Among so many other technological advances, the Cold War saw the advent of the ballistic missile submarine. The concept was simple—pack enough nuclear warheads to destroy a small civilization into a compact metal tube, and then hide it underwater. The oceans would act as a cloak for your fleet of world-enders, and keep your enemies forever on their toes. A terrifying machine that could both start and end a war with the push of a button.

Most nation states are populated by humans with the will to live. Thus, there has been a great incentive to find ways to keep tabs on these sunken doombringers. Great efforts have gone into improving sonar and magnetic detection methods over the decades, which are the bread and butter of sub hunting to this day. However, military researchers have also explored the prospect of whether submarines could be detected via their effect on the gravitational field alone.

Do You Feel It?

The simple matter is that every object with mass has its own gravitational field. We don’t typically think about it, because gravity is the weakest of the fundamental forces. On anything less than a planetary scale, it’s generally not obvious to us in our daily lives. However, submarines are quite heavy and large, particularly those that are armed with a complement of nuclear-capable ballistic missiles. Thus is raised the prospect of detecting these massive objects via their perturbations to the local gravitational field. This has been a hot-button news item in military commentary circles of late, with much bluster that advanced measurement equipment could potentially render the ocean transparent and reveal the locations of submarines at great distances.

Naturally, it’s difficult to comment accurately on top-secret military capabilities from a civilian viewpoint. Such a technology would be game-changing in a strategic sense, to the point that any nation state with such a capability would have great reason to keep its existence strictly hidden. However, there is some literature on the topic that is in the public domain, which discusses just how hard this feat would be to execute in practice. A great example is a report prepared by the Pacific-Sierra Research Corporation in 1989, under the sponsorship of the Naval Air Development Center.

How It Works

When it comes to detecting the gravitational anomaly of a submarine, you might think it would be easy given the sheer mass of such a craft. However, the way submarines operate frustrates this at a very fundamental level. In normal operation, a submarine is neutrally buoyant, displacing an amount of water roughly equal to its own mass. Thus, the submarine is not really distinguishable from the water around it in terms of its first-order effect on the gravitational field, being roughly as heavy as the water that would otherwise be there.

There is a wrinkle, though, in that a submarine is bottom-heavy for the sake of stability. This does create a variance in the gravitational field versus the otherwise uniform field in open water, and it’s one that could theoretically be detectable with a sensitive enough apparatus.

The device used for measuring gravitational variation is called a gravimeter. They are essentially a special-case variant of accelerometer, specifically designed to very accurately measure the local acceleration due to gravity at a single point. Then there is the gravity gradiometer, which measures the spatial rate of change of gravitational acceleration. By virtue of measuring acceleration gradients, a gradiometer is not sensitive to the acceleration perturbations of a moving platform, making it particularly useful for use in a moving frame of reference such as towing behind a ship or aircraft. Various types of each instrument exist, from portable units to high accuracy laboratory instruments; creating an exhaustive list of all  variants is outside the scope of this article. The real question is, based on the gravitational anomaly generated by a large submarine, to what useful range could a gravimeter or gradiometer detect one?

Unfortunately, the maths says that you have to get very, very close. In the 1989 study, calculations suggested the best gravimeters and gradiometers in the world would maybe be able to pick up a large submarine from a distance of tens of meters, at best. The simple problem being that the gravitational anomaly generated by an underwater submarine, and the gradient of that anomaly, are both so small, that even highly sensitive instruments would struggle to pick it up when the submarine is practically in visual range. Even if the problem were simplified, and one were trying to detect a submarine as a heavy point mass in empty space, detection ranges would stretch to somewhere in the range of 100 meters at most. Of course, this would be largely irrelevant due to the neutral buoyancy considerations explained above.

It’s true that technology has moved on since 1989. We have more advanced gravimeters and gradiometers available now, including quantum units with greater sensitivity than ever. And yet, even with these advances, it would be still be a struggle to detect a submarine at useful range. Sensitivities would have to jump by four or five orders of magnitude to enable detection at ranges of 1000 meters. Even still, if this were achieved with some highly classified system, it would still be relatively limited in capability versus more established techniques in magnetic or acoustic detection.

