The briefing for October 22, 2026, evening, covers sixty articles published since nine in the morning UTC.
The largest theme in the window is the cybersecurity crisis centered on two critical PHP vulnerabilities, discovered with the help of AI, and the broader question of what AI is doing to offensive and defensive security. Several articles belong to this theme, and they are treated together first.
New Serious PHP Vulnerabilities Released and Patched, from Schneier on Security, reports that PHP maintainers have released fixes for two critical vulnerabilities affecting PHP 8.5.12, 8.4.27, 8.3.36 and 8.2.35. The flaws can allow unauthenticated remote code execution. The first, identified as CVE-2026-109618, involves remote code execution via out-of-memory corruption: out-of-memory handling can interrupt PHP while internal request data is only partially initialized, and request cleanup may then treat invalid data as valid. The second, CVE-2026-109619, involves remote code execution via stale global state: request-owned internal state can survive after its backing memory has been freed, and a later request handled by the same worker may use that stale state before normal initialization resets it. The risk applies to all PHP installations serving web requests, including PHP-FPM, Apache with mod_php, php-cgi, FrankenPHP and LiteSpeed, with the highest risk for mod_php, where a demonstrated exploit can trigger the attack before any application code runs. The article states that updating PHP is the only way to regain security.
AI Is Changing Everything About Hacking. No One Knows What Happens Next, from 404 Media, places this moment in a longer arc. The article reports that for much of the last month, cybersecurity experts have been posting with alarm: Johns Hopkins professor Matthew Green wrote that we are in the part of the movie where you have to throw whatever you can in the path of the zombies, and that so many people are discovering the whole internet is dry tinder. Those posts came before the PHP flaws were found. The article says the recent discoveries in eMudhra, Node.js, GitLab, Plex, Railway, OpenCode and other services have gotten less attention than PHP2Shell, which researchers from Wiz, Calif and OpenAI discovered using AI. Cloudflare CEO Matthew Prince posted that in the ten hours before public announcement the network saw thousands of exploit attempts, that jumped to tens of thousands in forty-five minutes after announcement, hundreds of thousands within two hours, and millions now. The article argues that AI has fundamentally changed both offensive hacking and defensive cybersecurity, not because AI is sentient, but because much of the software running the internet is vulnerable, out of date, or installed by someone who left years ago. If every piece of software is a door, AI makes it exceptionally easy to turn every knob to see if it is unlocked. The article also discusses abliterated AI models designed to not say no, and open weights models usable for offensive work. OpenAI argued it did its research ethically and opened a defender's window, but the article notes that AI is shortening the window for disclosure and making bug finding and fixing much faster, while whether fixes are actually deployed by companies is a much more complicated question. The security company Chainguard was criticized for posting fixes before official disclosure. The article raises the question of whether the AI models that found the vulnerabilities incorporated that information, making it easier for others to find, or whether there was a leak. Some infosec experts call this Black October, and the NSA, the Prime Minister of Japan and various governments told people to patch immediately. The article says one of the biggest impacts could be that many historic websites cannot or will not be updated, may be pulled offline, and may never come back. It reports that a working exploit was built by Will Tan in one day with Codex, drawing on roughly ten years of system security research, and that AI accelerates exploit development while turning a bug into a shell still takes expertise. Elias Taylor's blog post titled There isn't an easy exploit for this PHP vulnerability went viral, and Taylor told 404 Media that lots of people are being tricked by AI into thinking they have an exploit when they don't, while still urging servers to be patched. The article closes by saying it remains unclear what the ultimate impact will be in one hour, one week, one year, or ten years, but AI is changing how cybersecurity and hacking works.
AI #191: The Hackening Begins, from Don't Worry About the Vase, is a wide-ranging digest that opens with the claim that it was probably the worst day in cybersecurity history so far. The newsletter notes the critical PHP flaw was announced a day before the fix, sending most of the internet scrambling and likely dooming a lot of long tail websites permanently, and that several unrelated incidents also occurred. It says the floodgates seem to have been opened. The newsletter then moves across many AI topics. It reports on a leaked draft bill Anthropic had been shopping around Congress, weaker than the FRONTIER Act, and says Anthropic's initial attempts to explain itself only made the situation worse. It discusses a new claim that individual neurons mostly have a single interpretable meaning, and that it is tractable to label them one by one, treating this as excellent news that lowers the estimate of difficulty. It reports on Claude Duet, DeepSeek V4.1 Flash, Muse Wonder, and the Muse Spark Open model. On agents, it notes that UPS, Delta, JP Morgan Chase, Charles Schwab and Amazon block AI agents, and that a practical barrier for agents is overrefusals and misclassifications requiring permissions. The newsletter covers deepfakes, a case of a Brazilian man traveling twelve hundred miles for a woman who turned out to be AI-generated, and a fabricated HOA letter. It discusses an alleged OpenAI spy who used his wife's ChatGPT to plan planting a bug, with signs the client was the Saudi government, and notes OpenAI is giving employees anti-honeypot training. It reports on employment at Anthropic, where AI agent hours outnumber employee hours six to one and technical hiring has fallen as a share of employment. It covers DeepSeek raising fifteen billion dollars, Microsoft canceling a UK data center deal with Nscale, and Anthropic being in talks for a forty-five billion dollar cloud deal with Nscale. It reports that OpenAI has pulled back from supporting AI chip controls, while Anthropic supports several chip bills, and that the Institute for Progress estimates America is currently six months ahead of China but could double that lead by enforcing existing laws and using existing authority. The newsletter also covers broad worries about AI killing everyone, interpretability, cooperative alignment, and a Bloomberg profile asking why Marc Andreessen is so angry. It ends with a lighter section including a joke about the alignment plan.
