Caught between a public opinion that is adverse to the Trump administration and the economic and defense needs of maintaining Washington’s support as much as possible, numerous European Nato (and EU) countries, spearheaded by France and the UK, are collaborating with other nations to establish international maritime patrol and mine-clearing operations for the Strait of Hormuz once hostilities have ceased. This initiative aims, among other objectives, to somewhat appease Trump.
Bloomberg reports that Google will commit at least $10 billion to Anthropic, with the potential for the total to climb to $40 billion if Anthropic achieves certain performance milestones.
This comes after Amazon’s initial $5 billion investment in Anthropic announced a few days earlier; the Amazon arrangement similarly leaves room for additional funding tied to performance. Both deals place Anthropic’s valuation at $350 billion.
Anthropic has seen rapid adoption of its Claude models and related offerings like Claude Code, which promises to greatly accelerate and streamline how companies and individuals develop software. (Outcomes range from substantial improvements to setbacks, depending on the project, the organization, how Claude Code is applied, and many other factors.)
Several elements have driven Anthropic’s recent gains, including controversies around OpenAI and ChatGPT, advances in agentic workflows, and new products such as Claude Cowork, which handles general knowledge-work tasks in a way similar to how Claude Code targets software development.
This week the Federal Communications Commission made clear that its broad prohibition on consumer routers produced abroad also covers portable hotspot devices.
The FCC inserted a new section into an FAQ labeled, “Is my device a consumer-grade router under the National Security Determination?” That addition states the category encompasses “consumer-grade portable or mobile MiFi Wi-Fi or hotspot devices for residential use.” The FAQ also notes the ban excludes “mobile phones with hotspot features.”
Consequently, manufacturers of consumer hotspot units will need a government exemption to import and market any new hotspots that lack prior FCC approval. As with routers, devices that were already cleared for sale in the US may continue to be imported and sold without a separate exemption.
The FCC’s definition of routers is intentionally expansive, enabling the agency to cover a wide range of consumer networking hardware under the prohibition. When the FCC announced the ban last month, it described routers as “consumer-grade networking devices that are primarily intended for residential use and can be installed by the customer,” and as devices that “forward data packets, most commonly Internet Protocol (IP) packets, between networked systems.”
An earlier version of the FAQ had said that cellphones offering mobile hotspot functionality were exempt, but it didn’t explicitly state that standalone portable hotspot units would be covered. Beyond hotspot units, the updated FAQ specifies the router prohibition also covers “consumer or small and medium-sized business routers sold or rented through retail and self-installable by end users”; “LTE/5G CPE [customer premises equipment] devices for residential use”; “residential routers installed by a professional or ISP”; and “residential gateways that combine modem and router functions.”
Welcome to Open Tab, a weekly collection of updates, rumors, and narratives that have lingered open in my tabs throughout the week. Last week, we discussed the president’s DoorDash delivery.
If you caught last week’s edition of Open Tab, you’re in for a treat as we kick off this week’s edition by wrapping up some unfinished business. First up, the Philz Coffee Pride flag situation—which I assure you will not be mentioned again—is finally resolved. A message from Philz Coffee that landed in my inbox last week stated, among other things, “every Pride flag that is currently displayed will remain in place, and any Pride flag that was taken down can be raised again.” Complaining pays off!
The discussion on rotisserie chicken prices is still making waves—this time in governmental circles. A bipartisan proposal named the Hot Rotisserie Chicken Act is being introduced in the US Senate. The proposed law would permit SNAP beneficiaries to purchase hot rotisserie chicken—as of now, they can only buy cooked but cooled chickens.
Lastly, Nara Smith, the original traditional wife, has revealed that she will be launching a cookbook this October. Finally, everyone will discover what I, the offspring of a granola mom, have always known: The homemade variations of junk food just don’t taste as good. They really don’t—unless you happen to be Claire Saffitz, perhaps.
This week also brings: Prego aims to record your dinner discussions, and Jia Tolentino engaged in some casual shoplifting. Chaos followed. Moreover, Wagyu has lost some of its former glory, and lab-grown meat encounters a few obstacles.
Prego, known for its pasta sauce, is launching a device akin to Alexa called The Connection Keeper, designed to record your conversations at the dinner table. It purports to offer a screen-free method to “capture the laughter, stories, and moments that occur organically over a meal,” yet it feels unmistakably like a surveillance initiative. It is not Wi-Fi enabled and lacks cloud connectivity, but users are encouraged to upload their discussions to a StoryCorps website.
Personally, the idea of recording my private conversations and sharing them anywhere feels unsettling—websites can (and often do!) get hacked, leading to data breaches or sales. Wait, let me readjust my tinfoil hat.
New Yorker staff writer Jia Tolentino (and occasional BA contributor) appeared on the New York Times podcast The Opinions this week, and in an episode literally titled “The Rich Don’t Play by the Rules. So Why Should I?” she confessed to occasionally pocketing one or two items from Whole Foods. The mega-chain (owned by Amazon), as Tolentino claims, factors in this type of minor shoplifting anyway. The New York Post took issue, describing it as a “stunning revelation” that Tolentino had taken a few lemons.