The parameters of the problem, combined with the sheer weakness of gravitational forces, means that gravitational detection is not some silver bullet for tracking enemy submarines at great range. While it would be desirable to have some kind of sensor that could reveal where these nuclear weapon platforms are lurking at all times, that technology seems beyond the reach of even the most capable navies at this time. For now, strategic planners will continue to sweat over the threat these weapons pose, never quite knowing whether they’re lurking just off the coast or half a world away.

For many participants across insurance and property ecosystems, current valuation challenges did not emerge from a single moment or event. Multiple economic cycles, demographic shifts, evolving construction practices, and changing consumer expectations have accumulated over time, creating a landscape that can sometimes feel difficult to fully trace. Limited visibility into how these layers developed has made present-day valuation conversations increasingly nuanced. Within this environment, e2Value has focused on studying the long arc of property intelligence, using historical context alongside modern technology to support a more informed view of valuation.

Current discussions around property valuation often begin with present conditions, although a broader historical lens offers an important perspective. “In earlier decades, people generally saw homeownership through a simpler financial lens,” Todd Rissel, co-founder and CEO of e2Value, says. “A house was viewed as simply a place to live, and estimating its value tended to feel more straightforward because prices, replacement costs, and household budgets usually stayed closer together.”

Over time, economic expansion, rising residential values, and changing patterns of household wealth introduced additional layers of complexity. According to Rissel, during the 1980s and 1990s, evolving building standards, energy-related considerations, and increased private investment into residential markets contributed to an environment where homes became larger financial commitments for many families. Property ownership increasingly carried dimensions extending beyond shelter and into long-term wealth accumulation.

As these changes unfolded, distinctions between purchase price, market value, replacement cost, and appreciation became more difficult for consumers to separate. A home’s purchase price could reflect market demand and neighborhood dynamics, while replacement cost involves an entirely different calculation based on the availability of labor, quality of materials, and local market conditions.

Rissel offers a broader perspective on this evolution. He says, “Every economic era leaves behind a set of assumptions that people tend to hold onto. Those assumptions can keep shaping decisions long after the conditions that formed them have shifted.”

Additional economic developments introduced another layer of complexity. For many years, rising incomes and broad access to capital moved alongside increasing property values. Following the financial disruption of the late 2000s, that relationship entered a different phase. Property values remained elevated across many regions while access to capital moved through a period of recalibration, creating a longer-term adjustment across housing and financial systems.

Valuation standards increasingly became part of a larger conversation involving affordability, underwriting, and risk assessment as those conditions continued to develop. Consumers often encountered difficulty distinguishing between the value associated with purchasing a home and the cost associated with rebuilding that same structure.

Present conditions reflect the accumulation of these historical developments. According to Deloitte’s 2026 Insurance Industry Outlook report, insurers are navigating expanding complexity linked to catastrophic events, changing customer expectations, and rapidly evolving technology environments. The report also notes that technology priorities are increasingly focused on strengthening data foundations and creating systems capable of supporting meaningful AI implementation.

This progression naturally brings AI into the discussion. Across insurance ecosystems, AI has generated significant interest because of its ability to process extensive amounts of information and identify patterns across large datasets. EY reports that many senior leaders are reevaluating enterprise AI strategies with greater emphasis on long-term value creation, data quality, and future adaptability.

Still, technology itself represents only part of the equation. AI systems rely heavily on the information used to train and guide them. Historical context becomes particularly important because valuation is influenced by decades of construction practices, regional patterns, economic cycles, and changing cost environments.

Rissel explains this relationship through a broader lens. “Technology can organize and process information at remarkable speed, although meaningful insight often begins with understanding where the information originated and why it matters,” he says. “Data gains value when context travels alongside it.”

This perspective aligns closely with e2Value’s own role within the valuation ecosystem. Since its founding in 2000, the company has focused on preserving historical property intelligence while integrating modern modeling capabilities into its valuation platforms. That role extends beyond creating software designed to generate estimates.

“We’ve spent years examining how structures behave as economic assets that are influenced by construction practices, local market dynamics, material costs, labor conditions, and broader macroeconomic patterns,” Rissel shares. Through that work, the company developed valuation methodologies intended to reflect the larger environment surrounding a property instead of viewing a structure only as an isolated collection of components.

That perspective has also influenced how e2Value interacts with different parts of the insurance ecosystem. Across underwriting, risk evaluation, claims environments, and portfolio management discussions, valuation increasingly functions as a source of insight that supports larger decisions. Replacement cost estimates can influence coverage discussions, portfolio assessments, and broader conversations surrounding risk exposure. As data expectations continue expanding across the industry, connecting those elements within a unified framework has become increasingly important.