Quoting Armin Ronacher, from Simon Willison's Weblog, is a single quote from Armin Ronacher: as fun as AI is, we built so much bad software over the last thirty years, and now the universe is punishing us for our sins.
Several Linux Kernel Vulnerabilities Accidentally Posted To Public Kernel Mailing List, from Slashdot, reports that Trend Micro's Zero Day Initiative appears to have accidentally disclosed several unpatched Linux kernel vulnerabilities on the public LKML instead of the private security list. Two of the flaws carry CVSS 9.0 scores and could enable remote code execution through SMB Direct and KSMBD. The article lists additional vulnerabilities, including a local privilege escalation in skbuff skb_zerocopy in network code, an out-of-bounds write in the VMWGFX VMware graphics driver, an uninitialized resource in NVMET code leading to information disclosure, and a stack buffer overflow in a search-domain TCP query leading to local privilege escalation. Greg Kroah-Hartman reportedly responded to each report asking about their submission to the public mailing list.
Zoe Kleinman: PHP Is the Cyber Equivalent of an Unexploded Bomb, from Slashdot, quotes the BBC technology editor saying the PHP story feels like the cyber equivalent of an unexploded bomb, and that she has never seen such a dramatic, public and global response to a cyber event that has not happened yet. The article says no successful attacks have been confirmed, and that one security researcher was up all night trying to exploit the flaw but found it very difficult. The news was leaked before the coordinated disclosure, and from that point the clock started ticking. Professor Alan Woodward describes the dilemma of alerting people while not alerting attackers. Cloudflare and Palo Alto Networks have seen millions of attempts. Gina Neff, Professor of Responsible AI at Queen Mary University, told the Today Program that a team supported by a large AI company used AI to find the problem, worked for months to solve it, and got everybody ready, describing enormous capability to handle these risks.
eMudhra to Exit Global SSL/TLS Certificate Business, from Hacker News: Newest, carries a press release title announcing that eMudhra will exit the global SSL/TLS certificate business and sharpen focus on core enterprise solutions and trust services other than SSL/TLS.
Indian tech company eMudhra announces withdrawal from certificate business after security incident, from The Record from Recorded Future News, adds the cause: the issue was publicized by cybersecurity researcher Scott Helme, who published a blog saying that instead of a random 384-bit number, one of the company's private keys was just the number four.
Developer platform Railway urges 'all users' to revoke secret keys after cyberattack, from TechCrunch, reports that the vibe-coding platform said all of its customers should revoke and rotate their private keys after hackers were able to decrypt some customer secrets.
Microsoft blames ConnectWise attacks on Chinese state hacker using ARTEX AI agents, from The Record from Recorded Future News, reports that a Chinese government-aligned hacking group used automated tools to carry out the recent supply chain attack on IT management software company ConnectWise and is using the same techniques to try to compromise organizations around the globe.
Chinese hackers spied on thousands of companies after breaching IT giant ConnectWise, from TechCrunch, adds that Microsoft points the blame at a China-aligned hacker who relied on AI agents to rapidly hack into thousands of its customers.
South Carolina officials offer assistance after state capital hit with cyberattack, from The Record from Recorded Future News, reports that the systems impacted include the MyColaSC platform that citizens use to pay utility bills, report non-emergency issues and access other city services.
Cybercrime crackdown across Africa leads to 1,462 arrests, $105 million seized, from The Record from Recorded Future News, reports that more than fourteen hundred suspected cybercriminals were arrested across Africa this summer as part of a crackdown on gangs perpetrating online fraud, ransomware and business email compromise attacks, according to Interpol.
