On Wednesday, Nvidia became the first company in history to reach $4 trillion market valuation as shares rose more than 2 percent, reports CNBC. The GPU maker’s stock has climbed 22 percent since the start of 2025, continuing a trend driven by demand for AI hardware following ChatGPT’s late 2022 launch.
The milestone marks the highest market cap ever recorded for a publicly traded company, surpassing Apple’s previous record of $3.8 trillion set in December. Nvidia first crossed $2 trillion in February 2024 and reached $3 trillion just four months later in June. The $4 trillion valuation represents a market capitalization larger than the GDP of most countries.
As we explained in 2023, Nvidia’s continued success has been intimately tied to growth in demand for hardware that runs AI models as capably and efficiently as possible. The company’s data center GPUs excel at performing billions of matrix multiplications necessary to train and run neural networks due to their parallel architecture—hardware architectures that originated as video game graphics accelerators now power the generative AI boom.
AI has become a dirty word across almost every discipline over the past few years. As big corporations keep pushing this technology forward, a vocal resistance among creatives, critics, and passionate communities has risen up in opposition. While every creative medium is at risk of AI influence now, gamers are particularly sensitive about this technology sucking the creativity and human element from our beloved medium. Even the mere mention of AI being used in game development triggers a massive backlash, but we need to start being more nuanced in how we talk about the ways AI should and should not be used. Because, like it or not, AI is going to become more ubiquitous in gaming. We can’t keep talking about AI as though it is a black-and-white thing. It is a tool, and like any tool, there are ways it can be used appropriately.
The question we need to ask ourselves now is, when is it ethical to use and what crosses the line?
A blurry line
Game development is complicated. I say that upfront to acknowledge that it is easy for us to play armchair developer and say that AI shouldn’t be used under any circumstances, but the reality of the situation is very different. Developers, pundits, and analysts have all been shouting from the rooftops for years now about how unsustainable the current AAA landscape is, so at the very least, we can say that publishers are looking for solutions that cut cost, time, or both.
AI is the big bet right now across multiple disciplines, and that includes gaming. We’re already seeing players like PlayStation experiment with things like AI characters, while Steam is setting up flags to let players know if a game includes AI-generated content. It has been reported that Microsoft’s massive 2025 layoffs were done in part to fund its $80 billion AI infrastructure initiative, which will no doubt seep its way into Xbox’s massive portfolio of studios. Unless there’s some major piece of regulation put in place (which I could never see our current administration doing), then it is only a matter of time before it becomes the norm.
So, when is it okay? There are some clear examples of when it isn’t, such as AI-generated art, writing, or even entire games. Anything that we would hope has a human touch that comes from a person’s vision to communicate something to the player. No one wants to play a game made by AI, right? Okay, so that’s the easy part. But what about the less obvious stuff? We all seem to be okay with AI upscaling. That doesn’t hurt anyone and can be a huge load off developers’ shoulders. What about AI creating code? That’s influencing the game, but is invisible to the player if they weren’t told. Odds are a ton of games are being coded with AI assistance right now to cut down on some of that time-consuming technical work. That’s another way to be more efficient, so should we accept that as well?
11bit studios
The Altersgot hit with a double-whammy of controversy recently over AI, and both are fascinating examples of how grey this entire issue is. The first is that one in-game display uses AI-generated text. This text is illegible under normal circumstances and was left in by mistake, with the intention of replacing it with randomized text before release. Is there so much difference between AI-generated garbage text and pre-generated text? I can understand how one feels worse, but isn’t the end result the same? The other example lands on the wrong side of the ethical line for most. Some of the in-game films the player can watch were added so late in development that 11 Bit Studios didn’t have time to localise them in all languages. So, they used AI to generate subtitles. That’s a bad practice that likely harmed the final product more than if those videos hadn’t been included, but it raises some interesting questions.
And then there’s testing. AI can stress test and find bugs thousands of times faster than a person, but now we’re threatening the jobs of QA testers. Replacing humans is where a lot of people draw that ethical line, so should we not use it here, despite the potential to speed up development? I would never call for people to lose their jobs, but it is a sad reality that some industries do die out as technology advances. If AI is best suited for brute force work like that, is that something that should be embraced? I don’t like slippery slope arguments, but I do think we need to be cautious as to what we support with AI, knowing that capitalism can and will push it to the limits. If these jobs are okay to replace, why not those?