Overall, the long arc of valuation shows that today’s complexity is a product of decades of shifting economics, construction practices, and consumer expectations. As Rissel emphasizes, assumptions formed in earlier eras still shape how people interpret value, even as conditions evolve. In this environment, e2Value helps insurers, underwriters, and property stakeholders navigate a landscape where replacement cost, market value, and risk are increasingly intertwined by grounding advanced modeling in the context that gives data meaning.

What do you do when you’re one of the world’s top Smart TV and Connected TV platforms? You double down on what people love about your system while making the whole user experience even more streamlined and personalized for each individual viewer.

While Android TV/Google TV may lead the global streaming TV platforms in overall market share, Roku is #1 in North America, with more than a third of all streaming viewers. With a combination of streaming sticks which can turn any TV into a Smart TV, Roku branded Smart TVs and partnerships with TV makers such as Hisense, Philips, and ONN, Roku has established itself as the go-to streaming platform for over 100 million TV viewers and movie lovers.

While alternative Smart TV platforms like Amazon’s Fire TV and Google TV tend to push certain streaming services or shopping services over others, Roku has earned high marks from its users for providing a more service-agnostic user experience. And the company seems to be leaning into this philosophy with significant updates to its user interface.

At an event in New York City this morning, Roku unveiled a new version of their Home Screen which promises a more dynamic, smarter experience. Featuring more relevant recommendations and faster pathways to content, the new Roku Home Screen is designed to reduce friction, maintain Roku’s signature simplicity, and help viewers find their next favorite movie or TV show with ease.

While Roku typically makes running tweaks to their user interface and home screen to improve the overall user experience, today’s update to the Home Screen is the most substantial and significant update we’ve seen to the platform in the past ten years.

Company reps told us that the change wasn’t made just for the sake of change, but was informed by direct customer feedback as well as detailed analysis of viewer behavior, with the end goal of streamlining and optimizing the user experience. The new personalized Home Screen tackles the biggest challenges in streaming, while offering a tailored, content-forward way to start watching.

Roku executives at the event stressed that they wanted to make sure they maintained the snappy, responsive feel that Roku users love while improving every element of the user experience. The company has done extensive testing with employees, and actual customers over its development, even allowing customers to opt in early to explore the new features and provide feedback. to the development team along the way.

“When we set out to rethink the Home Screen, we knew we should listen to the people who use it every day,” said Anthony Wood, Roku’s Founder and CEO. “So we talked to the viewers, we tested extensively, and we pushed until the design and the data lined up for a meaningful update. Now, our new Home Screen puts entertainment at the center of everything, while staying true to Roku’s simple, intuitive roots. More than 100 million households will feel the difference the moment they turn on their TV—and it opens up a better, more powerful experience for our partners as well.”

To take things from abstract to actual, one specific data point was critical in inspiring this set of changes: A majority of Roku customers (82%) said they would love it if they turned on their TV and the show they wanted to watch was right on their Home Screen. The new Roku Home Screen does just that, recommending content based on the viewer’s past activity and interests, helping the viewer get up and watching the content they crave quicker. Out of the billions of potential Home Screen layouts, the new system is designed to offer just what you want to see, when you want to see it, for each individual viewer.

Preston Smalley, Roku’s Vice President of Viewer Product, told event attendees that the net effect of the change has been moving from a static Home Screen to a fluid experience that is more dynamic and personalized around each individual user and which changes based on evolving tastes and needs. In Preston’s words, “We want the UI to recede into the background. We want the content to really shine.”

Highlights of the New Home Screen Include:

• Quick Access for your most used apps, continuously adapting to your routine
• An intelligence-driven and expanded content-first “Top Picks for You” section
• New genre-based destinations such as:
  • “For You,” built on your interests and filled with fresh personalized picks
  • “Subscriptions,” allowing for a convenient way to browse and discover from across all your subscriptions in one place
• Search in key destinations with relevant suggestions and results
• A streamlined collapsible menu
• Elevated shortcuts for everyday actions including Save List, Continued Watching, and more
• “Your Daily Scoop,” a dynamic row that brings you a curated digest of breakout shows and cultural trends
• A “Roku City” tile, taking you to an interactive version of your favorite screensaver

The new Roku Home Screen begins rolling out today (May 27, 2026) across all Roku TVs and streaming devices in the United States. Expansion to additional countries will follow in the coming months.