This Woman's Living a CashApp Scam Nightmare, from 404 Media, reports on Amanda Smith, who answered her door in upstate Pennsylvania and was arrested on a warrant for computer fraud allegedly committed more than a thousand miles away in Arkansas. A scammer had tricked someone on Facebook Marketplace into sending a thousand dollars to a CashApp account in her name, created using her Social Security number and date of birth. Smith spent three nights in jail and now faces an ongoing criminal case, while CashApp refuses to provide her any information about the account because she could not prove she was the owner. The article describes the situation as Kafkaesque.
A second theme concerns AI in warfare and the ethical reactions inside technology companies.
AI Is Already 'Amplifying the Destruction' of War, from 404 Media, reports on the UN meeting this week discussing autonomous weapons. Laurent Gisel, the head of the International Committee of the Red Cross's legal warfare team, told 404 Media that the technology is developing at a pace that is very concerning. Gisel addressed the UN about illegal arms trade, nuclear weapons, mines and AI. The article says Israel used an AI powered by Microsoft to pick targets in Gaza, AI-powered Russian drones are deciding which humans to kill in Ukraine, and the US military is pitching itself as an AI-first fighting force. Secretary of Defense Pete Hegseth recently announced Project Meridian, with three leads being Elon Musk, defense contractor Palmer Luckey and Newt Gingrich. Gisel says AI is amplifying the destruction, the violation of the law, and the humanitarian harm caused by conflicts. One of the biggest dangers is speed: targets are created at higher speed and frequency and prosecuted quicker, raising concerns about the level of human control retained. The Israel Defense Forces told journalists it did not thoroughly vet kill lists generated by Lavender. Gisel says humans tend to trust systems, more so when tired, stressed and needing to deliver quickly, which is exactly the situation soldiers are in. Gisel has red lines: any AI support system that removes a human being from the kill chain requires specific limits and guardrails, and autonomous weapon systems that target human beings directly should be prohibited on ethical, humanitarian and moral grounds. He also argues that banning unpredictable weapons should be a non-brainer, since military discipline requires delivering the commander's intent. Gisel opposes AI in nuclear command and control, citing Stanislav Petrov's decision in 1983. The article ends with his warning that once the genie is out of the bottle, it is very difficult to put back, and that otherwise it is the technology that will decide what is acceptable.
Google Staff Resign In Backlash Over Its AI Contracts With Military, from Slashdot, quotes a BBC report that Google is facing a backlash from staff over its deployment of AI for military use. Several tech workers have recently resigned from Google DeepMind or are leaving over military contracts. Sofia Liguori, a London-based research engineer, resigned this week and told the BBC that realizing her company is pushing more and more for AI systems in military operations makes her feel ashamed. She cited Google's work with the US and Israeli militaries and said she has lost faith in the company being true to its stated ethical values. Andreas Kirsch, an AI researcher who also resigned, said Google's governance and oversight are not sufficient or meaningful enough to deploy AI safely and responsibly in military and classified settings, and that Google seems to have no right to control or veto how the Pentagon uses its AI and limited visibility into how it is deployed. Employees told the BBC they are scared to speak out because contracts prohibit public criticism. Google said it follows a consistent process for reviewing contracts and has dedicated channels for employees to raise concerns.
A third theme covers AI products, agents and platform moves.
Lumo can now turn your ideas into creations, from The Proton Blog, announces that Lumo can create documents, code and interactive web pages in a canvas alongside chat. Users can ask it to draft a report, help with coding, or create an early 2000s-style video game. The article says Lumo never logs conversations or uses them to train models, and creations are protected with zero-access encryption. The new canvas gives work a dedicated view with conversation beside it, and Lumo distinguishes between conversational answers and creations. Revisions create new versions, and earlier versions remain available. The article explains the technical challenge of showing creations across conversations without servers being able to read them: Lumo saves a separate encrypted copy of each creation's latest version. Export conversion happens on device, with Word files generated locally and PDF downloads using the browser print dialog. Keeping conversion local means Save to Drive PDF currently produces image-based PDFs without selectable text. Creations are available now for all Lumo users, with presentations coming soon.
Claude Duet app is Anthropic's bet on agents, from TechCrunch, reports that Claude Duet is available through waitlist for Pro and Max plan users.
Anthropic Takes On Muse With Claude Duet, from Slashdot, adds detail from The Verge: Claude Duet is an AI agent aimed at professionals and small businesses that competes with Meta's Muse and OpenAI's Dots. Each morning it sends a short brief on important items across connected apps, keeps tabs on those apps, and messages only when something needs attention. It can handle administrative tasks, build dashboards with dozens of tools, create financial or sales forecasts, prototype and build web apps. It pulls in memory, skills and connected apps from Claude, and has a virtual machine with a browser. Before any big step like sending a payment or deleting something, it messages for approval. It has a dedicated classifier and safety system prompt designed to reject requests to share the strictest tier of information. Anthropic says it is an early version that might misread a thread or miscalibrate urgency, and it is not as strong at personal errands like booking tickets.