Perhaps an even bigger question we all need to wrestle with is the exceptions to those rules. If we say AI music is unacceptable in games, is there any exception for a solo developer self-funding their game who can’t afford to hire a musician? Would it be better to launch without music or not launch at all? There are arguments to be made on both sides. Going back to the subtitle example, what if a team can’t hire a localization team? Is it better to not let players who speak another language engage with the product at all over using AI as a necessity?
I pose all these questions without answering them because I can’t. I can tell you where I fall on each of these issues, but that isn’t the point. What I am hoping to present are the grey areas where we can have productive discussions about when and where AI is acceptable, if we’re willing to approach it in good faith.
We can’t afford to lump all AI into the same bucket of “AI bad” anymore. It is too nuanced a tool with too many factors to make a blanket judgment call on anymore. Yes, we don’t need AI to make games — we’ve been doing it that way for decades. The issue is that games are so complex, time-consuming, expensive, and risky that we’re in an era where even successful studios are getting closed down. If AI has the potential to ease some of that pressure and make game development a slightly safer industry, we need to start having deeper conversations about when and where it is appropriate to use it instead of vilifying it as a whole.
When is an AI system intelligent enough to be called artificial general intelligence (AGI)? According to one definition reportedly agreed upon by Microsoft and OpenAI, the answer lies in economics: When AI generates $100 billion in profits. This arbitrary profit-based benchmark for AGI perfectly captures the definitional chaos plaguing the AI industry.
In fact, it may be impossible to create a universal definition of AGI, but few people with money on the line will admit it.
Over this past year, several high-profile people in the tech industry have been heralding the seemingly imminent arrival of “AGI” (i.e., within the next two years). But there’s a huge problem: Few people agree on exactly what AGI means. As Google DeepMind wrote in a paper on the topic: If you ask 100 AI experts to define AGI, you’ll get “100 related but different definitions.”
In the wake of Microsoft’s layoffs and studio closures last week, rumors circulated that Romero Games — the studio founded by Doom creator John Romero — had also shut down after funding for its still-untitled game was pulled. Those rumors are false, and the studio is still alive and kicking, even if it isn’t out of the woods yet.
Romero Games shared an update on its social media accounts late yesterday afternoon that the studio “is not closed, and [they] are doing everything in [their] power to ensure it does not come to that.” Perhaps more exciting is the possibility that the cancelled game still has a lifeline.
“We’ve been contacted by several publishers interested in helping us bring the game across the finish line, and we’re currently evaluating those opportunities,” the message reads. The original publisher pulled funding last week, and Romero Games cannot publicly state who that publisher was due to confidentiality agreements. Considering funding vanished after Microsoft announced thousands of layoffs, you can make an educated guess.
— Romero Games Ltd. (@romerogamesltd) July 7, 2025
John Romero added his own comment to the statement, saying “Tl;dr – This is a team and a game worth fighting for.”
While there is some small hope for the game to continue, the studio saw the loss of multiple members (as reported by various social media announcements) and now says it has to “reassess the entire staffing of our studio.” It is unclear what that means. In its original announcement, the studio said its members were “heartbroken that it’s come to this.” Some of team had worked together for more than 20 years.
Next to nothing is known about the game, but it comes from an impressive pedigree. The announcement that other studios are interested in potentially funding it is a flicker of good news — something much needed in an industry that has seen so many jobs lost and livelihoods disrupted.
Microsoft has announced another wave of layoffs, starting with its European-based King subsidiary, the company behind Candy Crush. ZeniMax Studios has also been affected. According to Bloomberg, Microsoft is cutting staff by 10% (roughly 200 jobs), with a total expected loss of 9,000 positions throughout the day. US-based branches will be informed later today about cuts. Amid these layoffs, Jason Schreier reports that Rare’s troubled-but-anticipated game Everwild has also been cancelled.
This new round of layoffs follows a previous round in May. Roughly 6,000 people were let go, with most of the job losses landing in product and engineering departments. Microsoft Gaming Chief Executive Officer Phil Spencer sent an email to staff (as transcribed by Bloomberg): “To position Gaming for enduring success and allow us to focus on strategic growth areas, we will end or decrease work in certain areas of the business and follow Microsoft’s lead in removing layers of management to increase agility and effectiveness.”
