As more legislation sweeps the nation limiting children’s phone use in schools, a new report card shows not all laws are created equal.
The “Phone-Free Schools State Report Card,” released late last month, gave only two states “A” grades out of the 40 with phone-free legislation. North Dakota and Rhode Island both received high marks for their stringent laws, dictating that devices be stored in inaccessible spots during the entirety of the school day.
The report card was born after a massive uptick in the number of states tackling potential over-use of personal electronic devices in the classroom. It’s a collaboration between entities advocating for limiting children’s exposure to technology: the Institute for Families and Technology, Smartphone Free Childhood US, the Becca Schmill Foundation and The Anxious Generation, a new nonprofit that emerged out of the best-selling book of the same name.
“There’s been so much movement, which has been very encouraging, however not all laws are created equal,” says Kim Whitman, lead researcher on the report card and co-lead for Smartphone Free Childhood US.
Advertisement
Some states initially adapted more lax laws, namely banning devices during instructional time, but allowing access during lunch or passing periods. The early adopters, including Florida, Louisiana and Indiana, widened the scope to become bell-to-bell banners in recent years, with 17 states adapting bell-to-bell policies straight away, according to Whitman.
From the perspective of the report card authors, the more comprehensive the ban, particularly when it comes to keeping phones inaccessible, the better. In the latest report, a majority of states (17, plus the District of Columbia) received a “B” rating for their “bell-to-bell” mandate, which requires devices to be put away during instructional time, but lost points for keeping phones accessible.
“We know phones are addictive and it’s hard for adults, let alone kids, to resist the ping in their pocket,” Whitman says, pointing toward research that teacher retention goes up when phones are in inaccessible places, since teachers do not have to police students’ usage.
While a majority of states get passing grades for their device restrictions, only two states received “A” grades after requiring devices be placed in inaccessible areas.
Source: Phone-Free Schools State Report Card, compiled by the Institute for Families and Technology, Smartphone Free Childhood US, the Becca Schmill Foundation and The Anxious Generation
Advertisement
Students themselves support more of a ban in classrooms only (with 41 percent in favor) and are less in favor of an all-day ban (with only 17 percent in favor), according to new data from the Pew Research Center.
Eight states were not judged in the report because they are crafting current legislation. Only two states — South Dakota and Montana — received zero points for not having any legislation, with four states (Wyoming, Mississippi, Connecticut and Maryland) receiving an “F” on the report card after proposed legislation failed to pass.
Safety is the largest argument that opponents to bans make against keeping phones inaccessible, with parents voicing concerns about contacting their child during an emergency, such as a school shooting. Whitman pointed toward research from the National Association of School Resource Officers that states it is actually less safe for students to use cellphones in that scenario, as a phone chirping may alert a shooter to their location; distract students from listening to teachers during an emergency; or cause parents to flock to the school, impeding law enforcement.
“During school emergencies, worried parents understandably want desperately to contact their children and be reassured that the children are safe,” the National Association of School Resource Officers said in a statement. “The risks posed by phone access during school emergencies are even greater, however, than during normal times.”
Advertisement
There are laws that carve out exceptions for students with 504 plans and IEPs, which the report card does not fault, though Whitman says it becomes a slippery slope when the legislation begins to make exceptions for things like “educational purposes,” such as studying social media.
“If all teachers can decide when kids can use their phones for educational purposes, it erodes the policy,” she says.
Brian Jacob, the co-director of University of Michigan’s Youth Policy Lab, previously voiced concerns to EdSurge about legislation placing the onus on teachers.
“I fear a lot of schools will ban them but say ‘Kids have to keep them in their pockets and teachers have to police that,’ and that approach will be really tough to implement in any way,” he says, adding it is best to mandate keeping them in lockers or a centralized location.
Advertisement
As the organizations continue to advocate for phone-free schools, Whitman says there is also a focus on expanding that reach to school-issued technology in general. The Distraction-Free Schools Policy Project is working on introducing “Safe School Technology” legislation, which pushes for eliminating all screen technology in elementary schools, prohibiting sixth through eighth graders from taking their school-issued devices home with them and prohibiting technology that uses generative artificial intelligence across all grade levels.
“A lot of the issues with personal devices can move to the district-issued devices,” Whitman says, explaining if students do not have cellphones, they can still chat on their MacBooks, or through Google Docs. “There are definitely issues with school-issued devices as well. But removing phones is the first step.”
Most enterprise RAG pipelines are optimized for one search behavior. They fail silently on the others. A model trained to synthesize cross-document reports handles constraint-driven entity search poorly. A model tuned for simple lookup tasks falls apart on multi-step reasoning over internal notes. Most teams find out when something breaks.
Databricks set out to fix that with KARL, short for Knowledge Agents via Reinforcement Learning. The company trained an agent across six distinct enterprise search behaviors simultaneously using a new reinforcement learning algorithm. The result, the company claims, is a model that matches Claude Opus 4.6 on a purpose-built benchmark at 33% lower cost per query and 47% lower latency, trained entirely on synthetic data the agent generated itself with no human labeling required. That comparison is based on KARLBench, which Databricks built to evaluate enterprise search behaviors.
