Social media is going the way of alcohol, gambling, and other social sins: Societies are deciding it’s no longer kid stuff. Lawmakers point to compulsive use, exposure to harmful content, and mounting concerns about adolescent mental health. So, many propose to set a minimum age, usually 13 or 16.
In cases when regulators demand real enforcement rather than symbolic rules, platforms run into a basic technical problem. The only way to prove that someone is old enough to use a site is to collect personal data about who they are. And the only way to prove that you checked is to keep the data indefinitely. Age-restriction laws push platforms toward intrusive verification systems that often directly conflict with modern data-privacy law.
This is the age-verification trap. Strong enforcement of age rules undermines data privacy.
How Does Age Enforcement Actually Work?
Most age-restriction laws follow a familiar pattern. They set a minimum age and require platforms to take “reasonable steps” or “effective measures” to prevent underage access. What these laws rarely spell out is how platforms are supposed to tell who is actually over the line. At the technical level, companies have only two tools.
Advertisement
The first is identity-based verification. Companies ask users to upload a government ID, link a digital identity, or provide documents that prove their age. Yet in many jurisdictions, 16-year-olds do not have IDs. In others, IDs exist but are not digital, not widely held, or not trustworthy. Storing copies of identity documents also creates security and misuse risks.
The second option is inference. Platforms try to guess age based on behavior, device signals, or biometric analysis, most commonly facial age estimation from selfies or videos. This avoids formal ID collection, but it replaces certainty with probability and error.
In practice, companies combine both. Self-declared ages are backed by inference systems. When confidence drops, or regulators ask for proof of effort, inference escalates to ID checks. What starts as a light-touch checkpoint turns into layered verification that follows users over time.
What Are Platforms Doing Now?
This pattern is already visible on major platforms.
Advertisement
Meta has deployed facial age estimation on Instagram in multiple markets, using video-selfie checks through third-party partners. When the system flags users as possibly underaged, it prompts them to record a short selfie video. An AI system estimates their age and, if it decides they are under the threshold, restricts or locks the account. Appeals often trigger additional checks, and misclassifications are common.
TikTok has confirmed that it also scans public videos to infer users’ ages. Google and YouTube rely heavily on behavioral signals tied to viewing history and account activity to infer age, then ask for government ID or a credit card when the system is unsure. A credit card functions as a proxy for adulthood, even though it says nothing about who is actually using the account. The Roblox games site, which recently launched a new age-estimate system, is already suffering from users selling child-aged accounts to adult predators seeking entry to age-restricted areas, Wired reports.
For a typical user, age is no longer a one-time declaration. It becomes a recurring test. A new phone, a change in behavior, or a false signal can trigger another check. Passing once does not end the process.
How Do Age-Verification Systems Fail?
These systems fail in predictable ways.
Advertisement
False positives are common. Platforms identify as minors adults with youthful faces, or adults who are sharing family devices, or have otherwise unusual usage. They lock accounts, sometimes for days. False negatives also persist. Teenagers learn quickly how to evade checks by borrowing IDs, cycling accounts, or using VPNs.
The appeal process itself creates new privacy risks. Platforms must store biometric data, ID images, and verification logs long enough to defend their decisions to regulators. So if an adult who is tired of submitting selfies to verify their age finally uploads an ID, the system must now secure that stored ID. Each retained record becomes a potential breach target.
Scale that experience across millions of users, and you bake the privacy risk into how platforms work.
Is Age Verification Compatible With Privacy Law?
This is where emerging age-restriction policy collides with existing privacy law.
Advertisement
Modern data-protection regimes all rest on similar ideas: Collect only what you need, use it only for a defined purpose, and keep it only as long as necessary.
Age enforcement undermines all three.
To prove they are following age-verification rules, platforms must log verification attempts, retain evidence, and monitor users over time. When regulators or courts ask whether a platform took reasonable steps, “We collected less data” is rarely persuasive. For companies, defending themselves against accusations of neglecting to properly verify age supersedes defending themselves against accusations of inappropriate data collection.
It is not an explicit choice by voters or policymakers, but instead a reaction to enforcement pressure and how companies perceive their litigation risk.
Outside wealthy democracies, the trade-off is even starker.
Brazil’s Statute of Child-rearing and Adolescents (ECA in Portuguese) imposes strong child-protection duties online, while its data-protection law restricts data collection and processing. Now providers operating in Brazil must adopt effective age-verification mechanisms and can no longer rely on self-declaration alone for high-risk services. Yet they also face uneven identity infrastructure and widespread device sharing. To compensate, they rely more heavily on facial estimation and third-party verification vendors.
In Nigeria many users lack formal IDs. Digital service providers fill the gap with behavioral analysis, biometric inference, and offshore verification services, often with limited oversight. Audit logs grow, data flows expand, and the practical ability of users to understand or contest how companies infer their age shrinks accordingly. Where identity systems are weak, companies do not protect privacy. They bypass it.
The paradox is clear. In countries with less administrative capacity, age enforcement often produces more surveillance, not less, because inference fills the void of missing documents.
Advertisement
How Do Enforcement Priorities Change Expectations?
Some policymakers assume that vague standards preserve flexibility. In the U.K., then–Digital Secretary Michelle Donelan, argued in 2023 that requiring certain online safety outcomes without specifying the means would avoid mandating particular technologies. Experience suggests the opposite.
When disputes reach regulators or courts, the question is simple: Can minors still access the platform easily? If the answer is yes, authorities tell companies to do more. Over time, “reasonable steps” become more invasive.
Repeated facial scans, escalating ID checks, and long-term logging become the norm. Platforms that collect less data start to look reckless by comparison. Privacy-preserving designs lose out to defensible ones.
This pattern is familiar, including online sales-tax enforcement. After courts settled that large platforms had an obligation to collect and remit sales taxes, companies began continuous tracking and storage of transaction destinations and customer location signals. That tracking is not abusive, but once enforcement requires proof over time, companies build systems to log, retain, and correlate more data. Age verification is moving the same way. What begins as a one-time check becomes an ongoing evidentiary system, with pressure to monitor, retain, and justify user-level data.