Yes, theft is wrong. However, Tolentino is far from the only individual who has ever experimented with shoplifting at Whole Foods. Why else would they have established Whole Foods jail? Additionally, as Bon Appétit reported in 2023, some staff support your shoplifting. You’ve never tasted a grape in the produce section? You’ve never inadvertently scanned a banana without paying for it at self-checkout? Go ahead! Take a risk! Ed. note: For legal purposes, this is humorous.
If you’ve been dining out lately, it’s likely that Wagyu has appeared on your plate. It might have been presented as an expensive addition to an upscale tasting experience, or maybe you tried it when it was featured at Arby’s. It’s the protein of the moment, but as it has basked in the limelight, non-Japanese Wagyu has started to dominate the market.
“We aren’t relying on Europe, but they have a greater need for the Strait of Hormuz than we do, and may want to consider reducing their discussions and extravagant conferences in Europe and obtain a vessel. This is far more their battle than it is ours,” Hegseth stated.
DeepSeek claims the V4 model can stand shoulder to shoulder with top American systems from Google, OpenAI, and Anthropic.
DeepSeek claims the V4 model can stand shoulder to shoulder with top American systems from Google, OpenAI, and Anthropic.
The Chinese AI firm DeepSeek unveiled a preview of its highly awaited next-gen AI model V4 this past Friday, stating that the open-source model can rival leading closed-source systems from American competitors such as Anthropic, Google, and OpenAI.
DeepSeek asserts that V4 signifies a substantial upgrade over its previous models, particularly in coding, a feature that has become crucial for AI agents and has contributed to the popularity of tools like ChatGPT Codex and Claude Code. The launch is also a significant step for China’s chip industry, as DeepSeek specifically emphasizes compatibility with domestic Huawei technology.
This release follows a year after DeepSeek shook the US AI sector with R1, a model it claimed was trained at a fraction of the expense of top American systems. DeepSeek has not revealed the training costs for V4 or the hardware utilized. US officials have alleged that the company has utilized banned Nvidia chips and Anthropic asserts that DeepSeek misappropriated Claude to enhance its own offerings.
‘A photo captures illumination on a sensor or film. It serves as a record of a tangible moment in time.’
‘A photo captures illumination on a sensor or film. It serves as a record of a tangible moment in time.’
We enjoy pondering about how the concept of “real” photography is interpreted here at The Verge now that generative AI is widespread, and the World Press Photo contest might provide clarity. This renowned accolade honors the pinnacle of photojournalism, where depicting reality holds great importance.
The recipient of the 2026 award — “Separated by ICE,” taken by photojournalist Carol Guzy — was revealed yesterday. The distressing image portrays children holding onto their father following an immigration hearing. The image needed to conform to specific regulations concerning AI tool usage to qualify for the contest, with the independent nonprofit organization that oversees the award asserting that “AI-generated visuals do not count as photography.”
“A photo captures illumination on a sensor or film,” the organization articulated in its entry guidelines. “It serves as a record of a tangible moment in time.” The organization also specifies the precise tools and equipment allowed to qualify:
Every photograph submitted to the contest must be produced with a camera. No synthetic or artificially generated images will be accepted, and no use of generative fill is permitted in post-processing. Any use of these tools will automatically disqualify the submission from the contest.
Photos taken with smartphones are allowed only if captured in the standard shooting mode; entries produced using HDR, portrait mode, creative lighting effects, or panorama mode are ineligible.
The use of smart tools or AI-driven enhancement tools is allowed within the contest rules, as long as these tools do not result in significant alterations to the image overall, introduce new information to the image, or eliminate information from the image that was recorded by the camera.
Tools that immediately violate the contest regulations and are prohibited include all AI-powered enlarging tools like Adobe Super Resolution and Topaz Photo AI. These are based on generative AI models that add new information to enhance and enlarge images.
There remain some ambiguous areas, such as using AI tools for reducing noise, automatic color/contrast adjustments, and object selection, which are allowed as long as they’re not applied excessively and don’t “alter, add or remove details from the photograph that was captured by the camera,” per World Press Photo. The organization also provides a link to a page that outlines what constitutes photo manipulation according to its rules.
It features one of the most thorough sets of requirements we’ve encountered for outlining what qualifies as genuine photography, and could act as a standard for how we, the general public, contemplate this subject in the future.
I don’t have to remind you that AI is ubiquitous.
Nor that it is increasingly utilized in healthcare settings. Physicians are employing AI to assist them with documentation tasks. AI-driven applications are combing through patient files, identifying individuals who might need specific assistance or treatments. They are also employed to analyze lab results and X-rays.
A rising number of research efforts indicate that many of these applications can yield precise results. However, a more significant question arises: Does their utilization genuinely lead to improved health results for patients?