“A lot of the big reinforcement learning wins that we’ve seen in the community in the past year have been on verifiable tasks where there is a right and a wrong answer,” Jonathan Frankle, Chief AI Scientist at Databricks, told VentureBeat in an exclusive interview. “The tasks that we’re working on for KARL, and that are just normal for most enterprises, are not strictly verifiable in that same way.”
Those tasks include synthesizing intelligence across product manager meeting notes, reconstructing competitive deal outcomes from fragmented customer records, answering questions about account history where no single document has the full answer and generating battle cards from unstructured internal data. None of those has a single correct answer that a system can check automatically.
Advertisement
“Doing reinforcement learning in a world where you don’t have a strict right and wrong answer, and figuring out how to guide the process and make sure reward hacking doesn’t happen — that’s really non-trivial,” Frankle said. “Very little of what companies do day to day on knowledge tasks are verifiable.”
The generalization trap in enterprise RAG
Standard RAG breaks down on ambiguous, multi-step queries drawing on fragmented internal data that was never designed to be queried.
To evaluate KARL, Databricks built the KARLBench benchmark to measure performance across six enterprise search behaviors: constraint-driven entity search, cross-document report synthesis, long-document traversal with tabular numerical reasoning, exhaustive entity retrieval, procedural reasoning over technical documentation and fact aggregation over internal company notes. That last task is PMBench, built from Databricks’ own product manager meeting notes — fragmented, ambiguous and unstructured in ways that frontier models handle poorly.
Training on any single task and testing on the others produces poor results. The KARL paper shows that multi-task RL generalizes in ways single-task training does not. The team trained KARL on synthetic data for two of the six tasks and found it performed well on all four it had never seen.
Advertisement
To build a competitive battle card for a financial services customer, for example, the agent has to identify relevant accounts, filter for recency, reconstruct past competitive deals and infer outcomes — none of which is labeled anywhere in the data.
Frankle calls what KARL does “grounded reasoning”: running a difficult reasoning chain while anchoring every step in retrieved facts. “You can think of this as RAG,” he said, “but like RAG plus plus plus plus plus plus, all the way up to 200 vector database calls.”
The RL engine: why OAPL matters
KARL’s training is powered by OAPL, short for Optimal Advantage-based Policy Optimization with Lagged Inference policy. It’s a new approach, developed jointly by researchers from Cornell, Databricks and Harvard and published in a separate paper the week before KARL.
Standard LLM reinforcement learning uses on-policy algorithms like GRPO (Group Relative Policy Optimization), which assume the model generating training data and the model being updated are in sync. In distributed training, they never are. Prior approaches corrected for this with importance sampling, introducing variance and instability. OAPL embraces the off-policy nature of distributed training instead, using a regression objective that stays stable with policy lags of more than 400 gradient steps, 100 times more off-policy than prior approaches handled. In code generation experiments, it matched a GRPO-trained model using roughly three times fewer training samples.
Advertisement
OAPL’s sample efficiency is what keeps the training budget accessible. Reusing previously collected rollouts rather than requiring fresh on-policy data for every update meant the full KARL training run stayed within a few thousand GPU hours. That is the difference between a research project and something an enterprise team can realistically attempt.
Agents, memory and the context stack
There has been a lot of discussion in the industry in recent months about how RAG can be replaced with contextual memory, also sometimes referred to as agentic memory.
For Frankle, it’s not an either/or discussion, rather he sees it as a layered stack. A vector database with millions of entries sits at the base, which is too large for context. The LLM context window sits at the top. Between them, compression and caching layers are emerging that determine how much of what an agent has already learned it can carry forward.
For KARL, this is not abstract. Some KARLBench tasks required 200 sequential vector database queries, with the agent refining searches, verifying details and cross-referencing documents before committing to an answer, exhausting the context window many times over. Rather than training a separate summarization model, the team let KARL learn compression end-to-end through RL: when context grows too large, the agent compresses it and continues, with the only training signal being the reward at the end of the task. Removing that learned compression dropped accuracy on one benchmark from 57% to 39%.
Advertisement
“We just let the model figure out how to compress its own context,” Frankle said. “And this worked phenomenally well.”
Where KARL falls short
Frankle was candid about the failure modes. KARL struggles most on questions with significant ambiguity, where multiple valid answers exist and the model can’t determine whether the question is genuinely open-ended or just hard to answer. That judgment call is still an unsolved problem.
The model also exhibits what Frankle described as giving up early on some queries — stopping before producing a final answer. He pushed back on framing this as a failure, noting that the most expensive queries are typically the ones the model gets wrong anyway. Stopping is often the right call.
KARL was also trained and evaluated exclusively on vector search. Tasks requiring SQL queries, file search, or Python-based calculation are not yet in scope. Frankle said those capabilities are next on the roadmap, but they are not in the current system.
Advertisement
What this means for enterprise data teams
KARL surfaces three decisions worth revisiting for teams evaluating their retrieval infrastructure.