Donating Agent Skills to the Agentic AI Foundation, from Simon Willison's Weblog, reports on Anthropic's announcement that it is donating the Agent Skills standard to the Agentic AI Foundation under the Linux Foundation, which it co-founded with Block and OpenAI last year and to which it previously donated the Model Context Protocol. Simon Willison notes he called this last December when Anthropic first declared skills an open standard, and that he found that prediction by prompting Claude to use his blog's MCP plugin.
Spotify's current AI bet is about being the music service that agents choose, from TechCrunch, reports that Spotify co-CEO Gustav Söderström explains why the rise of AI agents could be good news for the music service.
1Password launches a way for AI agents to log in on your behalf without seeing your password, from TechCrunch, reports that the new feature allows agents to log into a customer's personal accounts while still protecting the private credentials that secure those accounts.
Canva releases its first image generation model, from TechCrunch, reports that Canva is making its new image model available to all users, promising better, more customizable performance at lower costs.
Uber puts all 30,000 employees on ChatGPT Work, from TechCrunch, reports that the ride share giant previously deployed a much smaller pilot for ChatGPT Work but is now taking the enterprise agent company-wide.
Arc browser sunsets December 22 as The Browser Company doubles down on Dia, from TechCrunch, reports that Arc was fiercely beloved by a dedicated group of enthusiasts but never quite became the mass-market Chrome alternative its makers once envisioned.
Nirva raises $33M for a pendant that only listens to you, from TechCrunch, reports that Nirva will ship its two hundred forty-nine dollar ambient AI pendant by the end of the year.
DeepSeek reportedly raises $15B days after US restrictions, from TechCrunch, reports that DeepSeek has reportedly raised fifteen billion dollars in one of China's largest funding rounds, just days after the US ordered federal agencies to remove its technology over national security concerns.
A fourth theme groups three long reports from New Lines Magazine on Middle East politics.
The Palestinian Politician Who Could Be Israel's Kingmaker, from New Lines Magazine, examines Mansour Abbas and the possibility that Arab parties could decide Israel's next government. The article reports that if polls are correct, Benjamin Netanyahu will lose the national election on October 27, with his coalition short of the necessary sixty-one seats, but the opposition bloc is also several seats short. The only way for the opposition to guarantee a majority is by aligning with one or more parties representing Palestinian citizens of Israel, yet the opposition parties are adamantly opposed to including the so-called Arab parties. All eyes are on Mansour Abbas, an Islamist politician aiming to break through the boycott. In 2020 he declared willingness to enter any government, divested his party's manifesto of Palestinian national aspirations, and recognized Israel as a Jewish state. In August he announced that Yoav Segalovitz, former chief of police, would join his party. The article traces Abbas's history: he comes from Maghar, began giving Friday sermons at seventeen, and became a disciple of Abdullah Nimar Darwish, a former hard-liner who became a moderate integrationist. The UAL became the first Arab party to join a governing coalition in 2021 under Naftali Bennett and Yair Lapid. Segalovitz commanded Lahav 433 and is positioned as the UAL's candidate to replace Itamar Ben-Gvir as minister of national security. The article says the move is cold political calculation: Segalovitz's inclusion was designed to break the Jewish-Zionist parties' boycott of Arab parties. But it is also aimed at Palestinian voters, because homicides within Palestinian communities have skyrocketed, with at least 157 of 201 homicides in Israel this year involving Arab citizens, even though they make up less than twenty percent of the population. A May poll found violence and crime is by far the top concern among Palestinian citizens at seventy-one percent, while settling the Palestinian issue ranked second at eight percent. The article compares Abbas's trajectory to Ayman Odeh and the Joint List, which reached fifteen seats in 2020, and describes how the Jewish majority's red lines were firm, preferring to allow Netanyahu to remain in power rather than head an Arab-backed government. A February 2025 survey found sixty-seven percent of Jewish Israelis oppose including Arabs in a governing coalition. The article reports that support for expelling Gaza's residents and Israel's own Palestinian citizens has risen sharply, and that the Central Elections Committee has routinely disqualified Palestinian candidates. It concludes that the Israeli opposition may well choose another round of elections over partnering with Arabs, even the good Arabs like Mansour Abbas.