The Bottom Line

Any time we hear about a major user interface redesign rolling out to all devices, we wonder whether it will be a home run, which users of the platform embrace, or it will lead to the disaster that plagued Sonos just a few years ago. From what we saw at the event, we believe the new Home Screen redesign builds on the simple service-agnostic user interface that has built up Roku’s loyal fan base over the years. If the rollout goes well, these changes should help cement the platform’s popularity in the coming years.

Related Reading:

Roku Unveils New TVs and Slimmest Streaming Sticks Ever

• Third-party UK Visa Portal website exposed 100,000 docs in an unsecured cloud repository
• Cybercriminals with access to the affected PII could conduct identity theft or fraud
• Victims advised to protect and monitor accounts, and await notification

UK Visa Portal, a third-party website separate from the official government offering, has reportedly left thousands of highly sensitive documents exposed in a major data leak.

Affected documents and details include passports, photos, verification selfies and other application information, leaving victims widely open to identity theft and potential financial fraud.

Karin Verspoor of RMIT University explores how AI is impacting research in STEM.

Many of the most exciting discoveries in science involve highly specialised knowledge and making connections between far-flung facts. Scientists must combine deep analysis with broad reasoning strategies.

As in many information-rich tasks, researchers are looking to artificial intelligence (AI) systems to speed up their work. AI tools may be able to support key steps such as generating ideas, reviewing existing work and analysing data.

The latest systems use large language models (LLMs) to allow scientists to interact naturally and directly with the vast body of knowledge captured in words in the scientific literature.

But as two new systems described in papers just published in Nature show, when it comes to science, language alone can only go so far.

What AI is doing to science

A number of organisations, such as Sakana AI, are trying to automate the entire scientific process. To date, these efforts have largely focused on computer science, where ‘experiments’ mainly involved designing and writing code.

However, the Agents4Science conference organised at Stanford last October showcased a broader range of AI-generated papers. They covered topics from mechanical engineering and protein design to a system called BadScientist which deliberately produced “convincing but unsound” research.

I have previously raised concerns about the impacts of AI scientists on the scientific ecosystem. Recent work validates these concerns, showing increased quantity but lower quality of both papers and peer reviews, identifying fabricated references in published works, finding fabricated and misleading images, and more.

What scientists are doing with AI

AI systems clearly can’t be trusted to conduct the full process of science on their own. But how about using AI to help scientists get more done more quickly?

This is the intent of the two new systems described in Nature: Robin, made by non-profit Future House, and Co-Scientist, from Google DeepMind.

Both systems aim to accelerate scientific discovery, working in collaboration with a scientist. Both are also ‘multi-agent’ AI systems, meaning they are built as a collection of specialised agents each targeting specific steps of the scientific discovery process, coordinated by a ‘supervisor’ agent.

The agents that comprise Co-Scientist aim to mirror abstract cognitive tasks, such as a ‘reflection agent’ that acts as a critical scientific peer reviewer assessing the quality of a hypothesis. ‘Ranking agents’ debate research hypotheses in ‘tournaments’, using multiple interacting LLMs to simulate a discussion about the relative merits of two hypotheses.

Robin’s agents, on the other hand, are more tuned to specific tasks relevant to drug repurposing, aiming to identify new drugs for a given disease. One agent focuses on selecting experimental tests, while another analyses complex biomedical data.

How do the results stack up?

Co-Scientist can assess the quality of its generated proposals, using a method called the Elo rating which is best known for ranking chess players. Co-Scientist’s self-ratings of the novelty and impact of its outputs align quite well with the preferences of human experts and judgements by other LLM systems.

In a drug repurposing experiment, Co-Scientist selected 30 drug candidates as promising treatments for a kind of cancer called acute myeloid leukemia. Expert (human) oncologists refined the list, and five drugs were tested in the lab. Of these, three showed some positive results and one seemed to show particular promise.

Other experiments showed the potential of Co-Scientist to explore combinations of multiple drugs.

Notably, the predictions of Co-Scientist were not compared with the plethora of targeted computational and machine learning methods for drug repurposing that have been developed over decades of computational biology research. This means we don’t know whether the new general-purpose tool outperforms more specific AI approaches.