The first is pipeline architecture. If your RAG agent is optimized for one search behavior, the KARL results suggest it is failing on others. Multi-task training across diverse retrieval behaviors produces models that generalize. Narrow pipelines do not.
The second is why RL matters here — and it’s not just a training detail. Databricks tested the alternative: distilling from expert models via supervised fine-tuning. That approach improved in-distribution performance but produced negligible gains on tasks the model had never seen. RL developed general search behaviors that transferred. For enterprise teams facing heterogeneous data and unpredictable query types, that distinction is the whole game.
The third is what RL efficiency actually means in practice. A model trained to search better completes tasks in fewer steps, stops earlier on queries it cannot answer, diversifies its search rather than repeating failed queries, and compresses its own context rather than running out of room. The argument for training purpose-built search agents rather than routing everything through general-purpose frontier APIs is not primarily about cost. It is about building a model that knows how to do the job.
It’s been a wild week for Apple. After announcing a slew of new hardware, the company capped things off with its cheapest laptop ever: the $599 MacBook Neo. It’s low on specs, but high on character and value. In this episode, Devindra and Engadget Deputy Editor Nathan Ingraham dive into the MacBook Neo, as well as the refreshed MacBook Air M5, MacBook Pro M5 Pro/Max, iPad Air M4 and iPhone 17e.
Also, Devindra chats with Spencer Ackerman, author of Forever Wars and recent Iron Man comics, about the ongoing battle between Anthropic and the Department of Defense. It turns out the DOD still used Claude for attacks on Iran, after banning Anthropic’/s AI last week. And really, what do these AI companies expect to happen when they jump at military contracts?
Subscribe!
Topic
Apple announces a the MacBook Neo priced at $599 and it’s shockingly great – 0:53
MacBook Air got the M5, MacBook Pro got the M5 Pro and M5 Max, and who needs the new iPad Air now? – 22:31
Anthropic vs. DoD with Spencer Ackerman, author of The Forever Wars – 30:34
Gemini encouraged a man to end his own life to be with his ‘AI wife’ – 58:53
Polymarket nixes bets on nuclear detonation after public outcry – 1:01:55
No Yōtei on PC: Sony closes down first party titles outside of PS5 – 1:03:56
Wildlight Studios’ Highguard shuts down after 46 days live – 1:08:23
Working on: Dell’s XPS 14 will be great when the keyboard fix comes through – 1:15:09
Pop culture picks – 1:15:58
Credits
Hosts: Devindra Hardawar and Nathan Ingraham Guest: Spencer Ackerman Producer: Ben Ellman Music: Dale North and Terrence O’Brien
Light aircraft often use a heading indicator as a way to know where they’re going. Retired instrumentation engineer [Don Welch] recreated a heading indicator of his own, using cheap off-the-shelf hardware to get the job done.
The heart of the build is a Teensy 4.0 microcontroller. It’s paired with a BNO085 inertial measurement unit (IMU), which combines a 3-axis gyro, 3-axis accelerometer, and 3-axis magnetometer into a single package. [Don] wanted to build a heading indicator that was immune to magnetic disturbances, so ignored the magnetometer readings entirely, using the rest of the IMU data instead.
Upon startup, the Teensy 4.0 initializes a small round TFT display, and draws the usual compass rose with North at the top of the display. Any motion after this will update the heading display accordingly, with [Don] noting the IMU has a fast update rate of 200 Hz for excellent motion tracking. The device does not self-calibrate to magnetic North; instead, an encoder can be used to calibrate the device to match a magnetic compass you have on hand. Or, you can just ensure it’s already facing North when you turn it on.
Advertisement
Thanks to the power of the Teensy 4.0 and the rapid updates of the BNO085, the display updates are nicely smooth and responsive. However, [Don] notes that it’s probably not quite an aircraft-spec build. We’ve featured some interesting investigations of just how much you can expect out of MEMS-based sensors like these before, too.
In the days leading up to one of the games industry’s bigger trade conferences, Microsoft has quietly unveiled the code name for its next-generation Xbox console: Project Helix.
The name appeared without initial fanfare in a post on X on Thursday morning.
Xbox CEO Asha Sharma, who just replaced longtime leader Phil Spencer, followed up in a post on her own account, in which she briefly discussed her team’s “commitment to the return of Xbox.” Sharma also noted that Project Helix will “lead in performance” and “play your Xbox and PC games.”
Next week marks the annual Game Developers’ Conference in San Francisco, which has gained some prominence for news and announcements in recent years. It’s possible that some new information about this next-gen Xbox will come out of this year’s GDC, which is both Sharma’s first time at the show and her first time attending as the head of Xbox. Sharma reportedly has plans to meet with both partners and studios while at GDC.
That marks the end of the information about Project Helix that’s currently publicly available. The most remarkable fact about it for now may simply be that it exists, in the face of persistent rumors that Microsoft’s executives would like to sunset Xbox entirely and an ongoing memory shortage caused by the rise of AI data centers.