Advertisement
The Choice We Are Avoiding
None of this is an argument against protecting children online. It is an argument against pretending there is no trade-off.
Some observers present privacy-preserving age proofs involving a third party, such as the government, as a solution, but they inherit the same structural flaw: Many users who are legally old enough to use a platform do not have government ID. In countries where the minimum age for social media is lower than the age at which ID is issued, platforms face a choice between excluding lawful users and monitoring everyone. Right now, companies are making that choice quietly, after building systems and normalizing behavior that protects them from the greater legal risks. Age-restriction laws are not just about kids and screens. They are reshaping how identity, privacy, and access work on the Internet for everyone.
The age-verification trap is not a glitch. It is what you get when regulators treat age enforcement as mandatory and privacy as optional.
The build is based around the ESP32-2432S028—also known as the CYD, or Cheap Yellow Display, for the integrated 320 x 240 LCD screen. [Jordan] took this all-in-one device and wrapped it in an attractive 3D-printed housing in the shape of an old-school CRT monitor, just… teenier. A special lever mechanism was built in to the enclosure to allow front panel controls to activate the tactile buttons on the CYD board. The ESP32 is programmed to check Open-Meteo feeds for forecasts and current weather data, while also querying a webcam feed and satellite and radar JPEGs from available weather services. These are then displayed on screen in a way that largely resembles the Windows 95 UI design language, with pages for current conditions, future forecasts, wind speeds, and the like.
Crypto-powered gift card store Bitrefill says that the attack it suffered at the beginning of the month was likely perpetrated by North Korean hackers of the Bluenoroff group.
During the investigation, the platform observed indicators similar to previous attacks attributed to the North Korean threat actor, like tactics, malware, IP and email addresses.
“Based on indicators observed during the investigation – including the modus operandi, the malware used, on-chain tracing and reused IP + email addresses (!) – we find many similarities between this attack and past cyberattacks by the DPRK Lazarus / Bluenoroff group against other companies in the crypto industries,” reads Bitrefill’s statement.
Bitrefill is a mid-sized e-commerce platform that enables people to pay in cryptocurrency for gift cards at stores in 150 countries. The gift cards can be used to pay for anything from clothing, food and groceries, health and beauty products to bills, services, gas, transportation, and electronics.
Advertisement
The platform supports more than 600 mobile operators and thousands of brands worldwide.
On March 1st, Bitrefill announced technical issues affecting access to its website and app. A day later, the company disclosed that it had identified a security issue and took all services offline.
Although user balances were not affected, the gradual restoration of all services still continues to this day.
The breach was discovered after Bitrefill noticed suspicious supplier purchasing patterns, exploitation of gift card stock and supply lines, and draining of some “hot” wallets.
Advertisement
The investigation the firm launched to determine the cause revealed that the attack originated on a compromised employee’s laptop.
The attackers stole legacy credentials and used them to access a snapshot with production secrets, later escalating access to the larger Bitrefill infrastructure, including parts of the database and some cryptocurrency wallets.
About 18,500 purchase records containing customer email addresses, IP addresses, and cryptocurrency payment addresses were exposed in the breach. For 1,000 purchases, customer names were also exposed.
Although this information is stored in encrypted form, Bitrefill notes that the attackers may have obtained the decryption keys.
Advertisement
Bitrefill says this was the most serious cyberattack it has suffered in its ten years of existence, but it survived with minimal losses, which will be covered from its capital.
Ultimately, Bitrefill believes that attackers were after cryptocurrency and gift card inventory, not customer information.
BlueNoroff, also known as APT38, is a cluster of the Lazarus group that has been active since at least 2014. It typically targets financial organizations, with a more recent focus on the cryptocurrency industry, the objective being crypto theft.
Bitrefill says this was the most serious cyberattack it has suffered in the ten years of its existence, but it survived with minimal losses, which will be covered from its capital.
Advertisement
Meanwhile, it is expanding security reviews and pen-testing, tightening access controls, improving logging and monitoring, and refining automated shutdown mechanisms.
At this time, most of its services have returned to normal operational status, and customers aren’t required to take any action beyond treating incoming communications with extra caution.
Malware is getting smarter. The Red Report 2026 reveals how new threats use math to detect sandboxes and hide in plain sight.
Download our analysis of 1.1 million malicious samples to uncover the top 10 techniques and see if your security stack is blinded.
GMKtec EVO-T2 mini PC reaches 180 TOPS using combined CPU, GPU, and NPU acceleration
Its PCIe 5.0 storage introduces data speeds exceeding 10GB per second
Local AI models run without relying on external cloud infrastructure
At a recent launch event, GMKtec introduced the GMKtec EVO-T2, a compact desktop system built for local AI computing.
According to the company, the device integrates third-generation Intel Core Ultra processors and claims up to 180 TOPS of compute capability.
It combines CPU, GPU, and NPU resources, and enables local execution of large language models up to 70B parameters without dependence on external cloud infrastructure.
Article continues below
Advertisement
Compute architecture and AI workloads
The EVO-T2 is based on Intel’s Panther Lake architecture and is manufactured using the 18A process node, incorporating RibbonFET transistors and backside power delivery.
These design elements are associated with improved efficiency and transistor density, although most performance data referenced remains tied to internal benchmarks.
Advertisement
The company claims that complex workloads such as code generation and document processing can be executed rapidly with this device.
For some tasks, GMKtec says the EVO-T2 completes them within seconds under controlled conditions.
Sign up to the TechRadar Pro newsletter to get all the top news, opinion, features and guidance your business needs to succeed!
Graphics capabilities are handled by the integrated Intel Arc B390 GPU, which includes twelve Xe cores and support for DirectX 12 Ultimate, real-time ray tracing, and AI-assisted upscaling.
Advertisement
This configuration allows the system to extend beyond AI inference into areas such as rendering and visual content workflows.
