence” AI lab and reboot Llama. His recruiting process typically goes like this: a cold outreach via email or WhatsApp that cites the recruit’s work history and requests a 15-minute chat. Dozens of researchers have gotten these kinds of messages at Google alone.

For those who do agree to hear his pitch (amazingly, not all of them do), Zuckerberg highlights the latitude they’ll have to make risky bets, the scale of Meta’s products, and the money he’s prepared to invest in the infrastructure to support them. He makes clear that this new team will be empowered and sit with him at Meta’s headquarters, where I’m told the desks have already been rearranged for the incoming team.

Most of the headlines so far have focused on the eye-popping compensation packages Zuckerberg is offering, some of which are well into the eight-figure range. As I’ve covered before, hiring the best AI researcher is like hiring a star basketball player: there are very few of them, and you have to pay up. Case in point: Zuckerberg basically just paid 14 Instagrams to hire away Scale AI CEO Alexandr Wang.

It’s easily the most expensive hire of all time, dwarfing the billions that Google spent to rehire Noam Shazeer and his core team from Character.AI (a deal Zuckerberg passed on). “Opportunities of this magnitude often come at a cost,” Wang wrote in his note to employees this week. “In this instance, that cost is my departure.”

Zuckerberg’s recruiting spree is already starting to rattle his competitors. The day before his offer deadline for some senior OpenAI employees, Sam Altman dropped an essay proclaiming that “before anything else, we are a superintelligence research company.” And after Zuckerberg tried to hire DeepMind CTO Koray Kavukcuoglu, he was given a larger SVP title and now reports directly to Google CEO Sundar Pichai.

I expect Wang to have the title of “chief AI officer” at Meta when the new lab is announced. Jack Rae, a principal researcher from DeepMind who has signed on, will lead pre-training. Meta certainly needs a reset. According to my sources, Llama has fallen so far behind that Meta’s product teams have recently discussed using AI models from other companies (although that is highly unlikely to happen). Meta’s internal coding tool for engineers, however, is already using Claude.

While Meta’s existing AI researchers have good reason to be looking over their shoulders, Zuckerberg’s $14.3 billion investment in Scale is making many longtime employees, or Scaliens, quite wealthy. They were popping champagne in the office this morning.

Then, Wang held his last all-hands meeting to say goodbye and cried. He didn’t mention what he would be doing at Meta. I expect his new team will be unveiled within the next few weeks after Zuckerberg gets a critical number of members to officially sign on.

Tim Cook.

Getty Images / The Verge

Apple is accustomed to being on top of the tech industry, and for good reason: the company has enjoyed a nearly unrivaled run of dominance.

After spending time at Apple HQ this week for WWDC, I’m not sure that its leaders appreciate the meteorite that is heading their way. The hubris they display suggests they don’t understand how AI is fundamentally changing how people use and build software.

Heading into the keynote on Monday, everyone knew not to expect the revamped Siri that had been promised the previous year. Apple, to its credit, acknowledged that it dropped the ball there, and it sounds like a large language model rebuild of Siri is very much underway and coming in 2026.

The AI industry moves much faster than Apple’s release schedule, though. By the time Siri is perhaps good enough to keep pace, it will have to contend with the lock-in that OpenAI and others are building through their memory features. Apple and OpenAI are currently partners, but both companies want to ultimately control the interface for interacting with AI, which puts them on a collision course.

Apple’s decision to let developers use its own, on-device foundational models for free in their apps sounds strategically smart, but unfortunately, the models look far from leading. Apple ran its own benchmarks, which aren’t impressive, and has confirmed a measly context window of 4,096 tokens. It’s also saying that the models will be updated alongside its operating systems — a snail’s pace compared to how quickly AI companies move.

I’d be surprised if any serious developers use these Apple models, although I can see them being helpful to indie devs who are just getting started and don’t want to spend on the leading cloud models. I don’t think most people care about the privacy angle that Apple is claiming as a differentiator; they are already sharing their darkest secrets with ChatGPT and other assistants.

