"Ubuntu's decision to switch to Rust-based coreutils in 25.10 hasn't been the smoothest ride," writes the blog OMG Ubuntu, "as the latest — albeit now resolved — bug underscores." [Coreutils] are used by a number of processes, apps and scripts, including Ubuntu's own unattended-upgrades process, which automatically checks for new software updates. Alas, the Rust-based version of date had a bug which meant Ubuntu 25.10 desktops, servers, cloud and container images were not able to automatically check for updates when configured. Unattended-upgrades hooks into the date utility to check the timestamp of a reference file of when an update check was last run and, past a certain date, checks again. But date was incorrectly showing the current date, always.

A fix has been issued so only Ubuntu 25.10 installs withrust-coreutils 0.2.2-0ubuntu2 (or earlier) are affected.

Slashdot reader BrianFagioli writes: Password manager 1Password's 2025 Annual Report: The Access-Trust Gap exposes how everyday employees are becoming accidental hackers in the AI era. The company's data shows that 73% of workers are encouraged to use AI tools, yet more than a third admit they do not always follow corporate policies. Many employees are feeding sensitive information into large language models or using unapproved AI apps to get work done, creating what 1Password calls "Shadow AI." At the same time, traditional defenses like single sign-on (SSO) and mobile device management (MDM) are failing to keep pace, leaving gaps in visibility and control.

The report warns that corporate security is being undermined from within. More than half of employees have installed software without IT approval, two-thirds still use weak passwords, and 38% have accessed accounts at previous employers. Despite rising enthusiasm for passkeys and passwordless authentication, 1Password says most organizations still depend on outdated systems that were never built for cloud-native, AI-driven work. The result is a growing "Access-Trust Gap" that could allow AI chaos and employee shortcuts to dismantle enterprise security from the inside.

The address bar/ChatGPT input window in OpenAI's browser ChatGPT Atlas "could be targeted for prompt injection using malicious instructions disguised as links," reports SC World, citing a report from AI/agent security platform NeuralTrust: NeuralTrust found that a malformed URL could be crafted to include a prompt that is treated as plain text by the browser, passing the prompt on to the LLM. A malformation, such as an extra space after the first slash following "https:" prevents the browser from recognizing the link as a website to visit. Rather than triggering a web search, as is common when plain text is submitted to a browser's address bar, ChatGPT Atlas treats plain text as ChatGPT prompts by default.

An unsuspecting user could potentially be tricked into copying and pasting a malformed link, believing they will be sent to a legitimate webpage. An attacker could plant the link behind a "copy link" button so that the user might not notice the suspicious text at the end of the link until after it is pasted and submitted. These prompt injections could potentially be used to instruct ChatGPT to open a new tab to a malicious website such as a phishing site, or to tell ChatGPT to take harmful actions in the user's integrated applications or logged-in sites like Google Drive, NeuralTrust said.
Last month browser security platform LayerX also described how malicious prompts could be hidden in URLs (as a parameter) for Perplexity's browser Comet. And last week SquareX Labs demonstrated that a malicious browser extension could spoof Comet's AI sidebar feature and have since replicated the proof-of-concept (PoC) attack on Atlas.

But another new vulnerability in ChatGPT Atlas "could allow malicious actors to inject nefarious instructions into the artificial intelligence (AI)-powered assistant's memory and run arbitrary code," reports The Hacker News, citing a report from browser security platform LayerX: "This exploit can allow attackers to infect systems with malicious code, grant themselves access privileges, or deploy malware," LayerX Security Co-Founder and CEO, Or Eshed, said in a report shared with The Hacker News. The attack, at its core, leverages a cross-site request forgery (CSRF) flaw that could be exploited to inject malicious instructions into ChatGPT's persistent memory. The corrupted memory can then persist across devices and sessions, permitting an attacker to conduct various actions, including seizing control of a user's account, browser, or connected systems, when a logged-in user attempts to use ChatGPT for legitimate purposes....