Despite its small footprint, the device includes dual M.2 storage slots supporting PCIe 5.0 and PCIe 4.0, with total capacity reaching up to 16TB.
PCIe 5.0 SSDs are theoretically capable of sequential speeds exceeding 10GB/s, with some exceeding 15 GB/s, while PCIe 4.0 drives typically reach around 7GB/s under optimal conditions.
Advertisement
For connectivity, it includes USB4 with 40Gbps bandwidth and OCuLink support for external GPUs.
In addition, the system supports dual Ethernet configurations, offering both 10GbE and 2.5GbE networking.
To address memory constraints, Phison collaborated with GMKtec to integrate aiDAPTIV+ AI SSD technology.
This system dynamically extends available memory by distributing workloads between DRAM and storage, allowing large models to be segmented during execution.
Advertisement
Active portions are processed on the GPU, while less active data remains stored across memory and SSD layers.
This “pseudo-memory” mechanism is described as reducing bottlenecks when processing large models.
However, its long-term performance implications under sustained workloads have not been independently verified.
GMKtec states that it “effectively breaks through traditional DRAM limitations,” a claim that may require independent validation.
Advertisement
The system ships with a pre-configured AI environment, allowing immediate access to AI tools and models without manual setup.
OpenClaw enables the EVO-T2 to run autonomous AI agents locally, performing tasks from data processing to content generation without relying on cloud services.
Every few years, a piece of open-source software arrives that rewires how the industry thinks about computing. Linux did it for servers. Docker did it for deployment. OpenClaw — the autonomous AI agent platform that went from niche curiosity to the fastest-growing open-source project in history in a matter of weeks — may be doing it for software itself.
Nvidia CEO and co-founder Jensen Huang made his position plain at GTC 2026 this week: “OpenClaw is the operating system for personal AI. This is the moment the industry has been waiting for — the beginning of a new renaissance in software.” And Nvidia wants to be the company that makes it enterprise-ready.
At its annual large GTC 2026 conference in San Jose this week, Nvidia unveiled NemoClaw, a software stack that integrates directly with OpenClaw and installs in a single command. Along with it came Nvidia OpenShell, an open-source security runtime designed to give autonomous AI agents — or “claws”, as the industry is increasingly calling them — the guardrails they need to operate inside real enterprise environments. Alongside both, the company announced an expanded Nvidia Agent Toolkit, a full-stack platform for building and running production-grade agentic workflows.
The message from Jensen Huang was unambiguous. “Claude Code and OpenClaw have sparked the agent inflection point — extending AI beyond generation and reasoning into action,” the Nvidia CEO said ahead of the conference. “Employees will be supercharged by teams of frontier, specialized and custom-built agents they deploy and manage.” Watch my video overview of it below and read on for more:
Advertisement
Why ‘claws’ — and why it matters that Nvidia is using the word
The terminology shift happening inside enterprise AI circles is subtle but significant. Internally, teams building with OpenClaw and similar platforms have taken to calling individual autonomous agents claws — a nod to the platform name, but also a useful shorthand for a new class of software that differs fundamentally from the chatbots and copilots of the last two years.
As Kari Briski, Nvidia’s VP of generative AI software, put it during a Sunday briefing: “Claws are autonomous agents that can plan, act, and execute tasks on their own — they’ve gone from just thinking and executing on tasks to achieving entire missions.”
That framing matters for IT decision-makers. Claws are not just assistants. They are persistent, tool-using programs that can write code, browse the web, manipulate files, call APIs, and chain actions together over hours or days without human input. The productivity upside is substantial. So is the attack surface. Which is precisely the problem Nvidia is positioning NemoClaw to solve.
Advertisement
The enterprise demand is not hypothetical. Harrison Chase, founder of LangChain — whose open-source agent frameworks have been downloaded more than a billion times — put it bluntly in a recent episode of VentureBeat’s Beyond the Pilot podcast: “I guarantee that every enterprise developer out there wants to put a safe version of OpenClaw onto onto their computer or expose it to their users.” The bottleneck, he made clear, has never been interest. It has been the absence of a credible security and governance layer underneath it. NemoClaw is Nvidia’s answer to that gap — and notably, LangChain is one of the launch partners for the Agent Toolkit and OpenShell integration.
What NemoClaw actually does — and what it doesn’t replace
NemoClaw is not a competitor to OpenClaw (or the now many alternatives). It is best understood as an enterprise wrapper around it — a distribution that ships with the components a security-conscious organization actually needs before letting an autonomous agent near production systems.
The stack has two core components. The first is Nvidia Nemotron, Nvidia’s family of open models, which can run locally on dedicated hardware rather than routing queries through external APIs. Nemotron-3-Super, scored the highest out of all open models on PinchBench, a benchmark that tests the types of tasks and tools calls needed by OpenClaw.
The second is OpenShell, the new open-source security runtime that runs each claw inside an isolated sandbox — effectively a Docker container with configurable policy controls written in YAML. Administrators can define precisely which files an agent can access, which network connections it can make, and which cloud services it can call. Everything outside those bounds is blocked.
Advertisement
Nvidia describes OpenShell as providing the missing infrastructure layer beneath claws — giving them the access they need to be productive while enforcing policy-based security, network, and privacy guardrails.
For organizations that have been watching OpenClaw’s rise with a mixture of excitement and dread, this is a meaningful development. OpenClaw’s early iterations were, by general consensus, a security liability — powerful and fast-moving, but essentially unconstrained. NemoClaw is the first attempt by a major hardware vendor to make that power manageable at enterprise scale.
The hardware angle: always-on agents need dedicated compute
One aspect of NemoClaw that deserves more attention than it has received is the hardware strategy underneath it. Claws, by design, are always-on — they do not wait for a human to open a browser tab. They run continuously, monitoring inboxes, executing tasks, building tools, and completing multi-step workflows around the clock.
That requires dedicated compute that does not compete with the rest of the organization’s workloads. Nvidia has a clear interest in pointing enterprises toward its own hardware for this purpose.