From Sisi's Satirist to Sisi's Guest: Bassem Youssef's Surprise Appearance, from New Lines Magazine, reports on Bassem Youssef's appearance by video at the inaugural Egypt Forum 2026, launched by President Sisi. The article recalls that in the 2013 season premiere of El Bernameg, Youssef warned that fascism in the name of religion will be replaced by fascism in the name of patriotism and national security. The day after that episode, he was hit with at least four legal complaints, and the broadcaster pulled the following week's episode. On October 19, thirteen years after leaving Egypt, Youssef appeared at the forum, discussing social media, urging Arabs not to let online feuds turn them against each other, and expressing longing for Egypt. The article says his image was splashed all over state media, with clips focusing on Sisi's reaction. Some saw it as an endorsement of the authoritarian president, and commentators portrayed the comedian as a self-seeking opportunist. The article traces Youssef's rise after Mubarak's ouster, his satirical commentary on Morsi and the Muslim Brotherhood, his arrest and release, and his celebration of the coup after the Rabaa massacre, which some critics saw as making light of the brutal crackdown. In the documentary Tickling Giants, Youssef says network executives gave him a choice to continue as long as he stopped making fun of authority, and he refused. After El Bernameg was canceled, Youssef said he was tired of fighting and fearing for his safety. The article says his appearance at Egypt Forum reinforced familiar regime rhetoric in which profound issues can be explained away by foreign meddlers rather than abuses of power at home. It notes this was not his first return to Egyptian media, and that the desire to return to his homeland has not waned. His father was killed in a car accident in 2015, and Youssef was unable to attend the burial. The article closes with his words: my daughter may not get to grow up in our country, I don't know if I will ever be able to go back, I want to go back.
A Saudi on the Silent Fear of the Houthi Strikes, from New Lines Magazine, reports on the latest Houthi escalation against Saudi Arabia, with drones and missiles reaching King Khalid International Airport in Riyadh and King Fahd International Airport in Dammam. The author, a Saudi, says the fear has a silent nature, leaving little trace on social media or public discussion, and that Saudis find themselves with no voice representing them despite being directly affected. Some people responded only with hints, signaling with eyes and lips not to speak. One person compared Crown Prince Mohammed bin Salman unfavorably with previous monarchs under whom Saudis trusted their leaders and did not experience economic uncertainty. Another said they have gotten used to hearing Houthi missiles, that hearing drones and explosions has become normalized. The article says excessive nationalism has taken hold, with demands that foreign workers publicly express support for the kingdom, and the arrest of a foreign teacher feeds into this. A liberal friend in Riyadh described how religious hegemony has been replaced by staunch state-mandated nationalism, with digital surveillance targeting anyone who expresses doubt in state policies. The old religious weapon of declaring someone an infidel has been replaced by declaring someone a traitor. The article closes by reflecting on the Kuwait war, the Khobar Towers bombing, 9/11, the invasion of Iraq and the 2011 uprisings, and says the uncertainty feels more tangible now, and the harsh tone is not one Saudis want their country to be known for.
A fifth theme covers business, funding and science in technology.
Fusion startup Helion is developing new semiconductors that could double output of its power plants, from TechCrunch, reports that Helion now has another two hundred twenty-five million dollars to help speed development of its first fusion power plant, due to provide Microsoft with electricity in 2028.
Investors break down how founders can raise a Series A today, from TechCrunch, reports that raising a Series A is not simple, and top investors break down what founders need to know now.
Elon Musk made a $1 billion investment in Skydance, report says, from TechCrunch, reports that the investment will provide a much needed financial boost as the company seeks to cut costs and lift its stock price.
Former OpenAI science lead launches Warp Science to automate wet labs with robots, from TechCrunch, reports that the new startup raised a two hundred fifteen million dollar round led by Lightspeed.
Microsoft's Mustafa Suleyman says its AI outperforms doctors on hard cases as he chases 'medical superintelligence', from TechCrunch, reports that Microsoft's new AI diagnostic system beat generalist doctors on hard cases, according to a study in Science, but most of the improvement seems to come from the AI models underneath, not from Microsoft's system itself.
The secret to fast startup growth isn't early sales, top names in tech say, from TechCrunch, reports that at this year's Disrupt, Lovable GTM lead Luke Arno, Polska's Ben Broca, and Theory Ventures' Tomasz Tunguz gave founders tips on how to achieve sales success in the first ninety days.
Moderna Shares Plunge After Leak of Cancer Vaccine Data, from Slashdot, reports that Moderna shares plunged as much as twenty-five percent after partial clinical trial data leaked for its personalized cancer vaccine developed with Merck, despite full results showing promising outcomes for melanoma patients. The personalized vaccine paired with Keytruda reduced the risk of recurrence or death in certain melanoma patients by forty-two percent compared with Keytruda alone, and reduced the risk of cancer spreading to distant parts of the body or death by thirty-eight percent. About eighty-two percent of patients who received the vaccine alongside Keytruda were alive without their melanoma returning at two years, compared with seventy percent for Keytruda alone. Analysts attribute the selloff largely to inflated investor expectations rather than disappointing clinical results.