"What makes this exploit uniquely dangerous is that it targets the AI's persistent memory, not just the browser session," Michelle Levy, head of security research at LayerX Security, said. "By chaining a standard CSRF to a memory write, an attacker can invisibly plant instructions that survive across devices, sessions, and even different browsers. In our tests, once ChatGPT's memory was tainted, subsequent 'normal' prompts could trigger code fetches, privilege escalations, or data exfiltration without tripping meaningful safeguards...."

LayerX said the problem is exacerbated by ChatGPT Atlas' lack of robust anti-phishing controls, the browser security company said, adding it leaves users up to 90% more exposed than traditional browsers like Google Chrome or Microsoft Edge. In tests against over 100 in-the-wild web vulnerabilities and phishing attacks, Edge managed to stop 53% of them, followed by Google Chrome at 47% and Dia at 46%. In contrast, Perplexity's Comet and ChatGPT Atlas stopped only 7% and 5.8% of malicious web pages.
From The Conversation: Sandboxing is a security approach designed to keep websites isolated and prevent malicious code from accessing data from other tabs. The modern web depends on this separation. But in Atlas, the AI agent isn't malicious code — it's a trusted user with permission to see and act across all sites. This undermines the core principle of browser isolation.
Thanks to Slashdot reader spatwei for suggesting the topic.

OpenAI has introduced Aardvark, a GPT-5-powered autonomous agent that scans, reasons about, and patches code like a human security researcher. "By embedding itself directly into the development pipeline, Aardvark aims to turn security from a post-development concern into a continuous safeguard that evolves with the software itself," reports InfoWorld. From the report: What makes Aardvark unique, OpenAI noted, is its combination of reasoning, automation, and verification. Rather than simply highlighting potential vulnerabilities, the agent promises multi-stage analysis -- starting by mapping an entire repository and building a contextual threat model around it. From there, it continuously monitors new commits, checking whether each change introduces risk or violates existing security patterns.

Additionally, upon identifying a potential issue, Aardvark attempts to validate the exploitability of the finding in a sandboxed environment before flagging it. This validation step could prove transformative. Traditional static analysis tools often overwhelm developers with false alarms -- issues that may look risky but aren't truly exploitable. "The biggest advantage is that it will reduce false positives significantly," noted Jain. "It's helpful in open source codes and as part of the development pipeline."

Once a vulnerability is confirmed, Aardvark integrates with Codex to propose a patch, then re-analyzes the fix to ensure it doesn't introduce new problems. OpenAI claims that in benchmark tests, the system identified 92 percent of known and synthetically introduced vulnerabilities across test repositories, a promising indication that AI may soon shoulder part of the burden of modern code auditing.

The FCC plans to repeal a Biden-era ruling that required ISPs to secure their networks under the Communications Assistance for Law Enforcement Act, instead relying on voluntary cybersecurity commitments from telecom providers. FCC Chairman Brendan Carr said the ruling "exceeded the agency's authority and did not present an effective or agile response to the relevant cybersecurity threats." Carr said the vote scheduled for November 20 comes after "extensive FCC engagement with carriers" who have taken "substantial steps... to strengthen their cybersecurity defenses." Ars Technica reports: The FCC's January 2025 declaratory ruling came in response to attacks by China, including the Salt Typhoon infiltration of major telecom providers such as Verizon and AT&T. The Biden-era FCC found that the Communications Assistance for Law Enforcement Act (CALEA), a 1994 law, "affirmatively requires telecommunications carriers to secure their networks from unlawful access or interception of communications."

"The Commission has previously found that section 105 of CALEA creates an affirmative obligation for a telecommunications carrier to avoid the risk that suppliers of untrusted equipment will "illegally activate interceptions or other forms of surveillance within the carrier's switching premises without its knowledge,'" the January order said. "With this Declaratory Ruling, we clarify that telecommunications carriers' duties under section 105 of CALEA extend not only to the equipment they choose to use in their networks, but also to how they manage their networks." A draft of the order that will be voted on in November can be found here (PDF).