Advertisement
NemoClaw is designed to run on Nvidia GeForce RTX PCs and laptops, RTX PRO workstations, and the company’s DGX Spark and DGX Station AI supercomputers. The hybrid architecture allows agents to use locally-running Nemotron models for sensitive workloads, with a privacy router directing queries to frontier cloud models when higher capability is needed — without exposing private data to those external endpoints.
It is an elegant solution to a real problem: many enterprises are not yet ready to send customer data, internal documents, or proprietary code to cloud AI providers, but they still need model capability that exceeds what runs locally. NemoClaw’s privacy router architecture threads that needle, at least in principle.
What claws actually look like in the enterprise
Before evaluating the platform, it helps to understand what a claw doing real work looks like in practice. Two partner integrations announced alongside NemoClaw offer the clearest window into where this is heading.
Box is perhaps the most illustrative case for organizations that manage large volumes of unstructured enterprise content.
Advertisement
Box is integrating Nvidia Agent Toolkit to enable claws that use the Box file system as their primary working environment, with pre-built skills for Invoice Extraction, Contract Lifecycle Management, RFP sourcing, and GTM workflows.
The architecture supports hierarchical agent management: a parent claw — such as a Client Onboarding Agent — can spin up specialized sub-agents to handle discrete tasks, all governed by the same OpenShell Policy Engine.
Critically, an agent’s access to files in Box follows the exact same permissions model that governs human employees — enforced through OpenShell’s gateway layer before any data is exchanged. Every action is logged and attributable; no shadow copies accumulate in agent memory. As Box puts it in their announcement blog, “organizations need to know which agent touched which file, when, and why — and they need the ability to revoke access instantly if something goes wrong.”
Cisco’s integration offers perhaps the most visceral illustration of what OpenShell guardrails enable in practice. The Cisco security team has published a scenario in which a zero-day vulnerability advisory drops on a Friday evening.
Advertisement
Rather than triggering a weekend-long manual scramble — pulling asset lists, pinging on-call engineers, mapping blast radius — a claw running inside OpenShell autonomously queries the configuration database, maps impacted devices against the network topology, generates a prioritized remediation plan, and produces an audit-grade trace of every decision it made.
Cisco AI Defense verifies every tool call against approved policy in real time. The entire response completes in roughly an hour, with a complete record that satisfies compliance requirements.
“We are not trusting the model to do the right thing,” the Cisco team noted in their technical writeup. “We are constraining it so that the right thing is the only thing it can do.”
An ecosystem play: the partners behind the stack
Nvidia is not building this alone. The Agent Toolkit and OpenShell announcements came with a significant roster of enterprise partners — Box, Cisco, Atlassian, Salesforce, SAP, Adobe, CrowdStrike, Cohesity, IQVIA, ServiceNow, and more than a dozen others — whose integration depth signals how seriously the broader software industry is treating the agentic shift.
Advertisement
On the infrastructure side, OpenShell is available today on build.nvidia.com, supported by cloud inference providers including CoreWeave, Together AI, Fireworks, and DigitalOcean, and deployable on-premises on servers from Cisco, Dell, HPE, Lenovo, and Supermicro. Agents built within OpenShell can also continuously acquire new skills using coding agents including Claude Code, Codex, and Cursor — with every newly acquired capability subject to the same policy controls as the original deployment.
Separately, Nvidia announced the Nemotron Coalition — a collaborative initiative bringing together Mistral AI, Perplexity, Cursor, and LangChain to co-develop open frontier models. The coalition’s first project is a base model co-developed with Mistral that will underpin the upcoming Nemotron 4 family, aimed specifically at agentic use cases.
What enterprise leaders should be watching
The NemoClaw announcement marks a turning point in how enterprise AI is likely to be discussed in boardrooms and procurement meetings over the next twelve months. The question is no longer whether organizations will deploy autonomous agents. The industry has clearly moved past that debate. The question is now how — with what controls, on what hardware, using which models, and with what audit trail.
Nvidia’s answer is a vertically integrated stack that spans silicon, runtime, model, and security policy. For IT leaders evaluating their agentic roadmap, NemoClaw represents a significant attempt to provide all four layers from a single vendor, with meaningful third-party security integrations already in place.
Advertisement
The risks are not trivial. OpenShell’s YAML-based policy model will require operational maturity that most organizations are still building. Claws that can self-evolve and acquire new skills — as Nvidia’s architecture explicitly enables — raise governance questions that no sandbox can fully resolve. And the concentration of agentic infrastructure in a single vendor’s stack carries familiar platform risks.
That said the direction is clear. Claws are coming to the enterprise. Nvidia just made its bet on being the platform they run on — and the guardrails that keep them in bounds.
Bowers & Wilkins isn’t pretending this is a breakthrough and that’s exactly the point. The British luxury audio brand has expanded its flagship Pi8 true wireless earbuds and Px7 S3 noise-cancelling headphones with a slate of new premium finishes, leaning into a trend that’s been quietly reshaping the high-end audio category: color as innovation. The Pi8 now arrives in Dark Burgundy and Pale Mauve, bringing the total to six finishes, while the Px7 S3 adds a new Vintage Maroon option to its growing lineup.
If that sounds familiar, it should. Last year, I pointed out how a long list of premium audio brands had started treating industrial design and colorways not as afterthoughts, but as a legitimate product cycle strategy; extending relevance without touching the underlying acoustics. Bowers & Wilkins is now fully committed to that playbook. The hardware hasn’t changed and it didn’t need to, but the visual refresh keeps both models firmly in the conversation in a market that’s running out of meaningful spec-sheet upgrades.
Available starting March 19, the new finishes don’t come cheap: $499 for the Pi8 in Pale Mauve or Dark Burgundy, and $479 for the Px7 S3 in Vintage Maroon. Same award-winning sound, new wardrobe. Whether that counts as innovation or just smart business depends on how easily you’re seduced by a better shade of red.
What Are the Bowers & Wilkins Pi8?
