So where do we go now?

The researchers said that both the RTX 3060 and RTX 6000 cards are vulnerable. Changing BIOS defaults to enable IOMMU closes the vulnerability, they said. Short for input-output memory management unit, IOMMU maps device-visible virtual addresses to physical addresses on the host memory. It can be used to make certain parts of memory off-limits.

“In the context of our attack, an IOMMU can simply restrict the GPU from accessing sensitive memory locations on the host,” Kwong explained. “IOMMU is, however, disabled by default in the BIOS to maximize compatibility and because enabling the IOMMU comes with a performance penalty due to the overhead of the address translations.”

A separate mitigation is to enable Error Correcting Codes (ECC) on the GPU, something Nvidia allows to be done using a command line. Like IOMMU, enabling ECC incurs some performance overhead because it reduces the overall amount of available workable memory. Further, some Rowhammer attacks can overcome ECC mitigations.

GPU users should understand that the only cards known to be vulnerable to Rowhammer are the RTX 3060 and RTX 6000 from the Ampere generation, which were introduced in 2020. It wouldn’t be surprising if newer generations of graphics cards from Nvidia and others are susceptible to the same types of attacks, but because the pace of academic research typically lags far behind the faster speed of product rollouts, there’s no way now to know.

Top-tier cloud platforms typically provide security levels that go well beyond those available by default on hobbyist and consumer machines. Another thing to remember: There are no known instances of Rowhammer attacks ever being actively used in the wild.

The true value of the research is to put GPU makers and users alike on notice that Rowhammer attacks on these platforms have the potential to upend security in serious ways. More information about GDDRHammer and GeForge is available here.

• HP EliteBook 6 G2q delivers up to 85 TOPS for local AI tasks
• Always-connected 5G experiences require specific hardware and preinstalled eSIM modules
• Service works only on compatible commercial PCs running Windows 11

HP has unveiled the EliteBook 6 G2q, an ultraslim AI PC that relies on Snapdragon X2 Elite or X2 Plus processors to deliver up to 85 TOPS of NPU performance for local AI tasks.

This lightweight laptop, up to 15% thinner than its predecessor, claims to offer always-connected experiences through HP Go 5G service.

• Supercapacitors turn charging time from hours into mere seconds
• Fast charging exposes the real limits of lithium-ion battery chemistry
• Supercapacitor technology lacks sufficient energy capacity for practical electric vehicles

Dell has introduced a keyboard and mouse combo that charges in five seconds and delivers a full day of use.

The new Dell Pro 7 Rechargeable Compact Keyboard and Mouse relies on supercapacitor technology rather than traditional lithium-ion batteries.

from the silly-times dept

Back in October, Meta announced that its new Instagram Teen Accounts would feature content moderation “guided by the PG-13 rating.” On its face, this made a certain kind of sense as a communication strategy: parents know what PG-13 means (or at least think they do), and Meta was clearly trying to borrow that cultural familiarity to signal that it was taking teen safety seriously.

The Motion Picture Association, however, was not amused. Within hours of the announcement, MPA Chairman Charles Rivkin fired off a statement. Then came a cease-and-desist letter. Then a Washington Post op-ed whining about the threat to its precious brand. The MPA was very protective of its trademark, and very unhappy that Meta was freeloading off the supposed credibility of its widely mocked rating system.

And now, this week, the two sides have announced a formal resolution in which Meta has agreed to “substantially reduce” its references to PG-13 and include a rather remarkable disclaimer:

“There are lots of differences between social media and movies. We didn’t work with the MPA when updating our content settings, and they’re not rating any content on Instagram, and they’re not endorsing or approving our content settings in any way. Rather, we drew inspiration from the MPA’s public guidelines, which are already familiar to parents. Our content moderation systems are not the same as a movie ratings board, so the experience may not be exactly the same.”

In Meta’s official response, you can practically hear the PR team gritting their teeth:

“We’re pleased to have reached an agreement with the MPA. By taking inspiration from a framework families know, our goal was to help parents better understand our teen content policies. We rigorously reviewed those policies against 13+ movie ratings criteria and parent feedback, updated them, and applied them to Teen Accounts by default. While that’s not changing, we’ve taken the MPA’s feedback on how we talk about that work. We’ll keep working to support parents and provide age-appropriate experiences for teens,” said a Meta spokesperson.

Translation: we’re still doing the same thing, we’re just no longer allowed to call it what we were calling it.

There are several layers of nonsense worth unpacking here. First, there’s the MPA getting all high and mighty about its rating system. Let’s remember how the MPA’s film rating system came into existence in the first place: it was a voluntary self-regulation scheme created in the late 1960s specifically to head off government regulation after the government started making noises about the harm Hollywood was doing to children with the content it platformed. Sound familiar? The studios decided that if they rated their own content, maybe Congress would leave them alone. As the MPA explains in their own boilerplate:

For nearly 60 years, the MPA’s Classification and Rating Administration’s (CARA) voluntary film rating system has helped American parents make informed decisions about what movies their children can watch… CARA does not rate user-generated content. CARA-rated films are professionally produced and reviewed under a human-centered system, while user-generated posts on platforms like Instagram are not subject to the same rating process.

Sure, there’s a trademark issue here, but let’s be real: no one thought Instagram was letting a panel of Hollywood parents rate the latest influencer videos.

Next, the PG-13 analogy never actually made much sense for social media. As we discussed on Ctrl-Alt-Speech back when this whole thing started, the context and scale are just completely different. At the time, I pointed out that a system designed to rate a 90-minute professionally produced film — reviewed in its entirety by a panel of parents — is a wholly different beast than moderating hundreds of millions of short-form posts generated by individuals (and AI) every single day.

So, yes, calling the system “PG-13” was a marketing gimmick, meant to trade on a familiar brand while obscuring how differently social media actually works — but the idea that this somehow dilutes the MPA’s marks is still pretty silly.

Then there’s the rating system’s well-documented arbitrariness. The MPA’s ratings have been criticized for decades for their seemingly incoherent standards. On that same podcast, I noted that the rating system is famous for its selective prudishness — nudity gets you an R rating, but two hours of violence can skate by with a PG-13.

There was a whole documentary about this — This Film Is Not Yet Rated — that exposed just how subjective and inconsistent the whole process was. Meta was effectively borrowing credibility from a system that was itself created as a regulatory dodge, is famously inconsistent, and was designed for an entirely different medium. And the MPA’s response was essentially: “Hey, that’s our famously inconsistent regulatory dodge, and you can’t have it.”

The whole thing was silly. And now it’s been formally resolved with Meta agreeing to stop doing the thing it had already mostly stopped doing back in December. So even the resolution is anticlimactic.

But there’s a more substantive point buried under all this trademark squabbling: the whole approach reflects a flawed assumption that one company can set a universal standard for every teen on the planet.

As I argued on the podcast, the deeper issue is that the whole framework is wrong for the medium. The MPA’s rating system was built to evaluate a single 90-minute film, reviewed in its entirety by a panel of parents. Applying that logic to hundreds of millions of short-form posts generated by people across wildly different cultural contexts — a kid in rural Kansas, a teenager in Berlin, a twelve-year-old in Lagos — was never going to produce anything coherent. Different kids, different families, different communities have different standards, and no single company should be setting a universal threshold for all of them. The smarter approach is giving parents and users real controls with customizable defaults, rather than having Zuckerberg (or a Hollywood trade association) decide what counts as age-appropriate for every teenager on the planet.

This whole dispute was silly from start to finish.

Filed Under: content moderation, movie ratings, pg-13, social media

Companies: meta, mpa