• Nvidia has commanding lead over rivals in latest Adobe After Effects benchmarks
• Even lower-performance Nvidia GPUs outpace Intel and AMD cards
• But to Apple’s credit, the M3 Max pulls ahead in 2D significantly despite its laptop form factor

Nvidia‘s GeForce RTX 40-Series GPUs has shown off some significant advantages when it comes to dealing with 3D workflows over comparable Intel and AMD cards, new figures have claimed.

The latest Puget Systems After Effects benchmarks say Nvidia’s flagship GeForce RTX 4090 delivered up to 20 times the performance of Apple’s MacBook Pro M3 Max in 3D tasks; reflecting the card’s technical design focus on GPU-intensive workloads.

The one exception to that is the UMG v. Anthropic case, because at least early on, earlier versions of Anthropic would generate the song lyrics for songs in the output. That’s a problem. The current status of that case is they’ve put safeguards in place to try to prevent that from happening, and the parties have sort of agreed that, pending the resolution of the case, those safeguards are sufficient, so they’re no longer seeking a preliminary injunction.

At the end of the day, the harder question for the AI companies is not is it legal to engage in training? It’s what do you do when your AI generates output that is too similar to a particular work?

Do you expect the majority of these cases to go to trial, or do you see settlements on the horizon?

There may well be some settlements. Where I expect to see settlements is with big players who either have large swaths of content or content that’s particularly valuable. The New York Times might end up with a settlement, and with a licensing deal, perhaps where OpenAI pays money to use New York Times content.

There’s enough money at stake that we’re probably going to get at least some judgments that set the parameters. The class-action plaintiffs, my sense is they have stars in their eyes. There are lots of class actions, and my guess is that the defendants are going to be resisting those and hoping to win on summary judgment. It’s not obvious that they go to trial. The Supreme Court in the Google v. Oracle case nudged fair-use law very strongly in the direction of being resolved on summary judgment, not in front of a jury. I think the AI companies are going to try very hard to get those cases decided on summary judgment.

Why would it be better for them to win on summary judgment versus a jury verdict?

It’s quicker and it’s cheaper than going to trial. And AI companies are worried that they’re not going to be viewed as popular, that a lot of people are going to think, Oh, you made a copy of the work that should be illegal and not dig into the details of the fair-use doctrine.

There have been lots of deals between AI companies and media outlets, content providers, and other rights holders. Most of the time, these deals appear to be more about search than foundational models, or at least that’s how it’s been described to me. In your opinion, is licensing content to be used in AI search engines—where answers are sourced by retrieval augmented generation or RAG—something that’s legally obligatory? Why are they doing it this way?

If you’re using retrieval augmented generation on targeted, specific content, then your fair-use argument gets more challenging. It’s much more likely that AI-generated search is going to generate text taken directly from one particular source in the output, and that’s much less likely to be a fair use. I mean, it could be—but the risky area is that it’s much more likely to be competing with the original source material. If instead of directing people to a New York Times story, I give them my AI prompt that uses RAG to take the text straight out of that New York Times story, that does seem like a substitution that could harm the New York Times. Legal risk is greater for the AI company.

What do you want people to know about the generative AI copyright fights that they might not already know, or they might have been misinformed about?

The thing that I hear most often that’s wrong as a technical matter is this concept that these are just plagiarism machines. All they’re doing is taking my stuff and then grinding it back out in the form of text and responses. I hear a lot of artists say that, and I hear a lot of lay people say that, and it’s just not right as a technical matter. You can decide if generative AI is good or bad. You can decide it’s lawful or unlawful. But it really is a fundamentally new thing we have not experienced before. The fact that it needs to train on a bunch of content to understand how sentences work, how arguments work, and to understand various facts about the world doesn’t mean it’s just kind of copying and pasting things or creating a collage. It really is generating things that nobody could expect or predict, and it’s giving us a lot of new content. I think that’s important and valuable.

The U.K.’s Competition and Markets Authority (CMA) is launching “strategic market status” (SMS) investigations into the mobile ecosystems of Apple and Google.

The investigations constitute part of the new Digital Markets, Competition and Consumers (DMCC) Act which passed last year and came into effect in January. The Act includes new powers for the CMA to designate companies as having strategic market status if they are deemed to be overly dominant, and propose remedies and interventions to improve competition.