Bowers & Wilkins Pi8 in new Dark Burgundy color for 2026
The Bowers & Wilkins Pi8 are the company’s flagship true wireless earbuds, positioned as a no-compromise attempt to deliver genuine high-end sound in a category that usually prioritizes convenience over fidelity. In our review, the Pi8 stand out for their refined tuning, clarity, and sense of control, offering a presentation that feels closer to a compact hi-fi system than a typical pair of wireless earbuds. They’re designed for listeners who actually pay attention to what they’re hearing and not just how easily it connects.
At their core, the Pi8 combine carbon cone drivers, advanced DSP, and support for aptX Lossless to push beyond the limitations that have traditionally defined Bluetooth audio. Bowers & Wilkins also includes a smart charging case with retransmission capability, allowing wired sources to be streamed directly to the earbuds; an unusually practical feature that adds real-world flexibility. It’s a more thoughtful approach than most, focusing on how people actually use their gear rather than chasing feature checklists.
Advertisement
Bowers & Wilkins Pi8 in new Pale Mauve color for 2026
That said, the Pi8 don’t try to win on every front. As we noted in our review, the emphasis is clearly on sound quality, materials, and overall refinement, rather than class-leading noise cancellation or mass-market pricing. If overall sound quality, comfort, and strong but not class leading ANC matter most, these are among the best options available.
What Are the Bowers & Wilkins Px7 S3?
The Bowers & Wilkins Px7 S3 are the brand’s latest over-ear wireless noise-cancelling headphones, sitting just below the Px8 S2 but very much aimed at the same crowd that shops Sony, Bose, and Sennheiser at the top of the category. In our review, they come across as a deliberate refinement of what Bowers & Wilkins has been building for over a decade; premium materials, a more mature design language, and a clear focus on sound quality first. This isn’t a lifestyle headphone trying to fake it. It’s a high-end hi-fi product that just happens to be wireless.
Bowers & Wilkins Px7 S3 in new Vintage Maroon color for 2026
Where the Px7 S3 separates itself is in how it sounds relative to its competition. As noted in the review, it delivers audiophile-grade clarity, deep and controlled bass, and a level of detail that outpaces most rivals in this price range, including the usual suspects from Sony, Bose, and Apple. Bowers & Wilkins has also refined the internal driver design and overall tuning, while adding modern essentials like aptX Lossless and a more flexible EQ. The result is a presentation that feels more composed and revealing than what you typically get from mainstream ANC headphones.
That said, like the Pi8, the Px7 S3 doesn’t try to dominate every category. The review makes it clear that while ANC is improved and competitive, it’s not the class leader, and comfort is very good without being the lightest or most effortless in the segment. This is a headphone built around priorities: sound quality, build, and long-term listening satisfaction. If that’s what matters most, it’s one of the strongest all-around options available right now and one of the few that still feels like it was tuned by people who actually prioritise sound quality over ANC and connectivity features.
The Bottom Line
New colors are not innovation, but they do make the Pi8 and Px7 S3 feel fresher and harder to ignore. More importantly, this kind of refresh signals longevity these models are not going anywhere. Same excellent sound, now with a little more swagger.
Where to buy:
Tip: These new finishes add to the existing colors which include back, white, blue and jade green. Currently the new colors are only available at the Bowers & Wilkins website.
The U.S. Supreme Court won’t touch the debate over whether drivers have a free-speech right to put whatever they want on their license plates. Back in December, the justices refused to hear an appeal from a Tennessee woman who had her controversial vanity plate revoked. In doing so, they left in place the legal framework that has existed for years in many jurisdictions: vanity plates are government speech, not a means of individual expression.
This whole thing started when Tennessean Leah Gilliam had her plate reading “69PWNDU” revoked after driving around with it for more than a decade. After originally approving it, Tennessee officials eventually got enough complaints to realize the message was referencing something sexual. For the record, Gilliam argued it was actually a reference to the year of the moon landing alongside some gaming slang. She sued, lost, and tried to take it all the way to the Supreme Court.
Alas, the Supreme Court’s decision not to make a decision still leaves vanity plates in the hands of lower court rulings. That means it’s up to the states to regulate what appears on plates. Unfortunately for Gilliam, it also means it’s time to start brainstorming something for her next regrettable vanity license plate.
Advertisement
The law hasn’t always been clear on vanity plates
RYO Alexandre/Shutterstock
The Supreme Court’s decision makes it all seem pretty cut and dried, but looking at past rulings, that’s not exactly the case. The Supreme Court declined to hear the case, letting the lower court’s decision stand. But those lower courts haven’t always been on the same page. In recent years, they’ve reached different conclusions on who ultimately has the authority over vanity plates.
One example: In a 2020 case in Rhode Island, a federal judge ruled that a ban on NSFW license plates would violate the First Amendment because it gave state officials overly broad discretion to reject messages. The court allowed him to keep his vulgar plate throughout the litigation, saying that revoking it would suppress his personal expression. That goes against what Tennessee decided in Gilliam’s case. The plate in question read “FKGAS,” which is pretty much on par with other rejected vanity plates from states like Illinois.
Advertisement
Not even the Supreme Court can get aligned on this. In the 1977 Supreme Court case Wooley v. Maynard, the justices decided that individuals can’t be forced to display ideological state slogans on their plates. That would suggest First Amendment protection (however limited). But in 2015’s Walker v. Sons of Confederate Veterans, the Supreme Court’s ultimate decision was that specialty plates were government speech. For now, the Supreme Court clearly has very little interest in ruling on the issue one way or the other.
Shortly after Trump took office for a second time, his administration made it clear that it felt constitutional rightswere merely privileges it would extend only to those who fully supported whatever the hell the administration happened to be doing.
At the time, the administration was not only engaged in a full-blown, bigoted war against migrants, but throwing all of its support behind Israel’s ongoing anti-Palestinian efforts, which look a whole lot like actual genocide.