Some of the new Apple Intelligence features I demoed this week were impressive, such as live language translation for calls. Mostly, I came away with the impression that the company is heavily leaning on its ChatGPT partnership as a stopgap until Apple Intelligence and Siri are both where they need to be.

AI probably isn’t a near-term risk to Apple’s business. No one has shipped anything close to the contextually aware Siri that was demoed at last year’s WWDC. People will continue to buy Apple hardware for a long time, even after Sam Altman and Jony Ive announce their first AI device for ChatGPT next year. AR glasses aren’t going mainstream anytime soon either, although we can expect to see more eyewear from Meta, Google, and Snap over the coming year.

In aggregate, these AI-powered devices could begin to siphon away engagement from the iPhone, but I don’t see people fully replacing their smartphones for a long time. The bigger question after this week is whether Apple has what it takes to rise to the occasion and culturally reset itself for the AI era.

I would have loved to hear Tim Cook address this issue directly, but the only interview he did for WWDC was a cover story in Variety about the company’s new F1 movie.

AI agents are coming. I recently caught up with Databricks CEO Ali Ghodsi ahead of his company’s annual developer conference this week in San Francisco. Given Databricks’ position, he has a unique, bird’s-eye view of where things are headed for AI. He doesn’t envision a near-term future where AI agents completely automate real-world tasks, but he does predict a wave of startups over the next year that will come close to completing actions in areas such as travel booking. He thinks humans will need (and want) to approve what an agent does before it goes off and completes a task. “We have most of the airplanes flying automated, and we still want pilots in there.”
Buyouts are the new normal at Google. That much is clear after this week’s rollout of the “voluntary exit program” in core engineering, the Search organization, and some other divisions. In his internal memo, Search SVP Nick Fox was clear that management thinks buyouts have been successful in other parts of the company that have tried them. In a separate memo I saw, engineering exec Jen Fitzpatrick called the buyouts an “opportunity to create internal mobility and fresh growth opportunities.” Google appears to be attempting a cultural reset, which will be a challenging task for a company of its size. We’ll see if it can pull it off.
Evan Spiegel wants help with AR glasses. I doubt that his announcement that consumer glasses are coming next year was solely aimed at AR developers. Telegraphing the plan and announcing that Snap has spent $3 billion on hardware to date feels more aimed at potential partners that want to make a bigger glasses play, such as Google. A strategic investment could help insulate Snap from the pain of the stock market. A full acquisition may not be off the table, either. When he was recently asked if he’d be open to a sale, Spiegel didn’t shut it down like he always has, but instead said he’d “consider anything” that helps the company “create the next computing platform.”

If you haven’t already, don’t forget to subscribe to The Verge, which includes unlimited access to Command Line and all of our reporting.

As always, I welcome your feedback, especially if you’re an AI researcher fielding a juicy job offer. You can respond here or ping me securely on Signal.

Click to comment

Technology

The best PS5 games from 2025

Published

44 minutes ago

December 28, 2025

Press Room

In this new era of console… inclusivity? driven by Sony and mostly Microsoft bringing their exclusives to other platforms, it’s getting harder to highlight games that make a singular console shine. Indeed, the top-performing games year after year on both PlayStation and Xbox are multiplatform releases, and often the best games of the year, like 2025 GOTY winner Clair Obscur: Expedition 33, similarly launch on most everything. (Once again, like in all things, Nintendo is the exception.)

But while lists like this will probably disappear as we get deeper into the era of everything being an Xbox, Sony still seems committed to high-quality exclusives. Here are its best of 2025.

Photo by Antonio G. Di Benedetto / The Verge

As software exclusivity goes the way of Sega’s console business, console makers have other ways to attract consumers to their part of the video game ecosystem. For Microsoft, it’s Game Pass; for Sony in 2025, it’s the PlayStation Portal. This year, Sony launched an update for the Portal that enables cloud streaming and making the device a must-have for PlayStation households. For $200, roughly half the price of a PS5, you essentially get a second one, able to play select games without the need to stream them from the console. The list of streamable games isn’t comprehensive, but there are over 2,000 of them, and the ability to play them on the go is a literal game changer.

Get into a discussion with true Final Fantasy sickos about the top 10 games in the franchise, and you better have a lot of time and earplugs on your hands as everyone debates entries two through 10 — because No. 1, invariably, is Final Fantasy Tactics. Having finally played the game for the first time, taking advantage of the remaster Square Enix released this year, I get it. The best video games are often timeless while also able to speak to a specific moment, and Tactics’ story is and does exactly that.

But more than narrative, Tactics’ brilliance is also in its expansive job system that grants players a facsimile of godhood that’s so much fun to wield. The arithmetician class has the ability to cast any learned spell in the game instantly and without requiring mana. To offset the class’s ruinously slow speed, pair it with white mage to create the most diabolical unit in the history of strategy RPGs. Ramza Beoulve is just a guy trying to save the world, and FFT made me believe he can because I made his ragtag band of misfits into walking nuclear weapons.

Sucker Punch’s sequel to Ghost of Tsushima plays things relatively safe. It doesn’t change things up dramatically but instead polishes the experience. Which means the experience isn’t all that different from its predecessor — that means lush landscapes to explore, lots of combat techniques to learn, and, uh, plenty of hot springs to bathe in — but everything feels more finely tuned this time around. That’s especially true of the game’s tale of revenge, which is both more gripping than the original and also gives the open-world experience some much-needed structure.

The original Death Stranding was so weird that half of the fun was just wrapping your head around the experience. I’m not sure I really understood it until I was almost done (if at all). The sequel is still plenty weird, but because it builds off of the previous game, it’s also much easier to understand from the get-go. It’s still a delivery simulator where you trek across a postapocalyptic landscape to get goods and people where they’re needed. But now the setting has shifted to Australia, which means there are new environmental elements to deal with — hello, sandstorms — along with a whole new cast of weirdo characters to meet, played by everyone from Elle Fanning to George Miller.

But don’t worry; the story is still as inscrutable as ever.

Bennett Foddy sure made you feel the “walking” part of this elaborate walking simulator. In Baby Steps, your job is to walk, and you do that by controlling every move involved in taking a step, down to the fine motor details. What results is a hilarious examination of human psychology, emotional maturity, and a newfound appreciation for just how hard it is to ambulate a body with nothing more than DualSense trigger buttons.

When Lumines Arise was first announced, I made the mistake of asking, “Oh, is it a series or something?”, unfamiliar with Lumines’ days as a PSP powerhouse. My elders have never let me live that down, so I thought I’d see for myself what the game is.

Lumines Arise is Evil Dr. Tetris Effect Mean Bean Machine. Mean Bean Machine in that the game is played by globbing together matching blocks in two-by-two configurations. It’s part Tetris Effect because the game has an incredible synth / pop-house soundtrack that’s synced to the action, creating a full-body puzzle-playing experience. And evil in that the game is hard as shit. My skills as a Tetris player, limited though they are, just do not transfer to Lumines Arise. But! There is a flow state I can reach. Throughout playing, the shapes of the blocks will change from blocks, to eggs, to clocks, whatever, and I’ve found some patterns are easier to group than others. It’s interesting and fun to figure out what shapes make the game easier for me and which ones are impossible to configure.

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

Ash Parrish

Technology

ChatGPT’s GPT-5.2 is here, and it feels rushed

Published

1 hour ago

December 28, 2025

Press Room

ChatGPT’s GPT-5.2 is here, and it feels rushed

Can ChatGPT Really Help With Parenting?

Jenny Failla and Madison Alworth join the panel on “Fox News Saturday Night” to share their thoughts on OpenAI CEO Sam Altman admitting he relies on ChatGPT to help with parenting.

NEWYou can now listen to Fox News articles!

OpenAI, the company behind ChatGPT, has moved at an unusually fast pace in 2025. According to the company, it launched GPT-5 in August, followed by GPT-5.1 in November. Now, just weeks later, GPT-5.2 has launched with familiar claims of being the smartest and most capable ChatGPT yet.

At first glance, the rapid rollout might seem surprising. But there’s context behind it. OpenAI CEO Sam Altman has reportedly called a “code red” inside the company, urging teams to move faster on improving ChatGPT. That push comes as competition heats up. Google recently released Gemini 3, which reportedly outperformed ChatGPT on several artificial intelligence benchmarks and delivered stronger image generation. At the same time, Anthropic’s Claude continues to advance quickly.

Against that backdrop, GPT-5.2 feels less like a routine upgrade and more like a strategic response. So what actually changed in GPT-5.2, and why does OpenAI say it matters?

Sign up for my FREE CyberGuy Report
Get my best tech tips, urgent security alerts and exclusive deals delivered straight to your inbox. Plus, you’ll get instant access to my Ultimate Scam Survival Guide – free when you join my CYBERGUY.COM newsletter.

AMAZON ADDS CONTROVERSIAL AI FACIAL RECOGNITION TO RING

OpenAI CEO Sam Altman looks on as he takes a lunch break, during the Federal Reserve’s Integrated Review of the Capital Framework for Large Banks Conference in Washington, D.C., July 22, 2025. (REUTERS/Ken Cedeno)

What exactly is GPT-5.2

GPT-5.2 is the newest version in OpenAI’s flagship 5-series of large language models. Like its predecessor, it includes two default variants. GPT-5.2 Instant is designed for everyday chatting and web searches. GPT-5.2 Thinking is meant for more complex tasks like long reasoning chains and multi-step problem solving. These two models are now the default for all ChatGPT users, including free users. They replace GPT-5.1 Instant and Thinking entirely. If you are using ChatGPT today, you are already using GPT-5.2, whether you realize it or not.

What OpenAI says GPT-5 brings to ChatGPT

At the same time, OpenAI continues to position GPT-5 as “expert intelligence for everyone.” The company says GPT-5 delivers stronger performance across math, science, finance, law and other complex subjects. In OpenAI’s view, ChatGPT now acts more like a team of on-demand experts than a basic chatbot. To support that claim, OpenAI points to practical examples. These include better coding help, more expressive writing support, clearer health-related explanations and improved safety and accuracy. The company showcases use cases such as generating app code, writing speeches, explaining medications and correcting mistakes in user-submitted images. In theory, GPT-5.2 builds on that same foundation. However, while OpenAI emphasizes deeper thinking and more reliable answers, those gains remain subtle for many everyday users.

What new features does GPT-5.2 add?

Here’s the short answer. None. GPT-5.2 does not introduce new tools, interfaces, or headline features. Instead, OpenAI describes a series of behind-the-scenes improvements that supposedly make ChatGPT faster, smarter and more capable. According to OpenAI, GPT-5.2 performs better at:

Building presentations
Completing complex projects
Creating spreadsheets
Understanding long context windows
Interpreting images
Using tools more effectively

Kurt Knutsson reviews the new features in ChatGPT-5.2. (Kurt “CyberGuy” Knutsson)

OpenAI also released new benchmarks showing GPT-5.2 outperforming GPT-5.1 and competing models by small margins. However, big numbers on charts do not always translate into noticeable improvements for real users.

NEW US MILITARY GENAI TOOL ‘CRITICAL FIRST STEP’ IN FUTURE OF WARFARE, SAYS EXPERT

Why testing chatbot improvements is tricky

Evaluating chatbot upgrades is harder than it sounds. Responses can vary widely even when prompts stay the same. A model might excel at one task and struggle with a nearly identical one just moments later. On top of that, OpenAI’s 5-series models already perform at or near the top of the field. When performance starts that high, meaningful gains become harder to detect. With that in mind, we tested GPT-5.2, and in most tests, it behaved almost identically to GPT-5.1.