The CMA announced its first such SMS investigation last week, launching a probe into Google Search’s market share which is reportedly around the 90% mark. The regulator announced at the time that a second one would be coming in January, and we now know that it’s using its fresh powers to establish whether Apple and Google have strategic market status in their respective mobile ecosystems, which covers areas like browsers, app stores, and operating systems.

‘Holding back innovation’

Today’s announcement doesn’t come as a major surprise. Back in August, the CMA said it was closing a duo of investigations into Apple and Google’s respective mobile app ecosystems, which it had launched starting back in 2021. However, the CMA made it clear that this would be more of a pause, and it would be looking to use its new powers to address competition concerns around the two biggest players in the mobile services market.

In November, an inquiry group set up by the CMA concluded that Apple’s mobile browser policies and a pact with Google were “holding back innovation” in the U.K. The findings noted that Apple forced third-party mobile browsers to use Apple’s browser engine, WebKit, which restricts what these browsers are able to do in comparison to Apple’s own Safari browser, and thus limits how they can effectively differentiate in what is a competitive market.

As part of its new probe, the CMA has now confirmed that it will look at “the extent of competition between and within” Apple’s and Google’s respective mobile ecosystems, including barriers that may be preventing others from competing. This will include whether either company is using their dominant position in operating systems, app distribution, or browsers to “favour their own apps and services” — many of which are bundled by default and can’t always be uninstalled.

On top of that, the CMA said it would look into whether either company imposes “unfair terms and conditions” on developers that wish to distribute their apps through their app stores.

Alex Haffner, competition partner at U.K. law firm Fladgate, said that today’s announcement was “wholly expected,” adding that the more interesting facet is how this new probe fits into the broader changes underway at the U.K. regulator.

Indeed, news emerged this week that the CMA had appointed ex-Amazon executive Doug Gurr as interim chair, constituting part of a wider shift as the U.K. positions itself as a pro-growth, pro-tech nation by cutting red tape and bureaucracy.

“What is more interesting is how this fits into the current sea change which is engulfing the broader organisation of the CMA and in particular the very clear steer it is getting from central government to ensure that regulation is consistently applied with its pro-growth agenda,” Haffner said in a statement issued to TechCrunch. “We can expect this to feature heavily once the CMA gets its teeth stuck into the specifics of the DMCC regime, and its dealings with the tech companies involved.”

Remedies

Today’s announcement kickstarts a three-week period during which relevant stakeholders are invited to submit comments as part of the investigations, with the outcomes expected to be announced by October 22, 2025. While it’s still early days, potential remedies — in the event that Apple and Google are deemed to have strategic market status — include requiring the companies to provide third-parties with greater access to key functionality to help them better compete. It also may include making it easier to pay for services outside of Apple and Google’s existing app store structure.

In a statement issued to TechCrunch, an Apple spokesperson said that it will “continue to engage constructively with the CMA” as their investigation progresses.

“Apple believes in thriving and dynamic markets where innovation can flourish,” the spokesperson said. “We face competition in every segment and jurisdiction where we operate, and our focus is always the trust of our users. In the U.K. alone, the iOS app economy supports hundreds of thousands of jobs and makes it possible for developers big and small to reach users on a trusted platform.”

Oliver Bethell, senior director for competition at Google, echoed this sentiment, noting that the company “will work constructively with the CMA.”

“Android’s openness has helped to expand choice, reduce prices and democratise access to smartphones and apps. It’s the only example of a successful and viable open source mobile operating system,” Bethell wrote in a blog post today. “We favour a way forward that avoids stifling choice and opportunities for U.K. consumers and businesses alike, and without risk to U.K. growth prospects.”

• Elon Musk slams Project Stargate initiative
• Claims backers of $500bn initiative “don’t actually have the money”
• Sam Altman, Satya Nadella, hit back at Musk claims

The global AI market appears to have descended into a playground battle of insults after Elon Musk, Sam Altman, Satya Nadella, and others all clashed over the launch of Project Stargate.