U.S. Citizenship and Immigration Services has announced it will begin screening immigrants’ social media for evidence of antisemitic activity as grounds for denying immigration benefit requests. The screenings will affect people applying for permanent residence status as well as foreigners affiliated with educational institutions. The policy will go into effect immediately.
In a statement issued Wednesday morning, the Department of Homeland Security said it will “protect the homeland from extremists and terrorist aliens, including those who support antisemitic terrorism, violent antisemitic ideologies and antisemitic terrorist organizations such as Hamas, Palestinian Islamic Jihad, Hezbollah, or [the Houthis].”
Advertisement
But, in true MAGA fashion. Trump’s anti-antisemitic efforts only affected people who were more brown than white. A month later, Trump was opening the immigration door to South African “refugees,” but only the white ones. This decision was presumably based on lizard brain analysis of out-of-context clips shown on social media and/or Fox News that pretended whites in South Africa were being persecuted by the Black residents they’d persecuted for decades under apartheid.
And that included white South Africans who engaged in antisemitic speech, who were given a free pass to play their version of the race card to gain unvetted admittance to the Land of Opportunity.
Nine months later, the doors have been thrown wide open for white South Africans, with the administration yet again claiming — without facts in evidence — that these particular South Africans were more deserving of expedited asylum proceedings than anyone from any country where actual violence and persecution targets residents who are not white enough for the administration to consider worthy of naturalization.
The U.S. aims to process 4,500 refugee applications from white South Africans per month, far above President Donald Trump’s stated refugee program cap, and is installing trailers on embassy property in Pretoria to support the effort, a U.S. contracting document said.
While both versions of the Trump administration may have rendered satire mostly obsolete, it can’t eradicate irony. And here’s where it gets absolutely hilarious. White South Africans are now thinking their homeland is a better option than living here under this administration, as Reuters reports.
Advertisement
Andrew Veitch left South Africa after being held up at gunpoint in his car. But now he feels there are greater threats in the United States, he said, citing mass shootings in public places as well as violence by U.S. immigration officers.
“People are being shot in broad daylight. American citizens are being shot and killed,” said the 53-year-old, who moved to California in 2003. “I don’t want to live in a place like this.”
It’s a valid point — one that people who have lived here for their entire lives are making, albeit without the easily available option of just pressing CTRL-Z on their temporary protected status. Veitch isn’t the only one wanting to return to the allegedly-hyper-violent country of South Africa, rather than continue to live in the Land of Opportunity that is daily being rebranded as the Land of Impending Martial Law.
Other South African (white) “refugees” are heading back home because the financial climate is preferable, even if they choose to ignore the threat Trump poses to every freedom Americans hold dear. After 20 years of US residence, it’s the Trump administration that’s encouraging South African (white) migrant Naomi Saphire to return to her homeland.
[Saphire] had been settled in the United States for two decades when she came back for a holiday and realized how much she missed home.
Last year, she left North Carolina for a seaside town in South Africa’s Western Cape, where she said her three children spend more time outdoors, health insurance is affordable and she prefers the schools.
Advertisement
“My heart is just full of gratefulness to be here,” the 46-year-old said from her home in Plettenberg Bay. “The U.S. has been really good to me (but) I just felt like I was depriving my kids of this life.”
As for the supposed violence targeting white South Africans Trump is now pretending to “save” from their misery with his “let’s get a bunch more whites in here” immigration policies, it’s as mythical as divorced from reality as Trump’s self-perception as the smartest, savviest businessman/politician to ever grace the Earth with his presence:
Crime and joblessness are major issues in South Africa, but the unemployment rate is 35% for Black people compared with 8% for whites, according to the latest figures from the national statistics agency Stats SA.
Police statistics released last year showed that even farm murders, which Trump has focused on, killed more Black people than whites. Reuters has found that photos and videos Trump has presented on the matter were taken out of context or misrepresented.
While this might seem like the most useless of anecdotal evidence, there’s reason to believe white South Africans will either ignore the Trump’s invitation to further whiten the US, or head back home where things are still pretty fucking good for whites, but without having to deal with a megalomaniac chaos agent who seems to believe World War III will finally allow Truth Social to turn a profit.
Advertisement
White South Africans have been rejecting the United States since 2022, when a law allowed them to regain their citizenship after it was stripped by a post-apartheid law passed in 1995. 15,000 white South Africans took advantage of that to return to their homeland. Now that Trump’s back in office, even more whites are exiting than entering the United States, despite the administration’s warm welcome of white migrants into its white Christian nationalism plans.
Home Affairs Minister Leon Schreiber said 1,000 people had reclaimed their citizenship, a number he expected to grow significantly as the programme takes off.
One of the main reasons for exiting the US was “lower cost of living,” something Trump has actively worked against achieving, starting with his indiscriminate (and apparently unlawful) tariffs and continuing through his destabilization of the world economy by (1) being buddies with Russia, (2) starting yet another forever war in the Middle East, (3) refusing to engage honestly with inflation and joblessness, and (4) decreasing American productivity by forcibly ejecting hundreds of thousands of people who work harder, pay more in taxes, and commit fewer crimes than US citizens.
And now it’s clear this administration is so inept and inherently dangerous it can’t even convince white people to live here. Let that sink in for a bit.
But sir, it is wafer-thin. That’s how they get you! Just when you couldn’t possibly justify building another keyboard, let alone owning one, along comes the Kambala by [aroum2].
Now, ‘Kambala’ means a few things, but here it refers to fish, as evidenced by logo and matching themed PCB key chain shown in the gallery.
This catch is so flat because of the switches: PG1316S, and 42 of them. These are better known to some as Kailh butterfly switches, and are meant for laptops. But, this is Hackaday.
No matter what you call them, those switches are controlled by a nice!nano V2-compatible controller, which allows for ZMK firmware support. There’s a 110 mAh battery and four status LEDs, and best of all, the charging indicator is in the fish’s eyes.
[aroum2] might share the files later. Here’s hoping!