TechCrunch Disrupt 2026 Startup Battlefield 200: Celebrating outstanding achievements, from TechCrunch, reports that among the Startup Battlefield 200 were many industry-defining companies exhibiting and pitching on the Showcase Stage, with standouts highlighted.
A separate article on space follows.
An Interview with Craig Moffett About SpaceX and Bubbles, from Stratechery by Ben Thompson, is an interview with Craig Moffett about SpaceX, from launch to broadband to mobile, plus how he is applying the lessons of the dot-com bubble to today.
Four articles from Nautilus form a science group.
Watch Humpback Whales Team Up With Sea Lions to Hunt, from Nautilus, reports on a beneficial partnership in Monterey Bay.
A New Fossil Hints at the Surprising Origins of Snakes, from Nautilus, reports that by the Cretaceous, snakes had already tried on several different lifestyles and sensory styles.
We Now Know What Denisovans Looked Like, from Nautilus, reports that an analysis of their skulls paints a picture of our archaic cousins.
LSD May Speed Up Brain Activity and Make it Fractal, from Nautilus, reports that a new study suggests a more complex reshuffling of experience.
A consumer and hardware group follows.
Airbnb Bans Making Guests Do Dishes, Take Out the Trash, from Slashdot, reports that Airbnb is banning hosts from requiring guests to perform time-consuming checkout chores, including washing dishes, stripping beds, doing laundry and taking trash off-site. The updated policy applies to reservations booked on or after November 20. Airbnb is displaying hosts' requests in a new rules and requests section, and hosts cannot sneak in extra rules in messages or signs. Some hosts pushed back, saying the requirements pose health risks. Exceptions include asking guests to dispose of trash in an animal-proof container in bear country.
Spotify confirms Family Plan crackdown, but calls it a 'routine improvement', from TechCrunch, reports that Spotify told investors its rollout of QR code verification to ensure Family Plan members live in the same household is not a crackdown, just a routine improvement.
Tesla is prepping a Cybercab onslaught, from TechCrunch, reports that the company is looking to roll the gold two-seater robotaxis out to other U.S. cities this year, even as it goes mum on the number of paid rides.
Bose Is Buying Skullcandy, from Slashdot, reports that Bose announced it is acquiring Skullcandy, the lifestyle brand known for bass-heavy headphones. Skullcandy will remain a separate company within Bose's Premium Consumer Audio business and will stay in Park City, Utah. The article says Bose seems to be after Skullcandy's connection to its community and a customer base Bose largely misses, and that the acquisition lets Bose compete at both ends of the market.
Gadget Review: Mokin 17 Port USB-C Tower, from Terence Eden's Blog, is a two-star review of the MODK2403 docking station with triple display support, four USB-A ports, two USB-C ports, Ethernet, SD and microSD, and a suction cup base. The reviewer notes it worked perfectly with a Linux laptop, driving two external screens, and achieved 850 megabits per second through Ethernet. However, the review raises significant concerns about the always-on WiFi with a weak default password, the lack of a privacy policy, and the presence of a JTAG debug unit and an unexplained HID device. The manufacturer said the password can be changed and the entries are reserved for future firmware updates, but the reviewer does not find this good enough and would be reluctant to connect it to a home WiFi network.
Qotom Q51551UT 4x 10GbE and 6x 2.5GbE Mini PC Review, from ServeTheHome, is a review of a network-focused industrial mini PC from Qotom with ten Ethernet ports supporting data rates as high as ten gigabits per second.
On data centers in Texas.
Portal Texas Used to 'Rubber Stamp' Data Centers Down After 404 Media Reporting, from 404 Media, reports that Texas has shut down the web portal for its automated construction permit filing system and suspended construction permits of seventeen data center projects. The change came after 404 Media reported that the state had continued to issue construction permits related to data centers despite Governor Greg Abbott calling for a pause in September. Watchdog groups FracTracker Alliance and Public Citizen published reports detailing more than two dozen permits issued to Texas data center projects. The article reports that before the shutdown, 404 Media viewed seventeen permits related to data center projects that had been moved from active to suspended, including for OpenAI and Amazon. Sam Law of FracTracker Alliance said the portal was rubber stamping construction approvals and seemed to have been quietly issuing approvals since the governor's order, and that the irony of suspending approvals and switching to human review is that some work is better left to humans.
A single compliance article follows.
