Is Marvel Tokon: Fighting Souls one of the best fighting games of recent times? Yes, I would say so. Given its developer is Arc System Works — creators of Guilty Gear, Blazblue, and the excellent Granblue Fantasy Versus: Rising — it was a pretty safe bet that we’d be served a high quality fighting game with mountainous respect and reverence for Marvel and its roster of supers, mutants, and weirdos.

I love fighting games, but I’m not usually one for tag fighters — often because I find learning more than one character intimidating and time-consuming. So, when Marvel Tokon: Fighting Souls was first revealed last year as a 4v4 tag fighter, I’d already resigned myself to thinking: ‘well, that’s probably not for me.’

Parents will no longer have to rely on texts or phone calls to know how their teens’ Uber rides are going. Instead, they’ll be able to see for themselves.

Uber on Tuesday launched a live video streaming feature for teen accounts that lets parents or guardians check in on how a trip is going in real time. It uses the selfie camera on the driver’s phone, so family members can keep an eye on anything happening in the vehicle.

The live video feature is being piloted in nine areas: Atlanta; Phoenix; Orlando, Florida; Raleigh-Durham, North Carolina; Portland, Oregon; El Paso and San Antonio, Texas; Cleveland; and Rhode Island. It’s slated to roll out across the US in the coming weeks, Uber says.

“The feature is designed for quick check-ins, giving parents another way to feel confident everything on their teen’s trip is going smoothly,” the company said in a release.

Uber teen accounts launched in 2023, with the goal of making it easier for parents and guardians to keep tabs on their kids’ rides. They can follow along in the app to track a trip or delivery, and receive real-time status updates. It’s one of several features Uber and other ride-hailing platforms have rolled out over the years to assuage concerns about rider safety. That includes audio recordings and the option to easily favorite and block drivers on Lyft.

To add an Uber teen account, go into the app and tap Account > Family to create a family profile. Then tap Invite family > Teen > Choose contact and tap on their name from your contacts. Your teen can then request a ride, so long as your payment information is up to date.

To maintain privacy, only a teen’s linked parent or guardian can watch a live stream during a ride. That video is also encrypted, Uber says, so that “neither the driver, guardian or teen has access to the recording of the live video stream after the trip ends.”

Ride-hailing companies have also rolled out features catered to older riders. Last year, both Uber and Lyft simultaneously debuted simplified versions of their apps with larger text and live phone support, in an attempt to boost ridership among adults over 65 years old.

Abrar Al-Heeti

Senior Technology Reporter

Abrar’s interests include phones, streaming, autonomous vehicles, internet trends, entertainment, pop culture and digital accessibility. In addition to her current role, she’s worked for CNET’s video, culture and news teams. She graduated with bachelor’s and master’s degrees in journalism from the University of Illinois at Urbana-Champaign. Though Illinois is home, she now loves San Francisco — steep inclines and all.

See full bio

Variability in different math libraries, which are used when audio is produced through browsers, were once high enough to provide a massively large number of uniquely different signatures when combined with different CPUs and other system differences. After the audio soundprinting technique became well known, Firefox implemented a fix, starting in version 118 released in 2023. Beginning then, the browser used its own unique math libraries, rather than relying on ones that shipped with the OS it ran on. Tom Ritter, a Firefox developer who has also volunteered for the Tor Project, said the move to constant libraries reduced the entropy enough to stop the technique from working.

The technique is ineffective in Chrome because it, too, ships with its own libraries, a Google spokesperson said. Safari users are likely safe for the same reason, although Apple didn’t immediately confirm that.

An artifact from years past

AliExpress’ use of an obsolete fingerprinting method raises the question: Why is it bothering to use it? The likely answer has to do with the more than a dozen other fingerprinting methods the online retailer is using. Callaghan said the other techniques included:

• canvas rendering and toDataURL()
• WebGL renderer information, extensions, and shader precision
• audio oscillator and analyzer output
• screen and viewport dimensions
• device pixel ratio
• hardware concurrency and device memory
• installed browser plugins
• supported audio and video formats
• WebRTC behavior
• browser performance timing
• mouse, touch, focus, and scroll events
• device motion and orientation
• properties commonly associated with browser automation

It’s likely the web audio trick is left over from years earlier and no one noticed it until now.

It’s comforting to know that browser makers have taken precautions. At the same time, it’s not yet clear how effective the rest of the metrics AliExpress is using are. Less reassuring still: it’s almost a certainty that thousands of sites are employing similar tracking. Browser developers are in a dynamic race with site publishers, who are constantly looking for new ways to break the latest protections.

Security teams have spent years hardening authentication, with controls like multi-factor authentication (MFA) and conditional access now commonplace. While stronger authentication can make traditional credential theft less effective, it doesn’t solve every identity problem.

There are several points in the identity lifecycle where trust is established or re-established:

• When a new employee joins.
• When someone loses access to their account.
• When a password or MFA factor needs to be reset.
• When the service desk is asked to make a sensitive change to an account.

Rather than stealing credentials or bypassing MFA, an attacker can instead try to convince the service desk that they are the account holder, using social engineering to exploit legitimate processes.

That puts greater pressure on organizations to secure both the login as well as the processes around account creation and recovery.

How Attackers Exploit Identity at Onboarding and Recovery

In late July 2026, the US Department of State and allies including Japan, Canada and the UK issued a joint alert warning that North Korean IT workers were impersonating foreign nationals to secure employment.

Their tactics focus on falsifying identity documents, such as using images supplied by a third party based in another country to register accounts. The North Korean then carries out the actual work.

These workers typically target technology companies, so the important point is not that every organization should expect the same type of campaign. It is that onboarding creates a moment when trust is established for the first time.

If identity checks fail at that stage, the attacker can enter the environment with access that appears legitimate.

The same issue can occur during the recovery process. Threat actor groups like Scattered Spider are proficient at social engineering, impersonating employees and calling the service desk to reset passwords that gift access to an account.

This tactic was linked to the 2025 M&S ransomware breach, which contributed to an estimated $400 million hit to the retailer’s operating profit through lost sales.

The security question in these scenarios is the same: how confidently can the organization verify that the person making the request is who they claim to be?

Strong Authentication Still Depends on Strong Identity Checks

When someone calls the service desk claiming that they’ve forgotten their password or lost access to their authenticator, the agent needs to be able to confidently verify the person calling is the real account owner.

However, in many organizations identity checks can still rely on relatively weak signals. A service desk might ask for an employee ID or phone number. Security questions are still common, asking the caller the name of their first pet or where they went to school.

The problem is that many of these checks can be researched, stolen or manipulated. Attackers can find personal information through data breaches or social media.

Even in instances where stronger checks are in place, the North Korean remote worker campaigns demonstrate how documents and other identity evidence can be altered or fabricated.

AI is making impersonation more convincing, too. Attackers can use synthetic profiles, manipulated images, cloned voices and deepfake video to support a false identity or make a social engineering attempt more believable.

All of these make it harder for agents to act with confidence. As it’s vital for organizations to verify users during onboarding and recovery events, they need stronger measures to deliver that verification.

Strengthen Verification During High-Risk Identity Events

Solutions like Specops Verified ID add another layer of assurance and helps service desk agents confidently confirm identity before sensitive actions take place.

It does this by combining government document scanning and validation with biometric liveness detection.

Document checks help confirm that the ID being presented is legitimate, while liveness detection helps verify that a real, present person is completing the process rather than relying on a static image or other replayed evidence.

During onboarding, this gives organizations a stronger way to verify new employees before granting access to corporate systems. That can reduce the risk posed by fraudulent applicants and impersonation attempts, including tactics seen in North Korean remote worker campaigns.

The same approach can be applied when high-assurance verification is needed, such as password resets for privileged accounts.

Rather than adding complexity to every identity event, Specops Verified ID applies stronger verification where the consequences of getting it wrong are highest.

Protect Your Service Desk with Specops

Strong authentication remains essential, but attackers will continue looking for ways around the controls that are hardest to break. Increasingly, that means targeting the processes used to establish or recover identity rather than attacking the login itself.

Whether an organization is onboarding a new employee or helping an existing one recover their account, the challenge is ensuring the right person is granted access.

Specops Verified ID adds government ID validation and biometric liveness detection to these high-risk identity events, helping organizations make that decision with greater confidence.

If you’re interested in learning more about how Specops can strengthen identity verification at the service desk, contact us today to speak to an expert.

Sponsored and written by Specops Software.

Perplexity is launching Portable Computer today, a version of its agentic “Computer” platform that runs entirely on hardware users already own — starting with Nvidia’s DGX Spark desktop supercomputer and Linux machines equipped with Nvidia RTX GPUs.

The launch, developed in close partnership with Nvidia, is one of the most aggressive attempts yet to move serious AI agent workloads off the cloud and onto local devices. The model, the user’s files, and the work itself can all stay on the machine. Work completed locally consumes no billing credits, and the company says every task starts on the device by default — with the system asking permission before sending any individual step to a more powerful frontier model in the cloud.

“We’ve basically brought the exact same UI to a fully local app,” said Nate, Perplexity’s vice president of engineering for infrastructure and enterprise, during a press briefing Monday. “This incorporates the entirety of the agent harness and inference and everything needed to do work locally.”

For Nvidia, which has spent the past two years selling the world on trillion-dollar AI data centers, the announcement signals something subtler but strategically important: the chipmaker believes local AI has crossed a threshold from hobbyist curiosity to practical tool — and it wants to sell the hardware that runs it.

“Local AI reached an inflection point,” said Nader, Nvidia’s director of developer technology, who focuses on developer tooling and open source. “For the longest time, it was hobbyists and enthusiasts, and they were running these quantized models that were quantized down to be super tiny… And while that’s cool, it’s not super practical. But all that changed with a lot of these new open source models that have come out that are super useful.”

How Portable Computer packages a full local AI stack into a single app

Perplexity Computer, the company’s agentic platform for knowledge work, orchestrates AI models, files, tools, and web access to complete multi-step tasks — reviewing folders of documents, analyzing data, producing reports, and pushing results into business systems. Portable Computer replicates that experience locally: the local models, agent harness, inference engine, tools, app connectors, and a security sandbox come packaged together in a single system. That bundling is the point. With most local AI stacks today, users must assemble and operate those pieces separately — downloading model weights, standing up an inference server, wiring together tools, and tuning performance.

“Historically it’s just been really painful to bring up the local AI stack,” Nate said. “With Portable Computer, we really focused on just making this a really straightforward experience where you can get up and running very quickly.”

In one demo Monday, the system played the role of a retail investor reviewing a folder of 1099s and investment documents — the kind of sensitive financial material many users would hesitate to upload to a cloud service. Running a 27-billion-parameter Qwen model at full GPU utilization on a DGX Spark, the agent reviewed each document and flagged cases where the hypothetical investor was paying unnecessary fees. The interface element that normally displays a running tally of cloud credits “is just parked at zero,” Nate noted, “because all of this is happening on the device.”

A second demo showed the hybrid side of the product. Playing a startup founder, Nate asked the agent to analyze a CSV of user funnel data locally, then push the finished analysis to a Slack channel using Perplexity’s connector ecosystem — proof that local-first does not mean disconnected.

The system also connects to Google Drive, Gmail, and GitHub, and can escalate to a frontier cloud model when the local model hits its limits. At launch, users can set up Qwen 3.8 27B or PPLX 27B, a version Perplexity has post-trained on its own harness, with Nvidia’s Nemotron 3.5 Lightning coming soon.

Portable Computer arrives today for Pro, Max, Enterprise Pro, and Enterprise Max subscribers on Linux, with Windows support following in September. Any RTX GPU with at least 24GB of VRAM — roughly a GeForce RTX 3090 or newer — clears the bar, a threshold Nate called “sort of the floor where we really want to make sure that we can deliver a great experience, but balance that with making it broadly available.”

Why co-designing the model and agent harness beats general-purpose frameworks

Alongside the launch, Perplexity published a research paper arguing that effective local agents require the model and the agent harness — the scaffolding of prompts, tools, and orchestration logic around the model — to be designed together. The core insight: general-purpose harnesses assume a frontier model that can absorb enormous contexts, navigate sprawling tool surfaces, and plan over long horizons. Small local models buckle under those demands.

Perplexity found empirically that although models like Qwen 3.8 27B advertise 260,000-token context windows, they begin to struggle beyond 100,000 tokens. So the company built a deliberately minimal harness: a succinct system prompt, a small set of core tools, and capabilities that load and unload as on-demand “skills” rather than sitting permanently in context. It converted popular connectors like Gmail and GitHub from token-hungry MCP servers into compact command-line tools, added self-verification hooks that monitor the health of a task, and enforced always-on OS-level sandboxing. If the sandbox is unavailable, the harness disables itself rather than running tools unprotected — a contrast with open-source harnesses that run commands with the user’s full permissions by default.

The benchmark results Perplexity reports are striking, though they come from the company’s own evaluations. On its internal Local Knowledge Work Bench — 53 tasks spanning deep research, financial analysis, and document creation, which Perplexity says it plans to open-source — Computer running Qwen 3.8 27B on a DGX Spark scored 82.6%, versus 77.6% for the open-source Pi harness and 74.0% for Hermes running the identical model. 

Perplexity’s post-trained PPLX 27B pushed the score to 85.4%. The gaps widen dramatically on harder tasks: on BrowseComp, a web research benchmark, Computer hit 66.7% accuracy versus 50.2% for Pi and 43.9% for Hermes, while using 51% less wall time and 70% fewer tokens than Pi. On multimodal document understanding, Computer scored 65.1% against Hermes’ 34.6% and Pi’s 13.9%.

The token economics driving AI agents from the cloud to local hardware

The strategic logic behind the launch becomes clear when you consider how AI workloads have changed. Chat was bursty — a question, an answer, done. Agents are different.

“With agents, you want these agents always on if you can. You want the agents to really consume as many tokens as they can,” Nader said. “What we’re seeing is an insatiable demand for tokens, and that’s something that makes local AI so great. As you saw through all these demos, you were not metered by the token. You were not paying for the token. So it’s really killer for agents.”

This reframes the value proposition of local hardware. An agent that runs for hours reviewing documents, verifying its own work, and iterating on analyses would rack up substantial API bills in the cloud. On a device the user already owns, the marginal cost of those tokens approaches zero. Perplexity’s paper makes the enterprise version of this argument explicitly: as agents scale across individual workflows and entire organizations, token expenditure and data movement “become increasingly difficult to govern.” Local-first execution addresses both at once — spend, because inference is free, and privacy, because sensitive tokens never leave the device boundary.

Perhaps the most commercially interesting result concerns the hybrid middle ground. On Terminal Bench 2.1, a challenging coding benchmark, the fully local Qwen model scored 59.6% at essentially zero marginal cost. Letting it escalate to a Claude Opus 5 “advisor” in the cloud raised the score to 73.0% at an estimated $0.415 per task. Running the frontier model alone scored 82.4% at $0.65 per task. Escalation, in other words, recovered roughly three-fifths of the gap to frontier performance at about two-thirds of the cost — and the user decides when that trade is worth making. Before any advisor call, the harness runs a PII classifier over the outgoing context and shows the user exactly what would leave the device. The remote model returns text guidance only; it never touches local files or tools.

Where Portable Computer fits against Ollama and the DIY local AI stack

Jason Hiner of The Deep View pressed the companies on how Portable Computer relates to existing local inference tools like Ollama. Nate’s answer drew a clear line: the tools solve different layers of the problem.

“The majority of the effort here has been at the agent harness level,” he said, noting that the system uses vLLM to host model inference underneath, with an advanced mode for users who want to plug in their own inference endpoint. “We’ve heavily post-trained both the Qwen and Nemotron models that we’re working with in order to really get the best possible results… Our focus has been on really honing the whole stack, top to bottom, of the model inference and the harness together.”

Nader put it more colorfully. “Just getting inference running really quickly on a Spark — there’s a smooth path. You can use Ollama. You can get that set up. But then, as you start to do more complicated, more agentic things, then suddenly you need more perf. You start looking at different models. You start looking at different harnesses, and it’s kind of like the ocean. The deeper you go, the deeper it gets.”

The appliance-like pitch appeared to land with at least one attendee. Ben, who described struggling to set up his own DGX Spark despite being an engineer — “this experience sucks, we have to fix it” — said the product feels like the unlock “needed for people to really feel and understand what agentic means, and you need the right UX to make it happen.” Nvidia also emphasized that the hardware scales: connecting two Sparks over shared memory runs frontier-class open models like DeepSeek’s latest, and four can run GLM 5.2 or Nemotron Ultra. “I’ve even seen eight Sparks get connected,” Nader said.

What the deepening Nvidia-Perplexity alliance means for both companies

The launch extends a partnership that has been building for more than a year. In June 2025, Nvidia and Perplexity announced a collaboration to bring sovereign AI models to European publishers and telecoms, part of CEO Jensen Huang’s continent-hopping campaign to convince governments that, as the Associated Press reported from VivaTech in Paris, “every country needs a national intelligence infrastructure.” The sovereign AI pitch — that data “belongs to your people, your country, your culture,” in Huang’s words — is philosophically the same argument Portable Computer makes at the scale of a single desk: intelligence you control, running on hardware you own.

There is a self-interested logic for both companies. Perplexity, which has raised capital at steadily escalating valuations while facing legal pressure from publishers over its content practices — including a lawsuit filed by The New York Times in December 2025 and an earlier public dispute with Forbes — gets a product whose economics don’t depend on metering every token, and a differentiated wedge into privacy-sensitive enterprises in law, healthcare, and finance. Nvidia gets a killer app for DGX Spark, a device that, by the admission of attendees at Monday’s briefing, has been easier to buy than to use. When one reporter asked whether a Spark might ship with Portable Computer and a Nemotron model preinstalled, Nader demurred without ruling it out: “That would be cool… the goal is just making sure that it’s a super smooth experience for every user.”

Questions remain. Perplexity’s most impressive numbers come from its own internal benchmark, and the company acknowledges that compact models still trail the frontier meaningfully on hard reasoning tasks — advisor escalation “narrows but does not fully close the gap.” The launch is Linux-only for now, the 24GB VRAM floor excludes the vast majority of consumer PCs, and Apple silicon — home to some of the most enthusiastic local AI tinkerers — is conspicuously absent from the roadmap. “We’re very focused right now on Nvidia hardware,” Nate said when asked.

But the direction of travel is unmistakable. Perplexity’s researchers describe the launch as part of “a broader shift in which increasingly capable agents move from remote infrastructure to individual and local devices,” and both companies are betting that advances in chips and open models will keep expanding what a box on a desk can do. During Monday’s demos, the most telling detail wasn’t a benchmark score — it was that credit counter in the corner of the screen, sitting motionless at zero while the agent churned through a folder of tax documents. For two years, the AI industry has measured its ambitions in gigawatts and tokens per dollar. Portable Computer proposes a different meter, one that never runs.

Ukraine and the United Kingdom have signed a joint declaration on artificial intelligence partnership covering the development and deployment of AI-enabled military capabilities, according to the Ukrainian presidency’s account of talks in Kyiv between Volodymyr Zelenskyy and Andy Burnham.

It formalises a direction British defence policy has already been moving in, having committed £2bn to an AI combat laboratory for the British Army.

“We will work together to build next-generation Armed Forces and accelerate development and deployment of AI-enabled military capabilities,” Zelenskyy said. Neither side has published the text of the declaration or indicated what funding, if any, sits behind it.

The visit was Burnham’s first foreign trip since becoming prime minister, which gives the AI component more weight than a routine bilateral would carry.

Ukraine has spent four years turning battlefield necessity into a software industry, and Britain is the partner most explicitly trying to buy into what it has learned.

Munitions were the other substantive item. The two discussed localising production of Storm Shadow and SCALP cruise missiles inside Ukraine, which would move the country from receiving weapons to manufacturing them under licence.

A missile built in Ukraine is not subject to the export licences and political reversals that have repeatedly interrupted supply, and it keeps industrial capacity in the country rather than in a donor’s factory.

Air defence dominated the rest of the agenda, with anti-ballistic interceptors singled out specifically. Zelenskyy has been pressing allies to close a substantial funding gap for air defence as Russian strikes have intensified, and the arithmetic of interceptor cost against drone cost remains the central problem.

Europe’s defence technology sector has been reorganising around exactly that asymmetry. Berlin’s Stark Defence raised €500mn at a valuation above €3.5bn building cheap strike drones, and Helsing has been taking its defence AI into Japan through a partnership with Rakuten.

Ukraine has become an unusual kind of exporter in this field. Its drone and electronic warfare firms iterate on a timescale Western procurement cannot match, because their feedback loop is measured in days of combat rather than quarters of testing, and that experience is the asset London is really contracting for.

The financial track ran alongside the military one. The two discussed Ukraine’s fiscal resilience through the end of 2026 and into 2027, and war-risk insurance for maritime exports, where the London market underwrites much of the shipping that keeps Ukrainian grain moving.

Insurance is an unglamorous but decisive piece of infrastructure here. Without cover written in London, vessels do not sail, and a significant share of Ukraine’s remaining export earnings depends on that market continuing to price the risk.

On sanctions, Burnham indicated British support for further measures aimed at Russian energy revenues, the shadow fleet moving oil outside Western insurance arrangements, the financial sector, and the defence-industrial complex.

“We will stand with you,” he said. “This has been the case under every government.”

Zelenskyy, for his part, thanked Britain for support extending back to the opening weeks of the invasion, saying it was something Ukraine would never forget. Both statements are the expected register for a first visit, and neither commits either government to a figure.

What the AI declaration actually produces will depend on procurement rather than diplomacy. Britain’s defence establishment has been slow to buy from startups even as capital has poured into European military technology, and a declaration signed in Kyiv does not by itself change a contracting process in Whitehall.

There is a precedent worth remembering here, since the two countries signed a hundred-year partnership agreement in January 2025 that was similarly broad and similarly light on enforceable commitments. Declarations are cheap; contracts are not.

No timeline, budget, or delivery milestones were attached to any of it. The signature is real, the substance behind it is still to be published, and the Storm Shadow localisation in particular would take years and a licensing agreement neither government has yet described.

Years of smartphone presentations treated screen-to-body ratio like a scoreboard. Smaller black borders meant progress. Xiaomi chased the idea with the Mi Mix series. Samsung briefly tried waterfall curves that bent the glass over the sides. One extreme concept even wrapped the display around the back so the number could climb past 100 percent. Those experiments eventually settled down. Curved edges brought reflections, weaker durability, and accidental touches. Premium phones today sit at roughly one to two millimeters of border all the way around. Some people still call those borders unacceptable. Most of them were never going to buy the phone in the first place.

Earlier this summer, Marques Brownlee (MKBHD) received an email from Tecno, a company that claimed to have a concept that would finish the job. This new phone has a premium design and a 6.7-inch screen that extends all the way to the edges. When held straight on, the black border that typically surrounds most current screens vanishes. From corner to corner, edge to edge, the glass meets the silver frame with no space in between. When compared to a Pixel 11 Pro, the contrast is obvious. The Pixel still has a little black strip between the active pixels and the metal, which is highly obvious. This Tecno gadget knocks it out totally.

Sale

Google Pixel 11 Pro – Obsidian 512 GB & Pixel Watch 5 Obsidian 45mm WiFi

• Order the new Google Pixel 11 Pro today and get $250 off Pixel Watch 5; valid 8/12/2026 until 8/27/2026 at 11:59pm PST, while supplies last and…
• Return of Pixel 11 Pro without Pixel Watch 5 results in charge; offer cannot be combined, is non-transferable and not valid for cash or cash…
• If a qualifying item in your order is returned, you’ll be reimbursed for what you return

Their approach for removing that strip is internal stacking, in which the screen remains flat but the metal frame protrudes just beyond the glass, allowing the black edge of the screen to be hidden underneath. From typical viewing angles, it’s obvious. Tilt the phone or picture lifting the frame off and you’ll notice a sliver of unused glass peeking out, and that raised lip becomes really crucial when you’re actually holding the thing.

Every time you swipe from the top to check your notifications or from the bottom to return home, your finger scrapes against a metal edge. It disrupts the fluid flow, which is far more upsetting than a little traditional bezel would be. It’s a shame because the phone itself still seems very striking when you look at it. Put it to use, and a modest border with rounded edges begins to sound appealing. One big issue with the concept is the number of cases. Any case that wraps around the sides will only cover part of the screen.

The rest of the design, however, is rather competent, with a brilliant screen that refreshes at 144Hz, a triple camera array on the back, and a 6000mAh battery inside. It’s so light that it nearly feels like one of those dummy devices used to showcase phones in store windows. The hole-punch selfie camera is in its customary location. Fair play to anyone who wants to chase that full-screen ideal; many people would prefer to hide the camera behind glass, but the rest of the gear does not feel like a toy.

The major surprise lies on the back, which Tecno hardly highlighted in the press materials, but it’s far more interesting than the front. There’s a holographic lenticular pattern going on, which gives this three-dimensional butterfly-wing appearance that appears to be lifting right off the surface. MKBHD struggled to get a good photo of it, but he believes it’s one of the best examples of that optical trick currently accessible. If this design were to make it onto store shelves, it would require a clear case to be bundled with the smartphone.