Advertisement
Let’s Talk DIY Palm Rests
Palm rests! Depending on the keyboard, they can be built right in. This here Kinesis Advantage comes to mind. That said, you can buy a pair of nice adhesive pads for your Kinesis once the ABS shine starts to bother you, or better yet, before that happens. Don’t make your own out of adhesive foam sheets. Just, trust me on this.
But oftentimes, especially with travel keyboards, palm rests aren’t included. And that’s fair, because people want different things. Before you go printing some, or even rendering a pair from zebrawood, consider cheap alternatives like a large car-washing sponge cut in half and covered with the fabric of your choice.
On the slightly more expensive side, many employ a pair of Purple mattress samplers, which have doubled in price since I bought some 2022, but are still worth it.
Depending on your desk, you could do something as simple as cutting a pool noodle in half and shoving it onto the edge. Maybe you’ve done something even more temporary that turned out to be permanent. Tell me in the comments!
Squishy. Image via Purple
Even if you have built-in palm rests, sometimes you need to temporarily insert something like a spiral notebook between your desk edge and keyboard, pushing the thing further away and putting your delicate elbows at risk. This is me right now, and each elbow is on a mouse bag. Simple and effective.
Another consideration is attached versus unattached. I mean, if a travel keyboard is going to have palm rests, they should attach rather than just be placed in front. Maybe that’s just me.
Advertisement
The Centerfold: Telegraph Key Macro Pad is Dashing
The system works! [Colin] sent a tip about his Telegraph Key Macro Pad, which is exactly what it sounds like. [Colin] says that his job these days mostly consists of copy/pasting from GPT, and it was quickly becoming a pain in the wrist. (Boy, can I relate.)
Using the thing is just as it should be: to copy, you long press the key like a Morse code dash. To paste, you do the short one. This enables [Colin] to paste many times, and quickly. [Colin] started with a Soviet-era telegraph key from the electronic bay, and a Pimoroni Tiny 2040 programmed with Arduino. It may be wildly overpowered for the application, but hey, it fits nicely in the base of the telegraph key.
The default is to make a sound when you do either action. [Colin] used a piezo disk so that it can handle different tones. This was done mostly for the luls, but it also lets him know when something is copied. There’s also a nifty silent mode that moves the mouse cursor in a quick loop-dee-loo when the deed is done.
Do you rock a sweet set of peripherals on a screamin’ desk pad? Send me a picture along with your handle and all the gory details, and you could be featured here!
Advertisement
Historical Clackers: the Crown Was a Machine for the Millions… Not!
You might wonder why I choose so many index typewriters for this portion of the program. I suppose it’s because they can be so differently designed, yet serve the same purpose. And that’s just cool to me.
Image via The Antikey Chop
The Crown index typewriter is no exception. Let’s start with the fact its creator, Byron Alden Brooks, was a celebrated inventor of early typewriters. You may have heard of the Brooks; he also had a hand in the People’s, the National, the Travis, and of course, the Crown index typewriter. Perhaps most unforgettable among his accomplishments, Brooks invented the Shift key.
The Crown was produced between 1888 and 1894, though it is thought that Brooks began work on it as early as 1881, evidenced by the date on the typewheel patent. It’s also thought that production really ceased in 1893.
That’s right, the Crown used a typewheel and a linear index from which the user selected each character. The ink came from a felt roller situated between the carriage and typewheel. Every time a character was selected, this roller would swing out of the way so the typewheel could strike the platen.
Originally, the Crown cost $20 (about $700 today), with the wooden case thrown in free. The price dropped to $16 by the middle of 1891. Despite being billed as ‘a machine for the millions’, the Crown was a failure.
Advertisement
Finally, There’s a Quiz To Find Your Switch Type
If you’re really up on things, you’re of course no stranger to KBD News and the corresponding newsletter. KBD is a great resource for all things keyboard, and now there’s a switch compatibility quiz to help you get started.
Image via KBD News
Of course, not all switches work with all PCBs, so you can’t begin this journey without knowing which path to head down. Choose MX, and you’ll have a bevvy of beauties to choose from. There are far fewer low-profile and Hall-effect switches out there, so keep that in mind.
Let’s say you go down the MX path. Your next choice is important: how much feedback do you need? None? A little? An audible click? Remember to keep your environment in mind.
If you’re me, you choose clicky. Now it’s time to think about actuation force. There are no light-force clicky switches; it’s just not a thing. So you can choose mid, heavy, or no preference, which takes you directly to RGB choices. Do you want a transparent housing? A light diffuser? Both? If you have no preference here, your final choice concerns factory lubrication. I ended up with 10 different switch recommendations, but of course, YMMV.
When an AI agent needs to log into your CRM, pull records from your database, and send an email on your behalf, whose identity is it using? And what happens when no one knows the answer? Alex Stamos, chief product officer at Corridor, and Nancy Wang, CTO at 1Password joined the VB AI Impact Salon Series to dig into the new identity framework challenges that come along with the benefits of agentic AI.
“At a high level, it’s not just who this agent belongs to or which organization this agent belongs to, but what is the authority under which this agent is acting, which then translates into authorization and access,” Wang said.
How 1Password ended up at the center of the agent identity problem
Wang traced 1Password’s path into this territory through its own product history. The company started as a consumer password manager, and its enterprise footprint grew organically as employees brought tools they already trusted into their workplaces.
“Once those people got used to the interface, and really enjoyed the security and privacy standards that we provide as guarantees for our customers, then they brought it into the enterprise,” she said. The same dynamic is now happening with AI, she added. “Agents also have secrets, or passwords, just like humans do.”
Advertisement
Internally, 1Password is navigating the same tension it helps customers manage: how to let engineers move fast without creating a security mess. Wang said the company actively tracks the ratio of incidents to AI-generated code as engineers use tools like Claude Code and Cursor. “That’s a metric we track intently to make sure we’re generating quality code.”
How developers are incurring major security risks
Stamos said one of the most common behaviors Corridor observes is developers pasting credentials directly into prompts, which is a huge security risk. Corridor flags it and sends the developer back toward proper secrets management.