Localizing compliance at scale: achieving South Korea's ISMS certification, from Cloudflare Blog, describes what it took for Cloudflare to obtain South Korea's Information Security Management System certification under the ISMS-P framework. For ten days, auditors appointed by the Korea Internet and Security Agency sat with Cloudflare and worked through eighty controls. The most visible result is the account sign-up page: users signing up from South Korea see four separately labeled consent items, three required and one optional, with the sign-up button disabled until the required items are checked. The hard part was sign-up paths that skip the form entirely, such as social sign-in with Google, Apple or GitHub. Cloudflare chose to remove social sign-up in South Korea so every account goes through the form. The article explains that the country signal is derived from the client IP address, but IP geolocation is an estimate, so the country a user selects at sign-up overrides the default. The article describes ISMS as a detailed, localized certification standard covering security policy, access control, encryption, incident response and other controls assessed against Korea's Personal Information Protection Act. Certification requires an independent audit appointed by KISA, with an annual audit to maintain it. The audit examined data flows with detailed diagrams, and scrutinized the language and terminology of consent collection. The timeline ran to a year and a half because preparation ran past one audit year and into the next. The certified scope is defined as a service boundary rather than an enumerated list of products, which differs from the shape of SOC 2 or ISO 27001 reports. The article concludes that ISMS is not a harder certification but a more nuanced one, built around the regulatory needs of one specific region, and that going through it made Cloudflare more global, not less.
The remaining articles are essays, community notes and assorted items.
The age of initiative, from Seth's Blog, argues that people grew up in an age of compliance and indoctrination, and that the long tail of the internet and the leverage of AI has made that training much less valuable. There is a machine standing by waiting for instructions, from you if you choose, or from your overlords if you don't. Following the playbook is not much help because the world is changing so fast. AI gave people seeking compliance a chance to wait for the AI to tell them what to do next, and gives bosses a cheaper, more compliant alternative to hiring humans. For those indoctrinated to comply, this is stressful, because they want to get back to normal, but right now is as normal as things are ever going to get. Good jobs are no longer sinecures where success comes from following instructions. The article says winging it is insufficient and handling the transition alone is too difficult, but there is a chance to find others and connect with a cohort that challenges, educates and encourages. The closing line: now there is a chance to lean into the hard work needed to thrive in an age of initiative.
The Science of Doing Your Best Work, from Farnam Street, is an announcement for a podcast conversation with Steven Kotler, bestselling author and executive director of the Flow Research Collective. Kotler has spent decades studying flow and makes the case that it is far more trainable than most people think. The conversation covers what brings flow on, how hard a task should be, why the hobby given up as an adult might matter, what frustration is really telling you when learning, how play compares with practice, how to focus on work you would rather not be doing, and how to think clearly when there is more information than ever and fewer people checking it. The public release is October 27, with members getting the extended version now.
Other people, from Aeon, is an essay by Elizabeth Oldfield about disconnection being at the heart of today's crises, and her and her husband's choice of a radical remedy: living communally.
Bhutan: the last place, from Aeon, is a video about what happened when people in Bhutan, the last country on Earth without television, were allowed to watch it in 1999.
San Francisco Meetup This Saturday, from Astral Codex Ten, announces a meetup at Mox in San Francisco on Saturday, October 24, from one in the afternoon to late afternoon. The author plans to attend and will do a Q and A if the sound system is up for it. Anyone is welcome, and anyone caught expressing policy opinions will be herded into a side room to help create a San Francisco voting guide.
The hotlist, from Seth's Blog, announces a page containing many of the author's books. Clicking on any one takes the visitor to a directory of where to get a copy, and the page re-indexes based on which books get the most clicks.
Modern Principles of Economics!, from Marginal REVOLUTION, is a promotional post for an economics textbook with over one hundred videos, iClicker, Achieve and additional support, directing instructors to contact Mike Krotine for an examination copy.
Thursday assorted links, from Marginal REVOLUTION, offers six links: one on the cherries trade, one on a new YouTube channel for Indian architecture, one on where Math Olympiad gold medal winners end up, noting that China has 105 out of a possible 120 gold medals over twenty years, one asking whether El Salvador is cost-competitive, one on Rebecca Lowe on Bach, and one asking whether the Democrats still care about the Ivy League.
DE STAAT - FREEDOM (Official Music Video), from Simon Willison's Weblog, notes a first for the author: at forty-four seconds into the music video by the Dutch alternative rock band De Staat, the visual effects are credited to Matties Grooten of Fake it or Leave it.
The sixty articles in this window are dominated by a single security story that touches several others: the PHP vulnerabilities, their AI-assisted discovery, the accidental disclosures around Linux and eMudhra, and the broader argument that AI has changed both hacking and defense. Around that core, the window carries substantial reporting on AI in warfare, on Israeli, Egyptian and Saudi politics, and on a broad set of AI product launches, funding rounds, and scientific findings.
- The age of initiative
- Miner Subsidy Crisis & Unit Bias: Why Bitcoin Needs a Transactional UoA
- The Science of Doing Your Best Work
- Other people
- An Interview with Craig Moffett About SpaceX and Bubbles
- This Day on SN: October 22
- Bhutan: the last place
- The Palestinian Politician Who Could Be Israel’s Kingmaker
- Samsung and SK Hynix are upgrading Chinese factories to produce advanced RAM
- One Year After $126K: Is the Bitcoin Treasury Playbook Still Working?