Revealed earlier this week to huge fanfare as part of the new Trump administration’s plans to boost AI across the US, Project Stargate is reportedly set to see as much as $500 billion invested into data centers to support the increasing data needs of Altman’s OpenAI.

We may see OpenAI’s agent tool, Operator, released sooner rather than later. Changes to ChatGPT’s code base suggest that Operator will be available as an early research preview to users on the $200 Pro subscription plan.

The changes aren’t yet publicly visible, but a user on X who goes by Choi spotted these updates in ChatGPT’s client-side code. TechCrunch separately identified the same references to Operator on OpenAI’s website.

Here are the three interesting tidbits we spotted:

• There are multiple references to the operator.chatgpt.com URL. This URL currently redirects to the main chatgpt.com web page.
• There will be a new popup that tells you to upgrade your plan if you want to try Operator. “Operator is currently only available to Pro users as an early research preview,” it says.
• On the page that lists the Plus and Pro plans, OpenAI will add “Access to research preview of Operator” as one of the benefits of the Pro plan.

Bloomberg previously reported that OpenAI was working on a general-purpose agent that can perform tasks in a web browser for you.

While this sounds a bit abstract, think about all the mundane things you do regularly in your web browser with quite a few clicks — following someone on LinkedIn, adding an expense in Concur, assigning a task to someone in Asana, or changing the status of a prospect on Salesforce. An agent could perform such multi-step tasks based on an instruction set.

More recently, The Information reported that OpenAI could launch Operator as early as this week. With today’s changes, it seems like everything is ready for a public launch.

Anthropic has released an AI model that can control your PC using a “Computer Use” API and local tools that control your mouse and keyboard. It is currently available as a beta feature for developers.

It looks like Operator is going to be usable on ChatGPT’s website, meaning that it won’t interact with your local computer. Instead, OpenAI will likely run a web browser on its own servers to perform tasks for you.

Nevertheless, it indicates that OpenAI’s ability to interact with computers is progressing. Operator is a specific sandboxed implementation of the company’s underlying agentic framework. It’s going to be interesting to see if the company has more information to share on the technology that powers Operator.

• Beyerdynamic’s new IEMs come in four specifications, for every band member
• The numbers you need to remember are 70, 71, 72 or 73
• …Oh, and $499, which is the price

Revered hi-fi brand Beyerdynamic (see the Aventho 300 for the firm’s most recent headphone hit, but that’s just for starters) has released a new line of professional in-ear monitors, and the company wants you to know that every member of the band has been specifically catered for here.

The DT 70 IE, DT 71 IE, DT 72 IE, and DT 73 IE (that’s the full quartet) all feature Beyerdynamic’s own TESLA.11 dynamic driver system, boasting a Total Harmonic Distortion (often abbreviated to THD) of just 0.02%, which is very low indeed – anything below 0.1% is typically considered gifted for an in-ear monitor. Beyer calls it “one of the loudest, lowest-distortion systems available”, but you also get five different sizes of silicone eartips and three pairs of Comply memory foam eartips to achieve a decent fit and seal (nobody wants distractions from the Amazon delivery guy outside while trying to lay down a particular riff).

The dream of battery-free devices has taken an unlikely turn, as Carnegie Mellon researchers debuted Power-Over-Skin. The technology allows for electrical currents to travel through human skin in a bid to power things like blood sugar monitors, pacemakers, and even consumer wearables like smart glasses and fitness trackers.

Researchers note the tech is still in “early stages.” At the moment, they’ve showcased the tech supporting low-power electronics like the LED earring pictured above.

“It’s similar to how a radio uses the air as the medium between the transmitter station and your car stereo,” notes CMU researcher Andy Kong. “We’re just using body tissue as the transmitting medium in this case.”

AI companies like Google, OpenAI, and Anthropic want you to believe we’re on the cusp of Artificial General Intelligence (AGI)—a world where AI tools can outthink humans, handle complex professional tasks without breaking a sweat, and chart a new frontier of autonomous intelligence. Google just rehired the founder of Character.AI to accelerate its quest for AGI, OpenAI recently released its first “reasoning” model, and Anthropic’s CEO Dario Amodei says AGI could be achieved as early as 2026.