We still lack a definitive answer.
That’s the position taken by Jenna Wiens, a computer scientist at the University of Michigan, alongside Anna Goldenberg from the University of Toronto, in a paper published this week in Nature Medicine.
Wiens shares that she has dedicated years to examining how AI could enhance healthcare. In the first decade of her work, she endeavored to advocate for the technology among clinicians. Over the past few years, she notes, it seems that “a switch flipped.” Healthcare providers not only seem significantly more interested in the potential of these technologies, but they have also started swiftly integrating them.
The challenge is that numerous providers aren’t thoroughly evaluating how effectively they function.
Consider “ambient AI” systems, for instance. Also referred to as AI scribes, they “listen” in on discussions between healthcare professionals and patients, subsequently transcribing and summarizing them. Various tools are accessible, and they are already being embraced widely by healthcare practitioners.
A few months prior, an employee at a prominent medical center in New York, who develops AI solutions for physicians, informed me that, based on anecdotal evidence, healthcare providers are “thrilled” with the technology—it enables them to concentrate entirely on their patients during consultations, and it spares them from a lot of tedious paperwork. Initial studies back up these anecdotes and propose that the tools could alleviate clinician fatigue.
That’s certainly positive. But what about the health outcomes for patients? “[Researchers] have looked at provider or clinician and patient satisfaction, but not really at how these tools influence clinical decision-making,” remarks Wiens. “We lack that knowledge.”
The same applies to other AI-driven technologies utilized in healthcare environments. Some are employed to anticipate patients’ health pathways, while others suggest treatments. They are crafted to enhance the effectiveness and efficiency of care.
However, even a tool deemed “accurate” may not necessarily lead to improved health outcomes. For instance, AI could accelerate the reading of a chest X-ray. But to what extent will a physician trust its evaluation? How will that tool influence the manner in which a physician engages with patients or suggests treatment? And ultimately: What implications will this have for the patients involved?
The responses to these queries may vary between healthcare facilities or departments and could hinge on clinical workflows, according to Wiens. They may also change among doctors at different points in their careers.
Take the AI scribes again, for example. Some research on AI in educational contexts indicates that such tools might alter how individuals cognitively process information. Could they influence how a physician interprets patient data? Will the tools modify the perceptions of medical students regarding patient information in a way that impacts care? These inquiries require investigation, states Wiens. “We appreciate tools that conserve our time, but we must consider the unanticipated effects of this,” she suggests.
In a study released in January 2025, Paige Nong at the University of Minnesota and her team discovered that approximately 65% of hospitals across the US utilized AI-assisted predictive tools. Only two-thirds of those institutions evaluated their precision. An even smaller proportion assessed them for potential bias.
The percentage of hospitals implementing these tools has likely surged since then, notes Wiens. These hospitals, or organizations other than those developing the tools, must assess how effectively they perform in specific environments. There’s a chance that they could leave patients in a worse situation, although it’s more plausible that AI tools simply aren’t as advantageous as healthcare providers may believe, according to Wiens.
“I genuinely believe in the ability of AI to significantly enhance clinical care,” states Wiens, emphasizing that she does not aim to halt the integration of AI tools in healthcare. She simply desires more insight regarding their impact on patients. “I have to be optimistic that in the future it’s not all AI or none AI,” she expresses. “It resides somewhere in between.”
This article initially appeared in The Checkup, MIT Technology Review’s weekly biotech newsletter. To receive it directly in your inbox every Thursday, and to read articles like this first, sign up here.
Money, a member of the Wiradjuri community, has been recognized for her exceptional poetry, including the esteemed 2025 Kate Challis RAKA Award, which honors Indigenous artists. Additionally, she has been awarded the First Nations Emerging Career Award by the Australia Council for the Arts.
Shares of Taiwan Semiconductor Manufacturing Co. jumped 5% to a new all-time peak on Friday following the island’s regulatory announcement to ease restrictions on fund allocations to individual stocks.
According to the new guidelines, domestic equity funds and actively managed ETFs that focus solely on Taiwanese equities will be permitted to direct up to 25% of their assets to any listed company with a weight exceeding 10% on the Taiwan Stock Exchange.
A longstanding regulation had limited fund managers’ investments in a single enterprise to 10% of their portfolio’s net asset value.
TSMC, whose stock also reached a record high on Thursday, last week disclosed a 58% rise in first-quarter earnings, surpassing forecasts as the surge in artificial intelligence heightens demand for chips.
TSMC’s net profit of 572.48 billion new Taiwanese dollars for the quarter ending in March marked a fourth consecutive quarter of record earnings.
The firm stands as Asia’s most valuable tech company, producing semiconductors utilized in various devices ranging from consumer electronics to extensive data centers.
The globe’s largest contract chip manufacturer continues to experience strong demand for advanced chips from leading clients like Apple, while also gaining from the swift growth of AI, manufacturing cutting-edge processors designed by companies such as Nvidia — now its top customer.