Microsoft is bringing shared audio to Windows 11, allowing you to stream audio across two pairs of wireless headphones, speakers, earbuds, or hearing aids. From a report: The feature is built using the Bluetooth Low Energy (LE) audio codec, and it's rolling out in preview to Windows 11 Insiders in the Dev and Beta channels. Shared audio comes in handy if you're watching a movie on a laptop with your friend or family member, or just want to show them new music that you can both stream inside your own wireless headsets. You can use shared audio by connecting Bluetooth LE-supported devices to your Windows 11 PC and then selecting the Shared audio (preview) button in your quick settings menu. Microsoft introduced an LE Audio feature on Windows 11 in August, enabling higher audio quality while using a wireless headset in a game or call.

Trevor McNally posts videos of himself opening locks. The former Marine has 7 million followers and nearly 10 million people watched him open a Proven Industries trailer hitch lock in April using a shim cut from an aluminum can. The Florida company responded by filing a federal lawsuit in May charging McNally with eight offenses. Judge Mary Scriven denied the preliminary injunction request in June and found the video was fair use.

McNally's followers then flooded the company with harassment. Proven dismissed the case in July and asked the court to seal the records. The company had initiated litigation over a video that all parties acknowledged was accurate. ArsTechnica adds: Judging from the number of times the lawsuit talks about 1) ridicule and 2) harassment, it seems like the case quickly became a personal one for Proven's owner and employees, who felt either mocked or threatened. That's understandable, but being mocked is not illegal and should never have led to a lawsuit or a copyright claim. As for online harassment, it remains a serious and unresolved issue, but launching a personal vendetta -- and on pretty flimsy legal grounds -- against McNally himself was patently unwise. (Doubly so given that McNally had a huge following and had already responded to DMCA takedowns by creating further videos on the subject; this wasn't someone who would simply be intimidated by a lawsuit.)

In the end, Proven's lawsuit likely cost the company serious time and cash -- and generated little but bad publicity.

An anonymous reader quotes a report from 404 Media: Someone recently managed to get on a Microsoft Teams call with representatives from phone hacking company Cellebrite, and then leaked a screenshot of the company's capabilities against many Google Pixel phones, according to a forum post about the leak and 404 Media's review of the material. The leak follows others obtained and verified by 404 Media over the last 18 months. Those leaks impacted both Cellebrite and its competitor Grayshift, now owned by Magnet Forensics. Both companies constantly hunt for techniques to unlock phones law enforcement have physical access to.

"You can Teams meeting with them. They tell everything. Still cannot extract esim on Pixel. Ask anything," a user called rogueFed wrote on the GrapheneOS forum on Wednesday, speaking about what they learned about Cellebrite capabilities. GrapheneOS is a security- and privacy-focused Android-based operating system. rogueFed then posted two screenshots of the Microsoft Teams call. The first was a Cellebrite Support Matrix, which lays out whether the company's tech can, or can't, unlock certain phones and under what conditions. The second screenshot was of a Cellebrite employee. According to another of rogueFed's posts, the meeting took place in October. The meeting appears to have been a sales call. The employee is a "pre sales expert," according to a profile available online.

The Support Matrix is focused on modern Google Pixel devices, including the Pixel 9 series. The screenshot does not include details on the Pixel 10, which is Google's latest device. It discusses Cellebrite's capabilities regarding 'before first unlock', or BFU, when a piece of phone unlocking tech tries to open a device before someone has typed in the phone's passcode for the first time since being turned on. It also shows Cellebrite's capabilities against after first unlock, or AFU, devices. The Support Matrix also shows Cellebrite's capabilities against Pixel devices running GrapheneOS, with some differences between phones running that operating system and stock Android. Cellebrite does support, for example, Pixel 9 devices BFU. Meanwhile the screenshot indicates Cellebrite cannot unlock Pixel 9 devices running GrapheneOS BFU. In their forum post, rogueFed wrote that the "meeting focused specific on GrapheneOS bypass capability." They added "very fresh info more coming."