Advertisement
Despite industry expectations, it looks like Microsoft’s games division plans to stick it out for at least one more console generation. The start of that generation may be pushed off a couple of years from its initially rumored late-2027 starting point, as RAM is currently getting scarcer on the market, but whenever it begins, it looks like Xbox will still be there.
You might think, “Oh, libusb handles that.” But, of course, it doesn’t do the actual work. In fact, there are two possible backends: netlink or udev. However, the libusb developers strongly recommend udev. Turns out, udev also depends on netlink underneath, so if you use udev, you are sort of using netlink anyway.
If netlink sounds familiar, it is a generic BSD-socket-like API the kernel can use to send notifications to userspace. The post shows example code for listening to kernel event messages via netlink, just like udev does.
When udev sees a device add message from netlink, it resends a related udev message using… netlink! Turns out, netlink can send messages between two userspace programs, not just between the kernel and userspace. That means that the code to read udev events isn’t much different from the netlink example.
Advertisement
The next hoop is the udev event format. It uses a version number, but it seems stable at version 0xfeedcafe. Part of the structure contains a hash code that allows a bloom filter to quickly weed out uninteresting events, at least most of the time.
The post documents much of the obscure inner workings of USB hotplug events. However, there are some security nuances that aren’t clear. If you can explain them, we bet [Arcanenibble] would like to hear from you.
If you like digging into the Linux kernel and its friends, you might want to try creating kernel modules. If you get overwhelmed trying to read the kernel source, maybe go back a few versions.
Pablo Casilimas (left), founding partner at OneSixOne Ventures, with Sudheesh Nair, co-founder and CEO of TinyFish. (GeekWire Photo / Taylor Soper)
The AI moment is not just another tech cycle — it’s one of the best openings founders have seen in years.
That was the message from Sudheesh Nair, a longtime Bay Area tech leader and co-founder of enterprise web agent startup TinyFish, speaking Thursday at a Seattle Enterprise AI Summit event hosted by OneSixOne.
“There is no better time to start companies than now,” he said. “It’s just magical.”
He believes the AI boom could produce the same kind of lasting infrastructure and category-defining companies that came out of earlier economic and technology shifts. Nair said this wave may be as significant as the internet, and possibly even bigger, because “for the first time, reasoning can be on tap.”
He added: “The way I think of it is, completely be constrained by your imagination — but nothing else.”
Advertisement
Nair previously helped scale Nutanix and ThoughtSpot. In 2024 he launched TinyFish, which raised $47 million last year to build infrastructure for AI agents to operate across the web. “I couldn’t stand on the sidelines,” he said.
He likened today’s moment to a gold rush, noting that most of the enduring outcomes from 1849 were second‑order products and infrastructure: durable jeans, safer elevators, modern banking systems. He said these were built not for the gold rush, but because of the gold rush.
Nair pushed back on the instinct to wait for clarity in a fast‑moving market where even frontier AI labs are still figuring out how their models behave. “No one who knows what the heck is happening,” he said.
But Nair also was careful not to romanticize startups. He said company-building is not for everyone, and noted that some people are better suited to join startups or build inside larger organizations. His broader point was that the tools, the pace of change, and the raw opportunity around AI have created a rare moment for people willing make the startup leap.
Advertisement
“If you just happen to have a pickaxe and shovel, the best thing might be to just jump in,” Nair said.
Fewer couples are getting married, and it has impacted The Chinese Wedding Shop’s sales
Marriage has long been seen as an important union between two families across cultures. But in Singapore, fewer couples are choosing to tie the knot.
Recently released figures show that marriages in Singapore fell by about 6.2%, from 26,328 in 2024 to 24,687 in 2025. This decline follows a broader drop after the country hit a record peak of 29,389 marriages in 2022.
After a 30% increase from 2020 to 2022, there has been an almost 16% drop in the total number of marriages in Singapore since 2022./ Data from the Singapore Department of Statistics
But this trend doesn’t just reflect shifting social priorities in the city-state—it’s forcing Singapore’s wedding industry, from banquet services to bridal studios, to rethink their strategies. And for niche retailers like The Chinese Wedding Shop, they need to find a way to balance tradition and staying relevant in a market where fewer people are saying “I do.”
Vulcan Post speaks to co-founder Michelle Neo on how The Chinese Wedding Shop, a specialist in traditional Chinese wedding products, is navigating a wedding recession.
The Chinese Wedding Shop has been around for almost 20 years
Michelle first established The Chinese Wedding Shop with her husband in 2009, investing S$400,000 of their savings to open their first store at Ang Mo Kio. From the start, they positioned the shop as a one-stop destination for couples seeking items for traditional Chinese wedding customs.
Advertisement
One example is the Guo Da Li (过大礼), a ceremony where the groom’s family presents wedding gifts to the bride’s family as a sign of respect as sincerity.
The Chinese Wedding Shop’s Guo Da Li package./ Image Credit: The Chinese Wedding Shop
Back then, the co-founder shared that there was strong demand for such products.