Spencer said that affected employees who apply for other positions within the company will be given priority review. This is the fourth wave of layoffs at Xbox within the last 18 months. The company has laid off more than 10,000 people since 2023. Nearly 2,000 people were affected in February 2024, with another 650 cuts in September 2024.
BREAKING: As part of today’s cuts, Xbox has canceled the troubled Rare game Everwild, according to people familiar. News on the job cuts is coming in drips — not sure why Xbox didn’t announce it all at once — but I’ll report what I can as I confirm it.
Everwild was described as “a brand new IP from Rare,” and many fans waited eagerly for its release since its announcement in 2019. However, little information has emerged since that time, with a report in 2021 announcing that development had restarted. In February 2025, Phil Spencer confirmed Everwild was still in development.
Everwild has been in development for more than a decade. The reasons for its cancellation are unclear, but the mysterious game has had a troubled development cycle with little to show for it in all this time.
This story is still developing. We will update throughout the day as more information comes in.
Even the fresh batch of AI features — such as live translations and intelligent Shortcuts — are fully supported on the machines that will soon be five generations old. I can’t say the same about Windows and its AI-powered rebirth with the Copilot package. Before confusion ensues, let me clear things up.
Copilot is a suite of AI features, just like Gemini or Apple Intelligence. Then we have Copilot+ machines, which is a branding for PCs that meet certain hardware-level requirements to enable AI-powered features on Windows laptops and PCs. Here’s the weird part. A healthy bunch of Intel silicon launched in 2025 — even those in the powerful “H” class — don’t meet those AI processing requirements.
Luke Larsen / Digital Trends
All of it has created a weird kind of divide in the Windows ecosystem where certain advanced AI features are locked to a handful of cheaper machines, even if you paid a much higher price to get a laptop with a far more powerful processor. Oddly, it’s not just the hardware, but the software experience that now feels different.
Copilot+ is not merely AI hype
Before we get into the hardware limitations, let’s break down the features. Copilot+ machines require a powerful hardware chip for AI acceleration to enable certain features, down to the OS level. For example, in the Settings app, Microsoft is pushing its own Mu small language model (SML) that runs entirely on the NPU.
The NPU on a chip, however, must meet a certain performance baseline, something not even Intel and AMD silicon launched in 2025 fulfill universally. Let’s start with the AI-powered Settings app interactions. It can now understand natural language queries and make suggestions so that users can directly take action with a click.
Microsoft
If you type something like “My screen doesn’t feel smooth,” the Settings app will show a dialog box underneath the search bar, where you get an actionable button to increase the refresh rate and make the interactions smoother. Apple is chasing something similar and has implemented it within the Spotlight system in macOS Tahoe.
Next, we have Recall. It’s like a time machine system that takes snapshots of your PC activity in the background and analyzes them contextually. In the future, if you seek to revisit or find something, you can simply type a natural language query and find a record of the activity, complete with a link to the webpage or app you were working with. It almost feels magical, and you can read more about my experience here.
The crucial benefit is that a healthy bunch of Copilot+ AI features will run on-device, which means they won’t require an internet connection. That’s convenient, but in hindsight, it’s a huge sigh of relief that all user activity remains locked to your device and nothing is sent to servers.
Microsoft
Copilot+ hardware also enables a bunch of creative features such as Cocreator and Generative Fill in Paint, Super Resolution, Image Creator, and Restyle in the native Photos app. But there are a few that are meaningful for day-to-day PC usage. With Click to Do in the Snipping Tool, the AI analyzes the text and image on the screen, somewhat like Google Lens and Apple Intelligence.
You can select text, look it up on the web with a single click, send email, open a website, summarize, rewrite, and take a wide range of image actions such as copy, share, visual search in Bing, erase objects, remove background, and do more — without ever opening another app.
Microsoft
On the more practical side of things, we have translated Live Captions that cover over 40 languages. The translation and captioning happen in real-time and work during video calls and video watching, too. Finally, we have Windows Studio Effects, which can perform chores such as automatic frame adjustment, portrait lighting tweaks, switch background effects, minimize noise, and even make gaze adjustment.