A critical security flaw in Chromium's Blink rendering engine can crash billions of browsers within seconds. Security researcher Jose Pino discovered the vulnerability and created a proof-of-concept exploit called Brash to demonstrate the bug affecting Chrome, Edge, OpenAI's ChatGPT Atlas, Brave, Vivaldi, Arc, Dia, Opera and Perplexity Comet.

The flaw, reports The Register, exploits the absence of rate limiting on document.title API updates in Chromium versions 143.0.7483.0 and later. The attack injects millions of DOM mutations per second and saturates the main thread. When The Register tested the code on Edge, the browser crashed and the Windows machine locked up after about 30 seconds while consuming 18GB of RAM in one tab. Pino disclosed the bug to the Chromium security team on August 28 and followed up on August 30 but received no response. Google said it is looking into the issue.

More than a half dozen federal departments and agencies have backed a proposal to ban future sales of the most popular home routers in the United States on the grounds that the vendor's ties to mainland China make them a national security risk, Washington Post reported Thursday, citing people briefed on the matter. From the report: The proposal, which arose from a months-long risk assessment, calls for blocking sales of networking devices from TP-Link Systems of Irvine, California, which was spun off from a China-based company, TP-Link Technologies, but owns some of that company's former assets in China.

The ban was proposed by the Commerce Department and supported this summer by an interagency process that includes the Departments of Homeland Security, Justice and Defense, the people said. "TP-Link vigorously disputes any allegation that its products present national security risks to the United States," Ricca Silverio, a spokeswoman for TP-Link Systems, said in a statement. "TP-Link is a U.S. company committed to supplying high-quality and secure products to the U.S. market and beyond."

If imposed, the ban would be among the largest in consumer history and a possible sign that the East-West divide over tech independence is still deepening amid reports of accelerated Chinese government-supported hacking. Only the legislated ban of Chinese-owned TikTok, which President Donald Trump has averted with executive orders and a pending sale, would impact more U.S. consumers.

An anonymous reader shares a report: The transition to the more-secure HTTPS web protocol has plateaued, according to Google. As of 2020, 95 to 99 percent of navigations in Chrome use HTTPS. To help make it safer for users to click on links, Chrome will enable a setting called Always Use Secure Connections for public sites for all users by default. This will happen in October 2026 with the release of Chrome 154.

The change will happen earlier for those who have switched on Enhanced Safe Browsing protections in Chrome. Google will enable Always Use Secure Connections by default in April when Chrome 147 drops. When this setting is on, Chrome will ask for your permission before it first accesses a public website that doesn't use HTTPS.

Blogger and technologist Anil Dash, writing about OpenAI's recently launched browser, Atlas: When I first got Atlas up and running, I tried giving it the easiest and most obvious tasks I could possibly give it. I looked up "Taylor Swift showgirl" to see if it would give me links to videos or playlists to watch or listen to the most popular music on the charts right now; this has to be just about the easiest possible prompt.

The results that came back looked like a web page, but they weren't. Instead, what I got was something closer to a last-minute book report written by a kid who had mostly plagiarized Wikipedia. The response mentioned some basic biographical information and had a few photos. Now we know that AI tools are prone to this kind of confabulation, but this is new, because it felt like I was in a web browser, typing into a search box on the Internet. And here's what was most notable: there was no link to her website.

I had typed "Taylor Swift" in a browser, and the response had literally zero links to Taylor Swift's actual website. If you stayed within what Atlas generated, you would have no way of knowing that Taylor Swift has a website at all.

Unless you were an expert, you would almost certainly think I had typed in a search box and gotten back a web page with search results. But in reality, I had typed in a prompt box and gotten back a synthesized response that superficially resembles a web page, and it uses some web technologies to display its output. Instead of a list of links to websites that had information about the topic, it had bullet points describing things it thought I should know. There were a few footnotes buried within some of those response, but the clear intent was that I was meant to stay within the AI-generated results, trapped in that walled garden.