“At that time, many of our friends who were getting married were extremely stressed trying to source traditional Guo Da Li items,” said Michelle. “They had to run from shop to shop, often with little guidance, and were worried about ‘doing it wrong’ in front of the elders.”
Beyond retail, the business also guides couples through traditional wedding customs. Each visit starts with a conversation to understand the couple’s background, which includes details like:
Dialect group
Family expectations
Wedding timeline
How traditional or modern they wish the ceremony to be
After consolidating this information, the team guides customers step-by-step through the customary sequence, explaining the essentials and optional items, and how certain practices can be simplified or adapted.
“Our focus is to ensure couples feel confident and reassured, rather than overwhelmed,” emphasised Michelle. These consultations helped the business build credibility and eventually expand to five locations across Singapore.
Advertisement
Adapting to a shrinking market
The Chinese Wedding Shop’s store at Ang Mo Kio./ Image Credit: Rong Yi Lim, Amy Yanling Charles via Google Images
But shifting wedding trends over the past few years have forced the business to adapt.
While she did not disclose figures, Michelle shared that these trends have gradually reduced overall sales volumes. Customers have also become more intentional with their spending, carefully weighing what’s essential and what’s not.
“Previously, couples were more worried about following traditions strictly. Today, they are more focused on why certain customs exist and how they can adapt them meaningfully without it being unnecessarily complex,” said Michelle.
Image Credit: The Chinese Wedding Shop/ Junhong Khang via Google Images
To combat the decline, the shop has gradually introduced new strategies: diversifying its curated traditional wedding sets, offering rentals of individual items like wedding baskets, and creating more flexible packages that let couples personalise dowry sets and other ceremonial essentials.
As more consumers shift their shopping habits online and value convenience in acquiring products, particularly after the COVID-19 pandemic, the business also started selling its products online in 2020, both through its own website and e-commerce platforms like Shopee and Lazada.
Advertisement
“These new streams helped offset the drop in traditional transactions,” said Michelle, adding that the shift has pushed the business to “innovate faster” and “serve couples better” rather than relying on tradition alone.
Beyond these initiatives, the shop has embraced a one-stop wedding approach, aiming to position itself as a go-to destination where couples can source more than just traditional items.
For instance, the company also facilitates wedding cake and pastry orders by partnering with local bakeries such as Baker’s Brew, Tong Heng, and Thye Moh Chan.
Another way the shop is positioning itself as a one-stop wedding destination is by expanding beyond retail into an advisory and educational platform. Social media has become a key channel for the business to educate younger couples about traditional wedding customs.
Advertisement
In this video, Michelle breaks down what is needed for a Teochew family to prepare for their Guo Da Li ceremony.
“The goal is to reduce stress for couples while keeping traditions meaningful, not burdensome,” she added.
Diversification is key to survival, but weddings remain their bread and butter
Michelle speaking at a wedding fair as a vendor./ Image Credit: The Chinese Wedding Shop
Since renewing its offerings post-pandemic, Michelle shared that they have been well received by both couples and parents alike, though she did not elaborate further.
Even so, the shrinking number of marriages means the overall market is likely to continue contracting, raising the question of whether the company should diversify beyond weddings.
Michelle and her team have explored this potential, considering expansions into other Chinese traditions—such as selling festive banners and red packets for Chinese New Year—but plans aren’t concrete yet, and any move into new areas would need the same level of cultural sensitivity, knowledge, and relevance.
Weddings continue to be the business’s bread and butter as of now, as the credibility they have gained over the years allowed them to establish a niche in Singapore’s crowded wedding scene.
Advertisement
“For now, our priority is to deepen our wedding-related offerings, such as rental sets for specific uses and modernised solutions, before extending into other areas.”
The market for high end wireless headphones has expanded rapidly over the past few years. What was once dominated by mainstream Bluetooth models has evolved into a category that now includes serious audiophile contenders from brands such as Focal, Bowers & Wilkins, DALI, Mark Levinson, and Sennheiser. These companies have demonstrated that wireless headphones can deliver a level of performance that appeals to listeners who once insisted on wired designs and dedicated headphone amplifiers.
HiFiMAN is now pushing deeper into that space with the introduction of the HE1000 WiFi and Arya WiFi, two open back planar magnetic headphones that rely on Wi-Fi rather than Bluetooth as their primary wireless connection. The company has experimented with wireless concepts before, but these new models represent a more ambitious attempt to bring high bandwidth wireless audio to planar magnetic designs.
HiFiMAN HE1000 WiFi
HiFiMAN has not announced pricing yet, but the company has indicated that both models will sit closer to the HE1000 Unveiled and Arya Unveiled in its lineup rather than its flagship tier. Both headphones are scheduled to begin shipping next month and will be demonstrated publicly at CanJam NYC 2026 this weekend, where we will have an opportunity to spend time listening to both models.
According to HiFiMAN, the key difference between these headphones and typical wireless designs is the use of Wi-Fi for audio transmission. Bluetooth’s limited bandwidth has long constrained wireless audio quality, while the Wi-Fi connection used here is designed to support full resolution lossless audio streams without compression.