Both systems stop short of validating their hypotheses directly, which would involve real physical experiments. Both also rely heavily on human input to define the key scientific question, sense-check predictions and prioritise predictions for further investigation.

Co-Scientist focuses primarily on generating hypotheses through elaborate reasoning agents, leaving validation and interpretation to subsequent steps. Robin also uses an agent to analyse data produced from real-world experiments.

Robin was used to propose 30 drug candidates for a condition called dry age-related macular degeneration. The top five were selected for testing.

Robin also made proposals for the experiments, with several suggestions overridden by the human scientists. Through several rounds of brainstorming and analysis, two drugs were identified as promising.

Testing of Robin’s individual agents showed those that dug through earlier research were better at the task than general-purpose LLMs. The analytical agent did less well on questions about statistics and bioinformatics, and relied heavily on human-supplied prompts.

The limits of language alone

AI can help scientists to navigate the vast amount of documented knowledge humans have acquired over the millennia. Use of computation to find patterns in large datasets, to integrate dispersed information, and to drive new discoveries from existing literature has already contributed to scientific progress for decades.

New models such as Robin and Co-Scientist represent a shift towards working directly in the realm of the language of science, rather than the realm of raw data. This allows more natural collaborations between scientist and machine, through language-based ‘discussions’.

However, more natural doesn’t necessarily mean more effective. Language-based communication can be imprecise and ambiguous, where science must be specific.

Models that combine the best of these worlds are on the horizon. These aim to link structured quantitative data to the concepts and relationships that describe the core facts beneath it.

Such models ground scientific reasoning in the structure of knowledge. They allow scientific evidence ranging from genomic sequences and protein structures to cellular imaging to be connected.

Words are how science is communicated. AI tools that facilitate making sense of the information that is hidden in all of those words are surely valuable. But the complexity of the natural world means that AI (co-) scientists will only be truly effective when they can go beyond connecting words together, to modelling the full complexity of the systems those words describe.

By Karin Verspoor

Karin Verspoor is dean of the School of Computing Technologies at RMIT University. She works at the intersection of science and technology, applying artificial intelligence methods to analysis and interpretation of biological and clinical data, focusing particularly on natural language processing of unstructured text data. She is a fellow of the Australian Academy of Technological Sciences and Engineering and the Australasian Institute of Digital Health. She is also the co-founder and Victoria node lead of the Australian Alliance for Artificial Intelligence in Healthcare.

Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.

As we get out of the house, the gear-obsessed WIRED Reviews team is writing about our favorite bags and EDCs. Today, reviewer Louryn Strampe raves about her Osprey Farpoint 40 backpack. You can also read other Bag Check stories where WIRED writers share their carryall of choice.

While planning our 20-day trip to Asia in 2018, my boyfriend at the time was adamant that neither of us would check any luggage. As a proud overpacker, this intention both shocked and horrified me. I love options and hate paring things down. I wanted to bring 30 pairs of shoes and 348 pairs of underwear; I certainly did not want to painstakingly build a capsule wardrobe and strategically compile packing cubes. Ultimately, though, I agreed to the single-bag trip, and the Osprey Farpoint 40 quickly converted me to legitimately loving the light-travel life.

If you want the perfect backpack, or the most durable suitcase, or the best tote for toting your toteables, my colleagues have plenty of recommendations worth browsing. But if you seek a bag that makes carrying-on (to the plane, train, or automobile) and carrying-off (from hotel to hostel to hotel again) a blissfully pain-free experience, the Farpoint is my favorite.

During that nearly three-week trip in 2018, my boyfriend and I were constantly on the move, visiting Shanghai, Tokyo, Osaka, Kyoto, Busan, Seoul, and Hong Kong. We took multiple flights and trains, and we stayed in both spacious rooms and cramped quarters. In the years since, I’ve taken my Osprey Farpoint to the East and West Coasts of the United States. It’s lived with me in three homes, joined me on dozens of road trips, and stayed with me in hundreds of hotels. I’ve used it for trips as long as three weeks and as quick as a single night. I’ve thrown it down flights of stairs, sat atop it on subways, used it to shield my head from the rain, crammed it into overstuffed vehicle trunks, and packed it to the brim nearly every time I’ve taken it anywhere. It’s still working just as well as it did on that first trip.

Needless to say, this bag is cavernous. Its 40-liter capacity lets you load it up to the brink of being too hefty to carry, but I’ve never had a problem bringing it onto a plane, partially thanks to the compression straps that help you squish down the silhouette. (Per Osprey’s website, the Farpoint meets domestic carry-on size requirements.) And it offers so many pockets, they’re difficult to track.