- A piece of Hope With ₿itcoin history deserves a special home.
- Deadly Shamonda virus detected in human for first time in France
- Airbnb Bans Making Guests Do Dishes, Take Out the Trash
- Several Linux kernel vulnerabilities accidentally posted to public mailing list
- 2026-27 UCL Pool: Week 3 recap and standings
- Modern Principles of Economics!
- Gadget Review: Mokin 17 Port USB-C Tower ★★☆☆☆
- Amazon identifies Oct 17 PHP2Shell exploitation prior to public disclosure
- Bitcoin Activism (Not a Social Movement)
- Unpatched CVSS 9.0 Linux kernel RCEs accidentally posted to public LKML
- South Carolina officials offer assistance after state capital hit with cyberattack
- Lumo can now turn your ideas into creations
- Localizing compliance at scale: achieving South Korea’s ISMS certification
- Woman faces six years in prison over Cash App account she never opened
- DeepSeek reportedly raises $15B days after US restrictions
- eMudhra to Exit Global SSL/TLS Certificate Business;
- New Serious PHP Vulnerabilities Released and Patched
- Nirva raises $33M for a pendant that only listens to you
- Cybercrime crackdown across Africa leads to 1,462 arrests, $105 million seized
- eMudhra to Exit Global SSL/TLS Certificate Business
- Claude Duet app is Anthropic’s bet on agents
- Developer platform Railway urges ‘all users’ to revoke secret keys after cyberattack
- From Sisi’s Satirist to Sisi’s Guest: Bassem Youssef’s Surprise Appearance
- AI Is Already 'Amplifying the Destruction' of War
- AI Is Changing Everything About Hacking. No One Knows What Happens Next
- Watch Humpback Whales Team Up With Sea Lions to Hunt
- Quad9 (DNS Resolver) to Exit France
- Fusion startup Helion is developing new semiconductors that could double output of its power plants
- Google Staff Resign In Backlash Over Its AI Contracts With Military
- Magic Patterns CEO Reviews Startup Websites
- Spotify’s current AI bet is about being the music service that agents choose
- AI #191: The Hackening Begins
- Investors break down how founders can raise a Series A today
- Spotify confirms Family Plan crackdown, but calls it a ‘routine improvement’
- Google cleared 300 hectares of Finnish forest
- This Woman's Living a CashApp Scam Nightmare
- Quoting Armin Ronacher
- Hashrate Hearth: How many home miners does it take to screw in a freedom-bulb?
- Tesla is prepping a Cybercab onslaught
- Thursday assorted links
- Uber puts all 30,000 employees on ChatGPT Work
- TechCrunch Disrupt 2026 Startup Battlefield 200: Celebrating outstanding achievements
- Several Linux Kernel Vulnerabilities Accidentally Posted To Public Kernel Mailing List
- A farewell to Arc browser
- Microsoft blames ConnectWise attacks on Chinese state hacker using ARTEX AI agents
- Chinese hackers spied on thousands of companies after breaching IT giant ConnectWise
- Anthropic Takes On Muse With Claude Duet
- A New Fossil Hints at the Surprising Origins of Snakes
- Elon Musk made a $1 billion investment in Skydance, report says
- Nvidia OpenShell
- Qotom Q51551UT 4x 10GbE and 6x 2.5GbE Mini PC Review
- Zoe Kleinman: PHP Is the Cyber Equivalent of an Unexploded Bomb
- We Now Know What Denisovans Looked Like
- Donating Agent Skills to the Agentic AI Foundation
- Former OpenAI science lead launches Warp Science to automate wet labs with robots
- Microsoft’s Mustafa Suleyman says its AI outperforms doctors on hard cases as he chases ‘medical superintelligence’
- A Saudi on the Silent Fear of the Houthi Strikes
- Indian tech company eMudhra announces withdrawal from certificate business after security incident
- Bose Is Buying Skullcandy
- Canva releases its first image generation model
- 1Password launches a way for AI agents to log in on your behalf without seeing your password
- DE STAAT - FREEDOM (Official Music Video)
- San Francisco Meetup This Saturday
- LSD May Speed Up Brain Activity and Make it Fractal
- The secret to fast startup growth isn’t early sales, top names in tech say
- Moderna Shares Plunge After Leak of Cancer Vaccine Data
- Arc browser sunsets December 22 as The Browser Company doubles down on Dia
- Portal Texas Used to ‘Rubber Stamp’ Data Centers Down After 404 Media Reporting
- The hotlist