Advertisement
Both headphones incorporate HiFiMAN’s proprietary Hymalaya R2R DAC, integrated amplification inside the earcups, and planar magnetic drivers based on the company’s established technologies.
HiFiMAN Arya WiFi
Hymalaya What?
The Hymalaya DAC is HiFiMAN’s proprietary digital to analog converter built around a classic R2R ladder architecture, a design approach favored in many high end audio systems for its natural timing and accurate signal conversion. Unlike the delta sigma DAC chips used in most modern headphones and wireless devices, an R2R DAC converts digital audio using a network of precision resistors that translate binary data directly into analog voltage. This approach can deliver excellent transient response and tonal accuracy, but it is traditionally more complex and power hungry than conventional DAC designs.
HiFiMAN developed the Hymalaya DAC to overcome those limitations by combining the R2R ladder with an FPGA controlled architecture and extremely low power consumption. The result is a compact DAC capable of supporting high resolution audio formats, including PCM up to 768 kHz and native DSD, while drawing far less power than traditional ladder DACs. That efficiency allows HiFiMAN to integrate the technology into portable gear and wireless headphones.
In products such as the HE1000 WiFi and Arya WiFi, the Hymalaya DAC works alongside a built in amplifier to convert digital audio directly inside the headphone. This self contained signal chain allows the headphones to operate more like a complete playback system, handling the digital conversion and amplification internally rather than relying on the DAC and amplifier inside a phone or computer.
Advertisement
HiFiMAN HE1000 WiFi
The HE1000 WiFi is an open back planar magnetic headphone that combines HiFiMAN’s familiar driver architecture with onboard digital processing and wireless connectivity. The design incorporates the company’s Nano Diaphragm driver paired with its Stealth Magnet system, a magnet structure intended to reduce wave diffraction and maintain a more consistent sound wave path.
Inside the earcups is a custom Class A/B balanced amplifier working alongside HiFiMAN’s Hymalaya R2R DAC, which converts incoming digital audio streams directly within the headphone. This approach allows the headphone to function more like a self contained playback system rather than relying on the amplification stage of an external device.
The HE1000 WiFi connects through a dedicated Wi-Fi network created by the headphone itself. Users simply select the headphone from the WiFi menu on a smartphone or tablet and stream audio directly to the headphone. This wireless link supports full resolution audio transmission and is capable of handling high bandwidth formats including PCM up to 768 kHz and native DSD up to DSD512.
Advertisement. Scroll to continue reading.
HiFiMAN lists a frequency response of 8 Hz to 65 kHz with THD + N rated at 0.009 percent at 32 ohms when the DAC and amplifier are operating together. Channel separation is specified at 105 dB at 1 kHz, and the headphone weighs 452 grams. The HE1000 WiFi can operate in Wi-Fi, USB audio, or Bluetooth modes and charges via USB Type-C with a typical charging time of three to four hours. Standby time is rated at more than 30 days.
HiFiMAN Arya WiFi
The Arya WiFi uses a similar architecture but incorporates HiFiMAN’s Super Nano diaphragm driver, a thinner variation of the company’s planar driver design intended to improve transient response and overall efficiency. As with the HE1000 WiFi, the Arya WiFi also uses Stealth Magnet technology, integrated amplification, and the Hymalaya R2R DAC.
Like its sibling, the Arya WiFi connects directly to smartphones and tablets through a Wi-Fi network created by the headphone. Once connected, audio streams are transmitted at full resolution without relying on Bluetooth compression. Support for high resolution audio formats is extensive, with PCM playback up to 768 kHz and native DSD up to DSD512.
Advertisement
HiFiMAN specifies a frequency response of 8 Hz to 55 kHz with THD + N rated at 0.009 percent at 32 ohms when the DAC and amplifier are operating together. Channel separation is listed at 105 dB at 1 kHz and the headphone weighs 452 grams.
The Arya WiFi supports Wi-Fi streaming, Bluetooth 5.1 connectivity with SBC, AAC, aptX, aptX HD, and LDAC codecs, and USB-C wired playback. Battery life is rated at approximately 6.5 to 7.5 hours when operating in Wi-Fi mode and up to 23 hours when using Bluetooth. Charging takes roughly three to four hours, and standby time is listed at more than 30 days.
The HE1000 WiFi and Arya WiFi take a different approach to wireless headphone design by focusing on bandwidth and signal integrity rather than convenience features. By combining planar magnetic drivers with a built-in Class A/B amplifier, HiFiMAN’s Hymalaya R2R DAC, and a Wi-Fi based audio connection, these headphones are designed to handle high resolution audio without relying solely on Bluetooth compression.
That said, these are clearly not lifestyle wireless headphones. Both models use open-back earcups, which means they leak sound and provide no isolation from the outside world. They are better suited for listening at home, where wireless freedom can be useful without the compromises that normally come with portable wireless designs.