It’s disheartening, because the Copilot+ experiences in Windows 11 are meaningful OS advancements. Most of them, at least. I have used a few of them extensively, and they feel like a practical evolution. Yet, depriving machines that merely miss out on a powerful NPU, despite packing plenty of compute and graphics processing power, is simply unfortunate.
Microsoft has laid out tight hardware requirements for machines that can bear the Copilot+ badge — 256GB of storage, 16GB DDR5 RAM, and a processor with a dedicated AI accelerator chip that can output a minimum of 40 TOPS performance. That’s a bottleneck from both ends.
Luke Larsen / Digital Trends
First, there are still a healthy bunch of machines that ship with 8GB of RAM, and that too, the DDR4 type memory. Take, for example, the Asus Vivobook 17, which costs $700 and ships with 8GB of DDR4 memory on the entry-point configuration, even with the variant that packs a 13th-generation Intel processor.
Let’s say you pay up to reach 16GB of RAM. Despite that added stress on your wallet, you are still limited by the RAM type and won’t be able to run Copilot+ tools on the machine. It’s worth mentioning that there are a LOT of Windows machines that still pack 8GB of RAM, and even when they go up to 16GB capacity, they still rely on the DDR4-type memory.
Now, it’s time to address the elephant in the room. The silicon situation. The latest from Intel is the Ultra 200 series processor family, which is bifurcated across Arrow Lake and Lunar Lake lines. These Ultra 200 series processors are available in four formats: V-series, U-series, H-series, HX-series, and H-series.
This Razer laptop costs 6x as much as the Surface Pro 12, but still can’t run Copilot+ suite.Razer
Out of the four brackets, only the V-series processors support Copilot+ experiences on Windows 11. Even the enthusiast-class H and HX series processors don’t meet the NPU requirements, and as such, they are devoid of the Copilot+ AI features. As perplexing as the situation remains with Intel Core 200 series silicon, the situation with AMD and its Copilot+ readiness isn’t too different.
At the moment, only AMD’s Ryzen AI 300 series processors fall under the Copilot+ bracket. That means if you invested in a top-shelf AMD silicon in the past few years, or even aim to build an AMD gaming rig this year, you either lose out on Copilot+ perks or must pick from the Ryzen AI 300 series line-up.
Even older Macs do better
The situation with Copilot+ is weird because it has created fault lines in the Windows 11 experience that don’t make sense, neither from a price perspective, nor from a firepower angle. It even makes one feel bad about spending a fortune on a top-tier Intel processor, only to find it locked beyond next-gen AI features in Windows 11 because the NPU isn’t up to the task.
Launched in 2020, the MacBook Air can run the full Apple Intelligence AI package.Andy Boxall / Digital Trends
The only other option is to pick a Qualcomm Snapdragon X-series processor. But in doing so, you run into the compatibility hurdles that come with Windows on Arm. Plus, the GPU limitations rule out gaming or other demanding tasks where you need a powerful GPU. Right now, it seems like Copilot+ is a bag of serious caveats.
And as Microsoft’s team comes with more AI-first experiences, the gulf within Windows 11 is only going to widen. An $800 Copilot+ machine will run native AI experiences that even a powerful desktop won’t be able to handle in the near future. The situation within the Apple ecosystem is just the opposite.
This $800 Asus laptop with Snapdragon chip is Copilot ready, but far pricier gaming laptops are not.Tushar Mehta / Digital Trends
Even if you have a nearly five-year-old M1 MacBook Air, you can run all the Apple Intelligence features just fine. Now, one can argue that AI is not the deciding factor for picking up a laptop. But as companies like Microsoft, Apple, and Google deeply integrate AI packages such as Copilot, Siri, and Gemini across their OS at the native level, these AI features will essentially serve as a key computing evolution.
Google has already given us a glimpse of how tightly interweaving Gemini across its Workspace tools can flesh out, and somewhat similar is the progress of Apple Intelligence within maCOS. But when it comes to the OS-level AI progress, it’s Microsoft that finds itself in an odd place where a huge chunk of Windows 11 users are going to feel left out, while macOS users will move forward just fine even on aging hardware.
In the summer of 2024, corporate anti-malware provider CrowdStrike pushed a broken update to millions of PCs and servers running some version of Microsoft’s Windows software, taking down systems that both companies and consumers relied on for air travel, payments, emergency services, and their morning coffee. It was a huge outage, and it caused days and weeks of pain as the world’s permanently beleaguered IT workers brought systems back online, in some cases touching each affected PC individually to remove the bad update and get the systems back up and running.