But here’s the uncomfortable truth: in the quest for AGI in high-stakes fields like medicine, law, veterinary advice, and financial planning, AI isn’t just “not there yet,” it may never get there.

The Hard Facts on AI’s Shortcomings

This prediction is based on several decades of research that my colleagues and I have been undertaking at the University of Oxford to establish what makes people willing to fight and die for their groups. We use a variety of methods, including interviews, surveys, and psychological experiments to collect data from a wide range of groups, such as tribal warriors, armed insurgents, terrorists, conventional soldiers, religious fundamentalists, and violent football fans.

We have found that life-changing and group-defining experiences cause our personal and collective identities to become fused together. We call it “identity fusion.” Fused individuals will stop at nothing to advance the interests of their groups, and this applies not only to acts we would applaud as heroic—such as rescuing children from burning buildings or taking a bullet for one’s comrades—but also acts of suicide terrorism.

Fusion is commonly measured by showing people a small circle (representing you) and a big circle (representing your group) and placing pairs of such circles in a sequence so that they overlap to varying degrees: not at all, then just a little bit, then a bit more, and so on until the little circle is completely enclosed in the big circle. Then people are asked which pair of circles best captures their relationship with the group. People who choose the one in which the little circle is inside the big circle are said to be “fused.” Those are people who love their group so much that they will do almost anything to protect it.

This isn’t unique to humans. Some species of birds will feign a broken wing to draw a predator away from their fledglings. One species—the superb fairy wren of Australasia—lures predators away from their young by making darting movements and squeaky sounds to imitate the behavior of a delectable mouse. Humans too will typically go to great lengths to protect their genetic relatives, especially their children who (except for identical twins) share more of their genes than other family members. But—unusually in the animal kingdom—humans often go further still by putting themselves in harm’s way to protect groups of genetically unrelated members of the tribe. In ancient prehistory, such tribes were small enough that everyone knew everybody else. These local groups bonded through shared ordeals such as painful initiations, by hunting dangerous animals together, and by fighting bravely on the battlefield.

Nowadays, however, fusion is scaled up to vastly bigger groups, thanks to the ability of the world’s media—including social media—to fill our heads with images of horrendous suffering in faraway regional conflicts.

When I met with one of the former leaders of the terrorist organization Jemaah Islamiyah in Indonesia, he told me he first became radicalized in the 1980s after reading newspaper reports about the treatment of fellow Muslims by Russian soldiers in Afghanistan. Twenty years later, however, nearly a third of American extremists were radicalized via social media feeds, and by 2016 that proportion had risen to about three quarters. Smartphones and immersive reporting shrinks the world to such an extent that forms of shared suffering in face-to-face groups can now be largely recreated and spread to millions of people across thousands of miles at the click of a button.

Fusion based on shared suffering may be powerful, but is not sufficient by itself to motivate violent extremism. Our research suggests that three other ingredients are also necessary to produce the deadly cocktail: outgroup threat, demonization of the enemy, and the belief that peaceful alternatives are lacking. In regions such as Gaza, where the sufferings of civilians are regularly captured on video and shared around the world, it is only natural that rates of fusion among those watching on in horror will increase. If people believe that peaceful solutions are impossible, violent extremism will spiral.

Samsung Unpacked’s “one more thing” was a bit of a weird one. After the presentation ended, the company rolled a brief pre-packaged video of the Galaxy Edge — not to be confused with the “Star Wars” theme park of the same name.

Though limited, the reveal was confirmation of earlier rumors that the hardware giant is working on an extra-thin version of its new S25 flagship. The Galaxy S25 Edge is, presumably, another tier for the line, slotting in alongside the S25, S25+, and S25 Ultra.

Key details, including pricing, availability, and actual thickness were not revealed, though the company did showcase what appeared to be dummy models at Wednesday’s event. Early rumors pointed to a 6.4 mm thickness, a considerable reduction from the base Galaxy S25’s 7.2 mm.

Samsung clearly wanted to avoid taking too much wind out of the Galaxy S25’s sails during the event, so it opted instead for a more cryptic reveal. Even so, the mere appearance of the device at Unpacked may be enough to keep early adopters from preordering the S25 ahead of its February 7 release.

After all, those are precisely the folks who get excited by things like a 0.8 mm profile reduction.