“The standard thing is you just go grab an API key or take your username and password and you just paste it into the prompt,” he said. “We find this all the time because we’re hooked in and grabbing the prompt.”
Wang described 1Password’s approach as working on the output side, scanning code as it is written and vaulting any plain text credentials before they persist. The tendency toward the cut-and-paste method of system access is a direct influence on 1Password’s design choices, which is to avoid security tooling that creates friction.
Advertisement
“If it’s too hard to use, to bootstrap, to get onboarded, it’s not going to be secure because frankly people will just bypass it and not use it,” she said.
Why you cannot treat a coding agent like a traditional security scanner
Another challenge in building feedback between security agents and coding models is false positives, which very friendly and agreeable large language models are prone toward. Unfortunately, these false positives from security scanners can derail an entire code session.
“If you tell it this is a flaw, it’ll be like, yes sir, it’s a total flaw!” Stamos said. But, he added, “You cannot screw up and have a false positive, because if you tell it that and you’re wrong, you will completely ruin its ability to write correct code.”
That tradeoff between precision and recall is structurally different from what traditional static analysis tools are designed to optimize for, and it has required significant engineering to get right at the latency required, on the order of a few hundred milliseconds per scan.
Advertisement
Authentication is easy, but authorization is where things get hard
“An agent typically has a lot more access than any other software in your environment,” noted Spiros Xanthos, founder and CEO at Resolve AI, in an earlier session at the event. “So, it is understandable why security teams are very concerned about that. Because if that attack vector gets utilized, then it can both result in a data breach, but even worse, maybe you have something in there that can take action on behalf of an attacker.”
So how do you give autonomous agents scoped, auditable, time-limited identities? Wang pointed to SPIFFE and SPIRE, workload identity standards developed for containerized environments, as candidates being tested in agentic contexts. But she acknowledged the fit is rough.
“We’re kind of force-fitting a square peg into a round hole,” she said.
But authentication is only half of it. Once an agent has a credential, what is it actually allowed to do? Here’s where the principle of least privilege should be applied to tasks rather than roles.
Advertisement
“You wouldn’t want to give a human a key card to an entire building that has access to every room in the building,” she explained. “You also don’t want to give an agent the keys to the kingdom, an API key to do whatever it needs to do forever. It needs to be time-bound and also bound to the task you want that agent to do.”
In enterprise environments, it won’t be enough to grant scoped access, organizations will need to know which agent acted, under what authority, and what credentials were used.
Stamos pointed to OIDC extensions as the current frontrunner in standards conversations, while dismissing the crop of proprietary solutions.
“There are 50 startups that believe their proprietary patented solution will be the winner,” he said. “None of those will win, by the way, so I would not recommend.”
Advertisement
At a billion users, edge cases are not edge cases anymore
On the consumer side, Stamos predicted the identity problem will consolidate around a small number of trusted providers, most likely the platforms that already anchor consumer authentication. Drawing on his time as CISO at Facebook, where the team handled roughly 700,000 account takeovers per day, he reframed what scale does to the concept of an edge case.
“When you’re the CISO of a company that has a billion users, corner case is something that means real human harm,” he explained. “And so identity, for normal people, for agents, going forward is going to be a humongous problem.”
Ultimately, the challenges CTOs face on the agent side stem from incomplete standards for agent identity, improvised tooling, and enterprises deploying agents faster than the frameworks meant to govern them can be written. The path forward requires building identity infrastructure from scratch around what agents actually are, not retrofitting what was built for the humans who created them.
If you’re ready to move your PC gaming experience to the next level, look no further than GIGABYTE’s X870E AORUS ELITE X3D.
Already known for their forward-thinking tech, GIGABYTE has outdone themselves with this one. Boasting state-of-the-art graphics capabilities, AI-enhanced X3D Turbo Mode 2.0, Zenith Memory Performance, and a comprehensive thermal design, X870E AORUS ELITE X3D offers everything you need to up your game.
Top-tier graphics performance for a competitive edge
The X870E AORUS ELITE X3D motherboard packs all the power of the AMD Radeon RX 9070 for top-of-the-line gaming graphics performance. If you’re an esports player, you’ll get the super-fast refresh rates you need to keep up with the competition. Or, if you’re into games with massive open-world maps, the X870E AORUS ELITE X3D will keep you exploring as far as the eye can see and beyond.
Advertisement
Switching between gaming and performance modes
GIGABYTE’s X870E AORUS ELITE X3D features X3D Turbo Mode 2.0, which offers an overall enhanced experience over the previous generation. Whether you’re gaming, multitasking, or a little of both, X3D Turbo Mode 2.0 delivers with a built-in AI model that automatically optimizes parameters in real time. You can switch easily between two modes, Extreme Gaming Mode & Max Performance Mode, to make sure you’re getting peak performance suited to your task.
Advanced DDR5 memory and AI-enhanced overclocking
Using a combination of advanced technologies, X870E AORUS ELITE X3D offers Zenith Memory Performance, an advanced AI-enhanced overclocking technology for DDR5 memory. To start, the PCB features advanced shielding to ensure clean and clear memory signals. The next innovation is daisy-chain routing, designed to remove signal bottlenecks. Finally, the PCB itself uses 8-layer server-grade materials to ensure DDR5 data is transmitted at blazing speeds.
Stay frosty with smart cooling and thermal guards
All this power demands cooling tech that’s up to the task of keeping your system running cool and efficiently. GIGABYTE’s X870E AORUS ELITE X3D has what it takes, with a full-metal thermal design and durable heat sinks. Add to this M.2 EZ-Flex, GIGABYTE’s exclusive patented design, plus their quiet and efficient Smart Fan 6 technology, and you can be sure your machine will run smoothly for hours on end.
With its top-of-the-line graphics capabilities, Turbo mode 2.0, Zenith Memory Performance, and compressive thermal design — not to mention its durable materials and ultra-connectivity — GIGABYTE’s X870E AORUS ELITE X3D should be at the heart of any true gamer’s rig.
You must be logged in to post a comment Login