The outage was ultimately CrowdStrike’s fault, and in the aftermath of the incident, the company promised a long list of process improvements to keep a bad update like that from going out again. But because the outage affected Windows systems, Microsoft often had shared and sometimes even top billing in mainstream news coverage—another in a string of security-related embarrassments that prompted CEO Satya Nadella and other executives to promise that the company would refocus its efforts on improving the security of its products.
The CrowdStrike crash was possible partly due to how anti-malware software works in Windows. Security vendors and their AV products generally have access to the Windows kernel, the cornerstone of the operating system that sits between your hardware and most user applications. But most user applications don’t have kernel access specifically because a buggy app (or one hijacked by malware) with kernel access can bring the entire system down rather than just affecting the app. The bad CrowdStrike update was bad mostly because it was being loaded so early in Windows’ boot process that many systems couldn’t check for and download CrowdStrike’s fix before they crashed.
Microsoft’s new 13-inch Surface Laptop is an odd one. It’s inarguably a step down in every respect from last year’s 13.8-inch Surface Laptop. But it’s also too good (and too expensive) to be considered a replacement for the Surface Laptop Go, the company’s perennially overpriced and underspecced entry-level laptop. It’s cheaper than last year’s Surfaces, but mostly because Microsoft gave those devices a de facto price hike by killing the entry-level configurations of those PCs.
We’re left with a laptop that’s perfectly fine or even great, depending on what you want. It’s relatively affordable for what is, a sort of MacBook Air-ish, just-the-basics portable computer. But it’s such a step down from the $999 laptop Microsoft released just last year that it’s hard not to see the entire laptop as one big series of compromises.
Where does the 13-inch Surface Laptop belong?
Despite just calling it the “Surface Laptop, 13-inch,” Microsoft clearly considers this to be an entirely different sub-class of laptop rather than a continuation of the flagship Surface Laptop or the lower-end Surface Go. The system identifies itself to Windows as “Surface Laptop 13in 1st Ed with Snapdragon,” where the 13.8- and 15-inch Surface Laptops are both labeled as 7th edition devices.
Nearly a decade ago, Ars testing found that Valve’s “Steam Machines”-era version of SteamOS performed significantly worse than Windows when SteamOS’ Linux game ports were tested on the same hardware as their Windows counterparts. Today, though, Ars testing on the Lenovo Legion Go S finds recent games generally run at higher frame rates on SteamOS 3.7 than on Windows 11. The performance advantage is yet another way that Valve’s upstart OS is differentiating itself from the “default” Windows installation used by most PC gamers for decades now.
While users have been able to install Windows on the Steam Deck since its 2022 launch, Valve doesn’t offer official “Windows on Deck” support for this alternative hardware use case. Lenovo’s Legion Go S, on the other hand, is the first gaming portable explicitly designed to work with either Windows 11 (in hardware first released in January) or SteamOS (in hardware first released in May, alongside a new version of SteamOS designed for non-Valve AMD hardware).
To test the performance impact of this operating system choice, we started with the SteamOS version of the Legion Go S (provided by Lenovo) and tested five high-end 3D games released in the last five years using built-in benchmarking tools and two different graphics/resolution tiers. We then installed Windows 11 on the handheld, downloaded updated drivers from Lenovo’s support site, and re-ran the benchmarks on the same games downloaded through Steam for Windows.
Last month, Microsoft released a modern remake of its classic MS-DOS Editor, bringing back a piece of computing history that first appeared in MS-DOS 5.0 back in 1991. The new open source tool, built with Rust and simply called “Edit,” works on Windows, macOS, and—in a twist that would have seemed unlikely three decades ago—Linux.
The cross-platform availability has delighted longtime users who never expected to see Microsoft’s text editor running on their preferred operating system. “30 years of waiting, and I can use MS Edit on Linux,” wrote one Reddit user, capturing the nostalgic appeal of running a genuinely useful version of a Microsoft DOS utility on a Unix-like system.
An animated GIF from Microsoft showing the modern “Edit” application in action. (Credit: Microsoft)
An animated GIF from Microsoft showing the modern “Edit” application in action. (Credit: