The immensely successful 1986 film Pretty in Pink gave birth to an equally popular hit soundtrack album back in the day. In honor of its 40th anniversary, Universal Music and A&M Records have issued a special limited pink vinyl remastered edition of the soundtrack including two additional songs from the film not on the original release.

A timecapsule, the music on Pretty in Pink clearly resonated with fans, catapulting the soundtrack to Gold Record status within two months of release. It reached the #5 spot in the Billboard Top 200 charts at the time and #3 for the year-end top soundtracks list that year.  

Official press materials frame the reissue’s significance this way: “As essential as the film itself, the music captures the story’s emotional truth—freezing moments in time with the vivid clarity of memories that linger for decades. Film soundtracks hold a rare power to shape identity, influencing who we are, where we belong, and the social experiences that define youth. Few filmmakers understood the coming-of-age experience quite like John Hughes, and Pretty in Pink stands as one of his most enduring cultural touchstones.”

The Pretty in Pink soundtrack plays like a who’s who of the Billboard charts and vintage MTV from the mid-’80s, when hair was high, synths were higher, and every hallway needed emotional damage with a backbeat. Included is Suzanne Vega’s hooky single “Left of Center,” featuring Joe Jackson on piano, along with The Psychedelic Furs’ updated re-recording of the 1981 track that inspired the film’s title, “Pretty in Pink.” OMD’s “If You Leave” became a No. 4 charting hit, while tracks from The Smiths, Echo & The Bunnymen, and INXS round out the pure ’80s vibe.

For this 40th Anniversary pink vinyl special edition, two songs from the film that were not included on the original soundtrack have been added: Otis Redding’s classic “Try a Little Tenderness” and Talk Back’s “Rudy.”

In general the sound quality on the Pretty in Pink soundtrack is very good as far as ’80s music productions go. The mastering by Bernie Grundman at Bernie Grundman Mastering in Hollywood no doubt helps bring out the best across all these different productions. The standard weight pressing is happily well centered and the opaque pink vinyl is actually quiet (which isn’t always the case with color vinyl variants). Original album graphics are in place and there is a new gatefold spread of photos from the film.   

While the pink vinyl seems to be sold out at the time of this writing, you can order the standard black reissue of the Pretty In Pink soundtrack for $27.98 at Amazon. Fans of the pink vinyl edition should check in at Interscope‘s website where they have a special portal to inquire about represses. Most sites I’ve seen are currently sold out but if there is enough demand maybe they’ll repress. The pink vinyl is already commanding crazy collector’s pricing on Discogs, so be forewarned! 

Whether you’re revisiting the ’80s or replacing a well-worn original LP or cast-off CD, Pretty in Pink remains a legitimate pop music cultural touchstone and a fun spin to boot.

Where to buy: $27.98 at Amazon

Mark Smotroff is a deep music enthusiast / collector who has also worked in entertainment oriented marketing communications for decades supporting the likes of DTS, Sega and many others. He reviews vinyl for Analog Planet and has written for Audiophile Review, Sound+Vision, Mix, EQ, etc.  You can learn more about him at LinkedIn.

Related Reading:

AI + ML

Text is the new attack

The adoption of AI agents has expanded the potential attack surface beyond code to natural language text.

AI agents – models wrapped in software that can use tools and perform multi-step tasks – often take direction from text-based skills. And researchers have demonstrated that skills can be weaponized.

“Many agent frameworks allow users to install skills from online registries so the agent can discover and use new capabilities on demand,” said Soheil Feizi, computer science professor at the University of Maryland (UMD) and founder/CEO of RELAI.ai, in a social media post. “This is powerful, but it also creates a new attack surface.”

Skills, Feizi explains, are not just code or dependencies. They’re also text instructions that tell agents what to do.

Skills, written out in a SKILL.md file, consist of text prompts with other data and resource references (e.g. URLs). They may get added to a user’s initiating prompt and pre-existing system prompts, all of which get fed to a model for a response. Typically, this happens when the user wants the model to perform a specific task that has been spelled out in a skill file, like conducting a code quality review.

When a model’s prompt – the combination of user input, instructions within skills, and system prompts – gets modified inadvertently or adversarially, that’s prompt injection. That can happen directly, if for example, a user submits a prompt that directs the model to ignore prior instructions. It can also happen indirectly, if for example, an AI agent visits a website and processes text on a page that the underlying model interprets as an instruction. 

A skill can effectively act as user-authorized prompt injection. And agents may also automatically retrieve and load third-party skills if their descriptions appear relevant to the task being pursued. And therein lies the problem.

The risk posed by skills has already been documented. In February, security biz Snyk found that 13.4 percent of skills on ClawHub and skills.sh (about 534 out of 3,984) “contain at least one critical-level security issue, including malware distribution, prompt injection attacks, and exposed secrets.”

In a preprint paper titled “Under the Hood of SKILL.md: Semantic Supply-chain Attacks on AI Agent Skill Registry,” Feizi and UMD co-authors Shoumik Saha and Kazem Faghih examine the role that skill registries play in the distribution of malicious skills. Specifically, they look at how adversarial skills get discovered, selected, and vetted before execution.

“An attacker may not need to hide malware in executable code,” Feizi said. “Small semantic changes to a skill description can affect how the skill is discovered in a registry, whether an agent selects it over alternatives, and whether it passes governance or safety checks.”

Those details matter, he argues, because the selection process may be automated – software agents like OpenClaw have the ability to fetch and use third-party skills.

The text that influences tool discovery and usage thus has security implications, which may not be addressed by traditional security scanning mechanisms that focus on code.

The three co-authors show that short 20-token triggers can be added to a SKILL.md file to influence the chance an agent will discover it in a registry, to influence the chance an agent will select that skill, and to avoid detection through semantic evasion strategies.

In terms of discovery, the researchers demonstrated they could induce an agent to discover their skill over an unaltered source skill 86 percent of the time. They also succeeded in making an agent select their skill over variants 77.6 percent of the time. And they were able to evade registry scanning defenses between 36.5 percent and 100 percent of the time.

The most successful strategy for evading detection was to overflow the context window of the scanner – making the skill too long for the scanner to handle. “In ClawHub-style review, only the first 10K characters of long SKILL.md files are passed to the LLM reviewer, so we place the malicious instruction beyond this boundary while keeping it in the submitted skill,” the authors explain.

“Our work shows that protecting agents requires treating natural-language specifications as security-sensitive objects,” said Feizi. “We hope this encourages more careful design of skill registries, ranking mechanisms, governance pipelines, and agent-side defenses.”

Source code and supporting documentation have been published on GitHub. ®

If you are an Audible member who also lives inside Apple Podcasts, this one is for you. Audible has officially launched a connected subscription that lets you stream nearly 700 premium titles directly in Apple Podcasts, across 135 countries, with more on the way.

What you get with Audible on Apple Podcasts?

The catalog is genuinely stacked. You get ad-free access to Audible Originals spanning true crime, investigative journalism, celebrity-led audio dramas, and personal growth series. That includes award-winning shows like Dr. Death and Hysterical, as well as star-powered productions like Kerry Washington’s The Prophecy and The Big Lie, starring Jon Hamm.

Marshall Lewy, Head of Audible Content for North America, says the goal is to meet listeners where they already are. He also said, “by making select shows and episodes available widely, we have the opportunity to introduce new listeners to the extraordinary audio storytelling Audible offers right inside the Apple Podcasts app.”

How to connect your Audible account to Apple Podcasts

If you are already an Audible member, open Apple Podcasts, and your subscription should link automatically. You can also search for any Audible premium show and connect from there at no extra cost.

New subscribers can sign up directly through Apple Podcasts by searching for a show like Dr. Death or OnlyFantasy and subscribing via the Audible app. Your membership then works across both platforms.

The connected subscription is currently live for US and UK members, with Australia, Japan, and Canada joining in June 2026.

Apple Podcasts has been on a serious upgrade spree lately, with its iOS 26.4 update bringing native video streaming, Picture-in-Picture, and offline downloads. That overhaul also introduced dynamic ad insertion for creators.

Audible, meanwhile, launched a new Standard membership plan at $8.99 per month for first-time listeners. Between the two platforms, there is a lot to look forward to as a listener.

London-based Humanoid announced a manufacturing agreement with Bosch that moves its HMND-01 robots closer to larger deployments across European factories and warehouses. The deal follows a successful proof-of-concept test completed in March at one of Bosch’s logistics sites in Bühl, Germany. During that trial, the robots handled box transfers from conveyors to trolleys, managing five different box sizes with varying weights and shapes without missing a step.

Humanoid develops robots for real-world applications, such as in factories and industrial environments. Their HMND-01 platform is available in two configurations: a 175 centimeter tall bipedal robot and a wheeled version. The bipedal one weighs roughly 70 kilograms and can move at a reasonable rate of 1.5 meters per second, which is quite quick for a robot. It can also carry a 15 kilogram payload with ease and run for approximately 4 hours on a single charge. There is also a wheeled variant, which is heavier, higher, and geared to handle demanding duties. Both versions run on Humanoid’s own KinetIQ system, which manages everything from perception to movement and actual task completion, and because they’re modular, they can be tailored to customers’ specific manufacturing or logistics needs.

Unitree G1 Humanoid Robot（No Secondary Development）

• Sleek & Durable Design: Standing at 132cm tall and weighing only approx. 35kg, the G1 is constructed with aerospace-grade aluminum alloy and carbon…
• High Flexibility & Safe Movement: Boasting 23 joint degrees of freedom (6 per leg, 5 per arm), it offers an extensive range of motion. For safety, it…
• Smart Interaction & Connectivity: Powered by an 8-core high-performance CPU and equipped with a depth camera and 3D LiDAR. It supports Wi-Fi 6 and…

Bosch has agreed to help Humanoid deliver these robots to Europe. With its extensive manufacturing facilities and experience creating extremely complicated systems, the German corporation brings significant expertise to the table. Bosch’s contract manufacturing agreement include assisting Humanoid with all of the technical aspects of constructing the robots. That begins with the design stage and continues all the way through to delivery, which includes supply chain coordination and, of course, cost control. Meanwhile, Humanoid is working on improving the design, with the goal of making it easier to make, more dependable for customers to use, simpler to fix if something goes wrong, and more economical in general.

The combined testing at Bühl produced some hopeful results, as the robots performed rather well in a real-world logistics scenario. They were scanning boxes, responding to changing conveyor belt inputs, and transferring products smoothly. That was enough to persuade both sides to take the project a step further and begin producing robots in larger quantities, rather than simply focusing on the prototype phase.

What the two companies want to accomplish next isn’t just to keep producing the same old robots. They also want to look into how Bosch-made components like as actuators, motors, and sensors may be employed in future versions of the HMND-01. The goal is to make the thing faster, stronger, and more tightly integrated with the equipment and systems it works with.
[Source]

SpaceX has launched the upgraded third version of its Starship rocket for the first time, though the test launch did not go perfectly for Elon Musk’s spaceflight company.

The 407-foot rocket — the most powerful ever built — lifted off from SpaceX’s company town Starbase, Texas, at 5:30 p.m. local time. Just a few minutes later, the upper stage ship separated from the Super Heavy booster and continued on into space.

The booster pitched away from the Starship vehicle and headed back to Earth, where it was supposed to perform a simulated landing in the Gulf of Mexico. But the booster’s engines did not properly re-ignite for the sustained burn that is meant to deliver it back to the launch site. The booster then tumbled down to the water, where it likely exploded.

Starship, meanwhile, lost one of its six Raptor engines as it ascended into space. But it successfully deployed all 20 of the Starlink satellite simulators along with two modified Starlink satellites meant to record footage of Starship’s exterior. Roughly one hour after liftoff, Starship simulated a landing in the Indian Ocean, before tipping over and exploding as expected.

While it didn’t go exactly according to plan, this was an important test launch for SpaceX. It was the first real shakedown of the upgraded Starship V3 hardware, which has been in development for months. The company was also testing out an all-new launchpad at Starbase that it’s been developing and building for years.

The test launch also comes at a historical inflection point for SpaceX as a company. Its IPO filing was made public this week, and SpaceX is expected to list on the Nasdaq in mid-June. The IPO is reportedly supposed to raise around $75 billion for SpaceX, which the company plans to use to fuel further development, massive AI ambitions, and to pay off some of the debt associated with xAI and Musk’s social media company X. (That means this could also be the last Starship test launch to happen without a stock market reaction.)

SpaceX has spent years and billions of dollars developing Starship, which it sees as crucial to its mission of making life multi-planetary. The company plans to use Starship for NASA missions to the moon, and eventually Mars. But the big job it needs to perform in the near-term is delivering more advanced Starlink satellites to Earth orbit, as Starlink is the only profitable part of SpaceX’s business.

This was the first Starship flight since October 2025. SpaceX had planned to attempt launching Starship V3 earlier, but one of the first upgraded boosters suffered an explosion during testing in November. SpaceX first attempted the launch on Thursday but ultimately had to delay it as a hydraulic pin on the launch tower arm refused to retract, according to Musk.

This new version of Starship features SpaceX’s third-generation Raptor engines, which have more thrust and a far simpler design. The new booster is designed for faster takeoffs and easier catches by the launch tower.

This story has been updated to include information about Starship’s simulated landing.

On Friday, the ultrafast-fashion giant Shein finalized its acquisition of Everlane, a US clothing retailer that made its name by promising “radical transparency” into how its clothes were made. Neither company disclosed the price of the deal, but Puck reported last weekend that it clocked in at $100 million.

Founded in 2010, Everlane became synonymous with a certain strain of millennial consumerism that was supposed to be the exact opposite of Shein. It mainly sold elevated basics, and told a generation of anxious and high-minded shoppers that they could feel morally good about buying yet another pair of plain ballet flats or black high-waisted skinny jeans. Shein, by contrast, became notorious by flooding the internet with astonishingly cheap, trendy clothing produced at enormous scale. It has been criticized for years for alleged poor labor practices.

Given how differently Shein and Everlane positioned themselves, many people online felt the acquisition fell somewhere between darkly ironic and outright dystopian. The fashion writer Derek Guy, better known online as the “menswear guy,” articulated the vibe in a post on X: “Under Shein,” he wrote, “Everlane’s ‘radical transparency’ means you get to read about the small child making your boring gray crewneck sweater.”

Really, though, the deal makes perfect sense. In the long run, it may end up looking like a preview of where Chinese consumer companies are heading next.

Chinese ecommerce giants conquered the global market largely by selling cheap stuff at eye-watering scale. Companies like Shein and Temu thrived in part because of the “de minimis” loophole, a US trade rule that allowed packages worth under $800 to enter the country tariff-free and with relatively little customs scrutiny. That system became the backbone of a new era of cross-border ecommerce, enabling Chinese companies to ship cheap stuff directly to American consumers faster and more efficiently than many traditional retailers could manage.

But after US president Donald Trump imposed sweeping new tariffs on Chinese imports and ended the de minimis exemption, the economics underpinning that model began to falter. Chinese companies quickly realized they could no longer rely solely on flooding Western markets with bargain-priced products. If they wanted to keep growing internationally, they needed something more durable: a good old-fashioned brand.

Shein buying Everlane, however culturally cursed it may appear, is part of a broader trend already unfolding across Chinese commerce and manufacturing. Increasingly, Chinese companies are trying to move beyond anonymous low-cost production and toward owning recognizable global brands associated with quality, lifestyle, and status.

One of the clearest examples comes from Temu’s parent company, Pinduoduo. In March, the company announced a major new initiative called New PinMu, a multibillion-dollar effort designed to help Chinese manufacturers build premium international brands. The project is part of a larger strategic vision outlined by Pinduoduo co-CEO Jiazhen Zhao, who has been hyping up the company’s ambitions to elevate manufacturing standards and create pathways for Chinese factories to move up the value chain.

Meanwhile, Luckin Coffee, a Chinese coffee chain that has become one of Starbucks’ biggest rivals, recently acquired Blue Bottle, the cultish specialty coffee brand that helped define American third-wave coffee culture. Anta Sports, a Chinese sportswear giant that began largely as a domestic sneaker company, has spent years buying into premium global sportswear brands, including controlling stakes in Arc’teryx and Salomon.

The trend also reflects broader political pressures inside China. The government has become increasingly critical of the brutal price wars and hypercompetition that dominate industries like ecommerce and electric cars, a phenomenon often referred to as “involution.” Beijing now wants companies to focus more on sustainable growth, higher-end manufacturing, and global competitiveness rather than an endless race to the bottom.

When you think of Memorial Day sales, you probably think of mattresses and other home goods. And while those items are definitely discounted, now is also a good time to purchase tech. Personally, I’m not buying anything right now unless it’s discounted—and fortunately many of our top picks are. Whether you’re shopping for a power bank, a new pair of headphones, or some other gadget, I’ve rounded up the best Memorial Day deals for your perusal. We’ll update this article again over the weekend.

Check out our buying guides for more recommendations, including the best headphones, the best laptops, and the best cheap phones.

Updated May 22: We’ve checked prices, removed expired deals, added ten new deals, and ensured accuracy throughout.

WIRED Featured Deals:

Sony WH-1000XM5 for $248 ($152 off)

The Sony WH-1000XM5 have a very frustrating name, but they’re the predecessor to our favorite wireless headphones, and they’re still an excellent pick if you don’t want to shell out for the new WH-1000XM6. They go on sale frequently, but rarely drop this low in price, which comes within $5 of their all-time low. If you’re in the market for over-ear headphones, they’re hard to beat. They’re comfortable, portable, lightweight, and stylish, and they’ll make your music sound great no matter what you like to listen to.

In the latest sign of these AI-heavy times, the National Transportation Safety Board temporarily removed access to its docket system after discovering that voices of pilots who were killed in a UPS plane crash last year had been re-created using AI and were circulating on the internet.

The NTSB is prohibited by federal law from including cockpit audio recordings in its docket system, which otherwise contains troves of data on investigations and has historically been open to the public. But the accident docket for this flight included a spectrogram file of the voice recorder. A spectrogram uses a mathematical process to turn sound signals, including low and high frequencies, into an image.

Scott Manley, a popular YouTuber whose channel combines physics, astronomy, and video games, noted on X that it could be possible to reconstruct audio from the megabytes of data encoded in that image.

And that’s what happened. People took the spectrogram, along with the publicly available transcript, to create approximations of the cockpit voice recorder audio from UPS Flight 2976 in Louisville, Kentucky, according to the NTSB. They used AI tools like Codex, according to posts on social media.

The agency restored public access to the docket system on Friday but kept 42 investigations closed pending review — including the one related to Flight 2976.