During its first run, there's a brief warning buried amidst all the other messages that says, "ChatGPT may give you inaccurate information", but nobody is going to think that means "sometimes this tool completely fabricates content, gives me a box that looks like a search box, and shows me the fabricated content in a display that looks like a web page when I type in the fake search box."

And it's not like the generated response is even that satisfying.

darwinmac writes: Ubuntu Unity is staring at a possible shutdown. A community moderator has gone public pleading for help, admitting the project is "broken and needs to be fixed." Neowin reports the distro is suffering from critical bugs so severe that upgrades from 25.04 to 25.10 are failing and even fresh installs are hit. The moderator admits they lack the technical skill or time to perform a full rescue and is asking the broader community, including devs, testers, and UI designers, to step in so Ubuntu Unity can reach 26.04 LTS. If no one steps in soon, this community flavor might quietly fade away once more.

An anonymous reader shares a report: After last week's major AWS outage took Signal along with it, Elon Musk was quick to criticize the encrypted messaging app's reliance on big tech. But Signal president Meredith Whittaker argues that the company didn't have any other choice but to use AWS or another major cloud provider.

"The problem here is not that Signal 'chose' to run on AWS," Whittaker writes in a series of posts on Bluesky. "The problem is the concentration of power in the infrastructure space that means there isn't really another choice: the entire stack, practically speaking, is owned by 3-4 players."

In the thread, Whittaker says the number of people who didn't realize Signal uses AWS is "concerning," as it indicates they aren't aware of just how concentrated the cloud infrastructure industry is. "The question isn't 'why does Signal use AWS?'" Whittaker writes. "It's to look at the infrastructural requirements of any global, real-time, mass comms platform and ask how it is that we got to a place where there's no realistic alternative to AWS and the other hyperscalers."

An anonymous reader quotes a report from BleepingComputer: The number of victims paying ransomware threat actors has reached a new low, with just 23% of the breached companies giving in to attackers' demands. With some exceptions, the decline in payment resolution rates continues the trend that Coveware has observed for the past six years. In the first quarter of 2024, the payment percentage was 28%. Although it increased over the next period, it continued to drop, reaching an all-time low in the third quarter of 2025.

One explanation for this is that organizations implemented stronger and more targeted protections against ransomware, and authorities increasing pressure for victims not to pay the hackers. [...] Over the years, ransomware groups moved from pure encryption attacks to double extortion that came with data theft and the threat of a public leak. Coveware reports that more than 76% of the attacks it observed in Q3 2025 involved data exfiltration, which is now the primary objective for most ransomware groups. The company says that when it isolates the attacks that do not encrypt the data and only steal it, the payment rate plummets to 19%, which is also a record for that sub-category.

The average and median ransomware payments fell in Q3 compared to the previous quarter, reaching $377,000 and $140,000, respectively, according to Coveware. The shift may reflect large enterprises revising their ransom payment policies and recognizing that those funds are better spent on strengthening defenses against future attacks. The researchers also note that threat groups like Akira and Qilin, which accounted for 44% of all recorded attacks in Q3 2025, have switched focus to medium-sized firms that are currently more likely to pay a ransom. "Cyber defenders, law enforcement, and legal specialists should view this as validation of collective progress," Coveware says. "The work that gets put in to prevent attacks, minimize the impact of attacks, and successfully navigate a cyber extortion -- each avoided payment constricts cyber attackers of oxygen."

Countries signed their first UN treaty targeting cybercrime in Hanoi on Saturday, despite opposition from an unlikely band of tech companies and rights groups warning of expanded state surveillance. From a report: The new global legal framework aims to strengthen international cooperation to fight digital crimes, from child pornography to transnational cyberscams and money laundering. More than 60 countries were seen to sign the declaration Saturday, which means it will go into force once ratified by those states. UN Secretary General Antonio Guterres described the signing as an "important milestone", but that it was "only the beginning".