HiFiMAN has not revealed pricing yet, but both models appear positioned for listeners who want something closer to a traditional audiophile headphone system that happens to operate wirelessly. We expect to learn more once we have the opportunity to spend time with both models at CanJam NYC 2026 this weekend.
We test a lot of laptop docking stations and, quite frankly, most of them are … fine. They’re fine! We get into the nitty-gritty for specific use cases to find the best, but that leaves a bunch of devices that are great options even if they don’t make our top picks. Here’s a selection of some of our favorites, past favorites, or just alternatives to our picks above.
Satechi Triple 4K Docking Station
Photograph: Eric Ravenscraft
Satechi Triple 4K Docking Station for $300: Satechi’s Triple 4K Docking Station supports three monitors, and while the first display output is HDMI-only, the two others can be connected via either HDMI or DisplayPort. Each display supports up to a maximum of 4K resolution at 60 Hz, which is more than enough for most office or media work, though don’t expect it to support the high frame rates you might want for gaming. —Eric Ravenscraft
Plugable USB-C Dual HDMI Display Dock for $120: Sometimes, all you need is a quick and easy way to plug your laptop into a couple of monitors—preferably without spending hundreds of dollars. It supports two monitors via HDMI and includes a healthy array of ports to connect the rest of your accessories. So, while there are certainly more up-to-date options out there, this is an affordable way to get some basic connectivity.
Advertisement
Kensington Triple Video Mobile Dock for $83: A mobile docking station might sound like a contradiction, but in the case of the tiny Kensington Triple Video Mobile Dock, it makes a lot of sense. Using the included two HDMI ports and the DisplayPort, this little device can power three 1080p displays or two 4K displays—all at 60 Hz. It also has a USB-C port with 85 watts of pass-through charging, which is enough to charge most laptops. The downside is that it only supports a single 4K monitor on MacBooks, as the dual 4K support is only for Windows devices.
Sonnet Echo 13 Thunderbolt 5 Dock
Courtesy of Luke Larsen
Sonnet Echo 13 Thunderbolt 5 Dock for $440: Sonnettech’s Echo 13 was one of the first Thunderbolt 5 docks out on the market. As it turns out, it’s also one of the most unique offerings out there, including just about every port imaginable as well as an integrated M.2 storage slot with a Kingston SSD inside. While handy, I don’t like that the drive isn’t user accessible. The cheap plastic chassis is disappointing for the price, too.
Advertisement
Ivanky FusionDock Max 1 Thunderbolt 4 Docking Station for $380: It’s hard to overstate how excessively luxurious this dock is. It’s specifically for MacBook Pro users and can tackle up to four 6K screens, something only recent MacBook Pros support. The Ivanky FusionDock Max 1 accomplishes this via four USB-C Thunderbolt 4 ports, each capable of 40 Gbps data transfer speeds. If you’re building the beefiest media workstation you can for the most powerful MacBook Pros on the market, this is it. Just put it all on the company card, because it’s expensive. —Eric Ravenscraft
Ugreen Revodok Max 213 Thunderbolt 4 for $228: Few people need an 8K display—or multiple 4K displays—but those who do know how difficult it can be to find gear that supports their needs. Fortunately, the Revodok Max 213 from Ugreen fits that bill. The DisplayPort 1.4 port can handle up to an 8K display at 30 Hz. It also comes with a Thunderbolt 4 upstream port that runs to your laptop, and, more importantly, a pair of downstream Thunderbolt 4 ports, which is another rarity among the docks I’ve tested. If you need to transfer a ton of media from various sources into one machine, connected to seriously high-res displays, this is the dock that can handle it all. —Eric Ravenscraft
Advertisement
Do You Need a Docking Station or a USB Hub?
This is the big question you’ll want to answer before moving forward. Chances are you know if you need a full-on docking station rather than just a USB hub, but I’ll explain the differences in case you’re on the fence. A simple USB hub will handle most people’s needs, as the device will expand the potentially very limited ports of your laptop. If you own a MacBook Air, for example, a USB hub functions as a multiport adapter to get you HDMI, USB-A, and more. They’re intended to be portable and many of these hubs even include HDMI to connect an external display.
A laptop docking station will do quite a lot more. These devices are meant to be stationary on a desk, enabling you to access your entire workstation setup with just a single USB-C cable. The docking station is meant to stay put and have all your monitors and accessories plugged into it. Because of that, they require significantly more power and are often bundled with a large power brick. They are often quite expensive. So, while both accessories connect your laptop to more ports, they serve two different functions.
There are now lots of docks and hubs that blur the lines, offering multi-monitor support in a very small package. These can be useful, but a full docking station will still give you the fastest transfer speeds, the most ports, and better external display support all through a single cable.
Advertisement
What Ports Should Your Docking Station Have?
Figuring out the right connections you need for your setup can be daunting, and the confusing, arcane USB terminology only makes it worse. You can check out our explainer on parsing USB terms here. For the short version, here are the basics you should keep in mind:
Check your ports’ speeds, and don’t rely on version numbers. For a lot of confusing reasons, ports labeled as USB 3.0, 3.1, and 3.2 can all have the same speed or wildly different speeds. For this reason, docking station manufacturers have recently started opting to add speeds (usually written like “5 Gbps”) directly onto individual ports. Use the faster ports for transferring data, and slower ports for things like your keyboard and mouse.