Why benchmarks don’t tell the full story

OpenAI’s benchmarks show modest gains for GPT-5.2. That matters for researchers and developers working at scale. Still, even advanced users may struggle to see practical benefits. Other companies have delivered clearer upgrades. Google’s Gemini Nano Banana Pro model shows obvious gains in AI image generation and editing. Those improvements are easy for anyone to test and verify. By contrast, GPT-5.2’s changes feel abstract. They exist mostly on paper rather than in daily use.

What this means to you

If you pay for ChatGPT, there’s little downside to using GPT-5.2. It replaces GPT-5.1 in the model lineup and generally performs at least as well in everyday use. Free users don’t have much choice either, as model access is handled automatically. For most people, the experience feels familiar and stable.

The picture shifts slightly for programmers and those who use it for business. Early pricing details suggest GPT-5.2 may cost roughly 40 percent more per million tokens than GPT-5.1, depending on usage tier and access method. That makes testing important before committing at scale.

ChatGPT-5.2 works fine but may not feel exciting, Kurt Knutsson writes. (Michael Nguyen/NurPhoto via Getty Images)

In short, GPT-5.2 works fine. It simply may not feel exciting.

KEVIN O’LEARY WARNS CHINA ‘KICKING OUR HEINIES’ IN AI RACE AS REGULATORY ROADBLOCKS STALL US

Take my quiz: How safe is your online security?

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

GPT-5.2 feels like a model released under pressure rather than inspiration. It performs well, stays reliable, and moves forward in measurable ways. Still, it doesn’t deliver the kind of clear progress many people expect from a new version number. OpenAI remains a leader in AI, but competition is closing in fast. As rivals roll out more noticeable improvements, small updates may no longer be enough to stand out. For now, GPT-5.2 feels less like a breakthrough and more like OpenAI holding its ground.

CLICK HERE TO DOWNLOAD THE FOX NEWS APP

Should AI companies slow down releases until improvements feel more meaningful? Let us know by writing to us at Cyberguy.com.

Kurt “CyberGuy” Knutsson is an award-winning tech journalist who has a deep love of technology, gear and gadgets that make life better with his contributions for Fox News & FOX Business beginning mornings on “FOX & Friends.” Got a tech question? Get Kurt’s free CyberGuy Newsletter, share your voice, a story idea or comment at CyberGuy.com.

Technology

Samsung will debut two new wireless speakers at CES 2026

Published

13 hours ago

December 27, 2025

Press Room

Samsung will debut two new wireless speakers at CES 2026

For years, Samsung has made products that try to camouflage what they are by displaying works of art. The Frame TV is the most famous example, but the company also released the Music Frame, a speaker disguised as a picture frame, at last year’s CES. Now, instead of hiding a speaker with a piece of art, Samsung worked with designer Erwan Bouroullec to make a speaker into a piece of art.

According to Samsung, the Music Studio line of speakers, debuting at CES 2026, draw inspiration from the “timeless dot concept” found throughout music and art. As a musician, the Music Studio 5 reminds me of a fermata, the symbol meant to hold a note or silence. It has a 4-inch woofer and dual tweeters with built-in waveguide.

The larger Music Studio 7 is a 3.1.1-channel speaker that can be used on its own, in a pair for wider stereo sound, or with compatible Samsung Wi-Fi speakers, soundbars, or TVs using the company’s Q-Symphony technology. The Studio 7 is capable of playing high-resolution audio up to 24-bit/96kHz, and it and the Music Studio 5 use AI Dynamic Bass Control to extend bass response.

Sometimes these aesthetic-first speakers forget about the most important part of a speaker — its sound quality. But Samsung has done an impressive job over the past few years with its audio tuning which makes me optimistic for the Music Studio’s performance.