"Every day, sophisticated scams, destroy families, steal migrants and drain billions of dollars from our economy... We need a strong, connected global response," he said at the opening ceremony in Vietnam's capital on Saturday. The UN Convention against Cybercrime was first proposed by Russian diplomats in 2017, and approved by consensus last year after lengthy negotiations. Critics say its broad language could lead to abuses of power and enable the cross-border repression of government critics.

Slashdot reader joshuark writes: Microsoft says that the File Explorer (formerly Windows Explorer) now automatically blocks previews for files downloaded from the Internet to block credential theft attacks via malicious documents, according to a report from BleepingComputer. This attack vector is particularly concerning because it requires no user interaction beyond selecting a file to preview and removes the need to trick a target into actually opening or executing it on their system.

For most users, no action is required since the protection is enabled automatically with the October 2025 security update, and existing workflows remain unaffected unless you regularly preview downloaded files.

"This change is designed to enhance security by preventing a vulnerability that could leak NTLM hashes when users preview potentially unsafe files," Microsoft says in a support document published Wednesday.

It is important to note that this may not take effect immediately and could require signing out and signing back in.

Has AI just become an easy excuse for firms looking to downsize, asks CNBC: Fabian Stephany, assistant professor of AI and work at the Oxford Internet Institute, said there might be more to job cuts than meets the eye. Previously there may have been some stigma attached to using AI, but now companies are "scapegoating" the technology to take the fall for challenging business moves such as layoffs. "I'm really skeptical whether the layoffs that we see currently are really due to true efficiency gains. It's rather really a projection into AI in the sense of 'We can use AI to make good excuses,'" Stephany said in an interview with CNBC. Companies can essentially position themselves at the frontier of AI technology to appear innovative and competitive, and simultaneously conceal the real reasons for layoffs, according to Stephany... Some companies that flourished during the pandemic "significantly overhired" and the recent layoffs might just be a "market clearance...."

One founder, Jean-Christophe Bouglé even said in a popular LinkedIn post that AI adoption is at a "much slower pace" than is being claimed and in large corporations "there's not much happening" with AI projects even being rolled back due to cost or security concerns. "At the same time there are announcements of big layoff plans 'because of AI.' It looks like a big excuse, in a context where the economy in many countries is slowing down..."

The Budget Lab, a non-partisan policy research center at Yale University, released a report on Wednesday which showed that U.S. labor has actually been little disrupted by AI automation since the release of ChatGPT in 2022... Additionally, New York Fed economists released research in early September which showed that AI use amongst firms "do not point to significant reductions in employment" across the services and manufacturing industry in the New York-Northern New Jersey region.

Critics question why basic flaws like buffer overflows, command injections, and SQL injections are "being exploited remain prevalent in mission-critical codebases maintained by companies whose core business is cybersecurity," writes CSO Online. Benjamin Harris, CEO of cybersecurity/penetration testing firm watchTowr tells them that "these are vulnerability classes from the 1990s, and security controls to prevent or identify them have existed for a long time. There is really no excuse." Enterprises have long relied on firewalls, routers, VPN servers, and email gateways to protect their networks from attacks. Increasingly, however, these network edge devices are becoming security liabilities themselves... Google's Threat Intelligence Group tracked 75 exploited zero-day vulnerabilities in 2024. Nearly one in three targeted network and security appliances, a strikingly high rate given the range of IT systems attackers could choose to exploit. That trend has continued this year, with similar numbers in the first 10 months of 2025, targeting vendors such as Citrix NetScaler, Ivanti, Fortinet, Palo Alto Networks, Cisco, SonicWall, and Juniper. Network edge devices are attractive targets because they are remotely accessible, fall outside endpoint protection monitoring, contain privileged credentials for lateral movement, and are not integrated into centralized logging solutions...