Thunderbolt is best for lightning-fast data transfers, or high-res displays. Thunderbolt is like a supercharged version of USB, and it even uses USB-C ports. However, Thunderbolt ports are capable of transferring massive amounts of data. This makes it ideal for things like moving uncompressed video files around, as well as things like 4K (or even 8K) displays or lower-resolution monitors with extra high refresh rates.
Advertisement
Keep in mind your power needs. Most laptop docking stations will have some form of power connector and USB Power Delivery (or USB-PD) that can send power through to your laptop. You’ll also sometimes see this referred to as “pass-through charging.” Most devices you connect will require their own power as well, especially if you want to connect monitors or charge your phone and tablet. If you plan to connect a lot of power-hungry devices, make sure your docking station can handle your needs.
Upstream and downstream ports. You’ll often see USB ports labeled either as upstream or downstream. The data either flows up to the source (your PC or docking station) or down from the source. An upstream USB port means it’s meant for transferring data from a peripheral (like an external drive) to your PC, whereas a downstream USB port only works in the opposite direction.
All the docks in our recommendations are compatible with both Mac and Windows, unless otherwise noted. But there are lots of hubs and docks out there that have certain limitations on Mac, such only supporting mirroring mode in dual monitors. That’s not a problem on Windows.
On lower-end Macs there is a limitation on the number of screens. There is a way aroudn this if you use a dock that supports DisplayLink. Software can create a “virtual GPU” that tricks the system into allowing for additional displays so you can drive more displays than is typically allowed on a MacBook Air, for example. In my experience, however, the performance can be shoddy, and you may run into issues with latency.
Advertisement
Is Thunderbolt 5 Worth It?
The first Thunderbolt 5-capable PCs, docks, and accessories came out in 2024. Thunderbolt 5 can now handle three 4K displays at 144 Hz (or two 4K displays at 240 Hz) and can deliver up to 240 watts of power. That’s dramatic, up from the 100 watts of Thunderbolt 4. Thunderbolt 5 allows you to fully juice up more powerful devices, such as gaming laptops or the 16-inch M4 Max MacBook Pro.
Thunderbolt 5 docks are all backward compatible, so there’s no worry about outdoing the peripherals you currently own. As is true in many scenarios, buying the latest specs are often worth it to avoid having to upgrade later. However, the adoption of Thunderbolt 5 has been slower than I’d hoped for. And if you don’t have a Thunderbolt 5 laptop to connect to, you won’t get the full benefit that Thunderbolt 5 offers.
While there’s a breadth of Thunderbolt 5 docs out in the world (many of which you’ll see on this list), the biggest disappointment has been in the lack of Thunderbolt 5 accessories to come out over the past year. It’s still very difficult, for example, to find a Thunderbolt 5 SSD, to achieve those faster speeds. Thunderbolt 5 docks are sometimes only marginally more expensive than previous-generation options, so they’re often worth buying for the improved display support or higher power delivery. For example, CalDigit’s TS5 is only $20 more than the TS4.
Advertisement
Power up with unlimited access to WIRED.Get best-in-class reporting that’s too important to ignore for just $2.50 $1 per month for 1 year. Includes unlimited digital access and exclusive subscriber-only content. Subscribe Today.
In the past, smartphones with smaller form factors usually came with a couple of compromises, and battery life was often the first casualty. But OnePlus’ next palm-friendly flagship might just outlast most of your big premium devices.
Going big on battery and charging
OnePlus 13TOnePlus
The company has confirmed that the upcoming OnePlus 15T will pack a massive 7,500mAh battery, which dwarfs the cells found in tall phones like the Galaxy S26 Ultra. After confirming a major camera update, OnePlus China president Li Jie Louis revealed the battery and charging specs of this device.
The details were shared in a post on the Chinese social media platform, Weibo. According to the executive, the OnePlus 15T will feature what the company calls a “Glacier Battery”, which is a silicon-carbon-based battery technology that is designed to pack higher capacity without making the device significantly thicker.
A big battery is only part of the story. The OnePlus 15T will also support 100W wired and 50W wireless fast charging, ensuring quick charging and shorter downtime. The giant battery is quite impressive considering the device is expected to sport a 6.3-inch display, similar to the Galaxy S26 and Pixel 10.
OnePlus 13TOnePlus
Compact size but big on features
Aside from the huge battery, this device will even offer flagship-level performance. It is expected to be equipped with Qualcomm’s Snapdragon 8 Elite Gen 5 SoC and will likely run on the Android 16 OS out of the box. Louis also confirmed a notable upgrade in design with the OnePlus 15T featuring IP66, IP68, IP69, and IP69K ratings for water and dust resistance. To recall, the OnePlus 13T only shipped with an IP65 rating for protection against the elements.