The outer shell has two mesh spaces that are perfect for shoes or water bottles, plus a smaller compartment where I like to stash my keys and passport. The pack itself features two main chambers, the first of which is a laptop compartment, complete with a zippered sleeve ideal for an e-reader or a tablet. The largest body pocket features two built-in compression straps to help you achieve that “sit on the top of the suitcase to close it” effect once you’re ready to zip up. On the opposite side, there’s another zippered mesh pocket that spans the entirety of the shell, which I use to stash my socks, underwear, toiletries, and other items that I need handy but don’t want mixed in with my clothes.

Somehow, the Farpoint makes it not only possible but also comfortable to carry everything I might need. Yes, when you wear a stuffed-to-the-brim Farpoint, you’ll resemble a turtle peeking out from under a shell. But you won’t need to move slowly, thanks to its stabilizing design. The shoulder and hip straps are padded, and there are clip straps for your hips and chest. The chest strap also has a built-in whistle, which won’t necessarily boost your comfort, but it might come in handy if you’re hiking. (I mostly use the whistle for funsies or to bother my campmates at festivals.)

Samsung’s 10.5%-of-profit bonus formula is only the second written profit-share agreement at a major Korean firm. Kakao’s union is already asking for more.

Samsung Electronics’ unionised workforce voted in favour of the government-mediated pay agreement on Wednesday, formally closing the deal that narrowly survived an injunction filing from a smaller non-chip union on Tuesday. The vote settles, in the immediate term, the largest labour dispute in the global semiconductor industry.

The wider effect is more durable: the agreement marks the first major win for a Samsung union in the company’s 56-year history, and it is being read across Korean industry as a structural shift in how labour bargaining works.

The substance of the deal is what makes it unusual. Samsung has agreed in writing to allocate 10.5% of its semiconductor operating profit to special bonuses for chip workers. It is, on Reuters’ count, only the second time a major South Korean company has put a fixed-percentage profit-share commitment into a binding labour agreement.

The 💜 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!

Some memory-division employees will receive total bonus packages worth around $416,000 over the agreement period. The non-chip Donghaeng union, which had filed the injunction at Suwon District Court, has signalled it intends to keep pressing for a revised allocation regardless of the vote.

The wider Korean labour picture has moved with it. Workers at Kakao and four of its affiliates have threatened to strike if their demands, including a 13–15% profit-share allocation, are not met. Other major Korean employers are reportedly fielding similar requests from their own unions.

The Samsung settlement has, in effect, created a precedent the rest of the chaebol system will now be benchmarked against.

Two structural conditions made the deal possible. The first is the AI-driven memory supercycle. Samsung’s memory division has been generating profits on a scale the company has rarely seen, and the gap between what the division produces and what its workers had been paid was visible to everyone involved.

The second is the loss of workforce to SK Hynix, where the AI-memory boom has been concentrated and where bonuses have been larger for years. According to the Samsung union, chip workers had begun leaving for SK Hynix in numbers that made the bonus gap commercially unsustainable.

The Korean chaebol bargaining model has historically been resistant to fixed-percentage profit-sharing on the grounds that it imports the cyclicality of the underlying business into the labour cost line. The Samsung deal accepts that trade-off: the bonus pool falls automatically when memory profits fall.

Workers, in turn, have agreed to the contingent floor on their payouts (memory must generate at least 200 trillion won between 2026 and 2028 and 100 trillion won between 2029 and 2035 for the full payout). It is a recognisably modern profit-share structure, of the kind Western technology firms have used for years, transplanted onto a chaebol balance sheet for the first time.

The political question is whether this is the start of a structural shift or a memory-cycle-specific anomaly. Korean economists have argued for years that the chaebol system’s relatively weak wage-growth performance during good years was a function of the labour-bargaining frame rather than of profitability.

The Samsung deal tests that argument empirically. If memory profits hold, the formula delivers genuinely large worker payouts and the new pattern spreads. If memory profits revert, the union side’s structural complaint, that a one-off cycle-linked bonus is not a sustainable wage policy, returns with it.

Samsung shares closed up modestly on Wednesday. The Korean labour ministry, which brokered the original agreement, said it expects similar mediated settlements at other major firms within months.