[R]esearchers have reported vulnerabilities in these systems for over a decade with little attacker interest beyond isolated incidents. That shifted over the past few years with a rapid surge in attacks, making compromised network edge devices one of the top initial access vectors into enterprise networks for state-affiliated cyberespionage groups and ransomware gangs. The COVID-19 pandemic contributed to this shift, as organizations rapidly expanded remote access capabilities by deploying more VPN gateways, firewalls, and secure web and email gateways to accommodate work-from-home mandates. The declining success rate of phishing is another factor... "It is now easier to find a 1990s-tier vulnerability in a border device where Endpoint Detection and Response typically isn't deployed, exploit that, and then pivot from there" [says watchTowr CEL Harris]...

Harris of watchTowr doesn't want to minimize the engineering effort it takes to build a secure system. But he feels many of the vulnerabilities discovered in the past two years should have been caught with automatic code analysis tools or code reviews, given how basic they have been. Some VPN flaws were "trivial to the point of embarrassing for the vendor," he says, while even the complex ones should have been caught by any organization seriously investing in product security... Another problem? These appliances have a lot of legacy code, some that is 10 years or older.
Attackers may need to chain together multiple hard-to-find vulnerabilities across multiple components, the article acknowleges. And "It's also possible that attack campaigns against network-edge devices are becoming more visible to security teams because they are looking into what's happening on these appliances more than they did in the past... "

The article ends with reactions from several vendors of network edge security devices.

Thanks to Slashdot reader snydeq for sharing the article.

The Washington Post reports on 996, "a term popularized in China that refers to a rigid work schedule in which people work from 9 a.m. to 9 p.m., six days a week..." As the artificial intelligence race heats up, many start-ups in Silicon Valley and New York are promoting hardcore culture as a way of life, pushing the limits of work hours, demanding that workers move fast to be first in the market. Some are even promoting 996 as a virtue in the hiring process and keeping "grind scores" of companies... Whoever builds first in AI will capture the market, and the window of opportunity is two to three years, "so you better run faster than everyone else," said Inaki Berenguer, managing partner of venture-capital firm LifeX Ventures.

At San Francisco-based AI start-up Sonatic, the grind culture also allows for meal, gym and pickleball time, said Kinjal Nandy, its CEO. Nandy recently posted a job opening on X that requires in-person work seven days a week. He said working 10-hour days sounds like a lot but the company also offers its first hires perks such as free housing in a hacker house, food delivery credits and a free subscription to the dating service Raya... Mercor, a San Francisco-based start-up that uses AI to match people to jobs, recently posted an opening for a customer success engineer, saying that candidates should have a willingness to work six days a week, and it's not negotiable. "We know this isn't for everyone, so we want to put it up top," the listing reads.

Being in-person rather than remote is a requirement at some start-ups. AI start-up StarSling had two engineering job descriptions that required six days a week of in-person work. In a job description for an engineer, Rilla, an AI company in New York, said candidates should not work at the company if they're not excited about working about 70 hours a week in person. One venture capitalist even started tracking "grind scores." Jared Sleeper, a partner at New York-based venture capital firm Avenir, recently ranked public software companies' "grind score" in a post on X, which went viral. Using data from Glassdoor, it ranks the percentage of employees who have a positive outlook for the company compared with their views on work-life balance.
"At Google's AI division, cofounder Sergey Brin views 60 hours per week as the 'sweet spot' for productivity," notes the Independent: Working more than 55 hours a week, compared with a standard 35-40-hour week, is linked to a 35 percent higher risk of stroke and a 17 percent higher risk of death from heart disease, according to the World Health Organization. Productivity also suffers. A British study shows that working beyond 60 hours a week can reduce overall output, slow cognitive performance, and impair tasks ranging from call handling to problem-solving.

Shorter workweeks, in contrast, appear to boost productivity. Microsoft Japan saw a roughly 40% increase in output after adopting a four-day work week. In a UK trial, 61 companies that tested a four-day schedule reported revenue gains, with 92 percent choosing to keep the policy, according to Bloomberg.