183623502
submission
BrianFagioli writes:
President Donald Trump has signed a new national security memorandum designed to accelerate the use of artificial intelligence across the US military and intelligence community. The directive specifically mentions adapting both commercial and open source AI technologies for mission use, while also pushing for faster deployment of advanced AI systems from multiple vendors. The administration says the move will help America maintain technological dominance and reduce dependence on any single AI company.
The memorandum replaces Biden-era AI guidance that Trump officials describe as outdated and overly restrictive. It also calls for updated rules on autonomous weapon systems and promises that military AI will not be used for censorship or unlawful domestic surveillance. From the report: âoeThe men and women who defend the United States deserve the most advanced AI in the world.â
183611036
submission
BrianFagioli writes:
Kraken says it plans to offer tokenized exposure to the eventual SpaceX IPO through its xStocks platform, allowing eligible users in more than 110 countries to trade around the clock, including weekends. The crypto exchange argues that traditional stock market hours are outdated and says blockchain-based equities could make IPO access more global and accessible. From the report: âoeThis isnâ(TM)t a friendlier version of the old system. Itâ(TM)s a new one: borderless, always on, and built on crypto rails.â
The move raises larger questions about whether tokenized equities could eventually pressure traditional exchanges to modernize. Kraken says the tokens will be backed 1:1 by underlying shares held by a regulated custodian, although users will not receive voting rights or dividends. The offering also excludes U.S. users despite Kraken being one of the best-known American crypto brands.
183609682
submission
BrianFagioli writes:
ReliaQuest says it uncovered a previously unknown China-linked hacking cluster called OP-512 that has reportedly been targeting outdated Microsoft IIS servers running unsupported .NET Framework software. According to the security company, the attackers used custom web shells, encrypted command channels, timestomping, and DNS-based âoephone homeâ techniques designed to evade traditional antivirus detection and maintain long-term access for espionage operations.
The company claims its âoeAgentic AIâ platform pieced together what initially looked like unrelated low-level security events into a single coordinated intrusion. ReliaQuest says the malware framework generates cryptographically unique deployments that make signature-based detection ineffective. The report also warns that organizations still exposing legacy IIS infrastructure to the internet remain attractive targets for increasingly sophisticated state-linked attackers.
183605890
submission
BrianFagioli writes:
Cloudflare has acquired VoidZero, the company behind Vite, one of the most widely used JavaScript build tools on the web today. The deal also brings related open source projects like Vitest, Rolldown, and Oxc under Cloudflareâ(TM)s umbrella, giving the company enormous influence over modern web development infrastructure. Cloudflare insists the projects will remain MIT licensed, vendor-neutral, and community driven, while also pledging $1 million toward an independent Vite ecosystem fund.
The acquisition highlights how AI-generated code is reshaping software development. Cloudflare says AI coding agents are increasingly choosing Vite-based workflows because of their speed and compatibility, with the Cloudflare Vite plugin now accounting for nearly 14 million weekly downloads. While developers may appreciate tighter integration and faster deployment tooling, some are likely to worry about another major piece of the open source web stack becoming closely tied to a giant infrastructure provider.
183593554
submission
BrianFagioli writes:
Some of the biggest names in artificial intelligence and biotechnology are now warning Congress that advanced AI systems could make biological weapons easier to create. Leaders from OpenAI, Google DeepMind, Anthropic, Microsoft AI, Meta, and several biotech firms signed a public letter calling for mandatory screening and recordkeeping for synthetic DNA orders in the United States. The group argues that AI systems are rapidly improving at answering complex biology and virology questions, potentially lowering the expertise barrier for dangerous research.
The proposal would require DNA synthesis companies to screen customer orders for sequences linked to pathogens and maintain records that could help investigators trace suspicious activity. While many companies already do this voluntarily, the signatories say federal rules are now urgently needed. Critics will likely see the effort as another step toward scientific surveillance, especially as the same companies building increasingly powerful AI systems are also warning about the risks those systems may create.
183587736
submission
BrianFagioli writes:
Fedora Linux 43 users upgrading to the latest Dovecot mail server discovered something rather unsettling: some older Microsoft Outlook configurations may have been silently ignoring SSL/TLS settings for POP3 email connections for years. According to a Fedora community blog post, affected Outlook clients reportedly continued using insecure port 110 connections even when encryption was enabled in the application settings. The issue surfaced after Dovecot 2.4 disabled plaintext authentication on non secure connections by default, causing Outlook users to suddenly lose mailbox access after the Fedora 43 upgrade.
The report suggests the behavior may date back as far as Outlook 2007, although modern Outlook builds were not fully tested. Fedora admins stress that the problem could be limited to legacy account configurations rather than current versions of Outlook itself. Still, the discovery has sparked discussion among Linux admins and security folks because many users likely assumed their email traffic was encrypted simply because Outlook claimed SSL/TLS was enabled. The incident also highlights how stricter defaults in modern open source infrastructure can expose ancient assumptions and questionable behaviors that quietly survived for decades.
183578840
submission
BrianFagioli writes:
Plex used to be the app self hosting nerds loved because it stayed focused on organizing and streaming personal media libraries. Now the company is adding discussions, emoji reactions, compatibility scores, social following features, and community conversations tied to movies and TV shows. Coming right after the controversial Plex Pass price increase, many longtime users are questioning whether Plex still understands why people liked the platform in the first place.
Some of the new tools could be useful, especially shared watchlists across streaming services, but a growing number of users seem far more interested in reliability, playback improvements, and core media server features than turning Plex into another engagement driven social platform. The bigger Plex gets, the more it risks becoming exactly the kind of bloated streaming ecosystem many users were trying to escape.
183577980
submission
BrianFagioli writes:
A new survey from AI career transition company Pelgo suggests AI may already be reshaping how unemployed Americans think about higher education. Only 20 percent of respondents said they would choose a four-year college degree again if given the chance to start over, while growing numbers said they would instead pursue trade school, entrepreneurship, or faster entry into the workforce through two-year programs.
The findings raise an uncomfortable question for universities: what happens when white collar jobs increasingly look vulnerable to automation while skilled trades still require humans physically showing up somewhere? As AI continues creeping into office work, many workers appear to be questioning whether massive student loan debt still makes sense in 2026.
183575010
submission
BrianFagioli writes:
Google is finally giving publishers actual visibility into how their content performs inside AI search features. The company announced new Search Console reports focused specifically on generative AI experiences like AI Overviews, AI Mode, and AI-powered Discover results. Site owners will now be able to see impressions, pages, countries, devices, and time-based performance data tied directly to AI-generated search experiences.
Google is also testing a new opt out toggle that lets publishers prevent their content from appearing in generative AI Search features entirely, without impacting normal Search rankings. For independent publishers and SEO folks, this is probably the most important part of the announcement. Until now, nobody really had clean data showing whether AI summaries were helping or cannibalizing visibility.
183454586
submission
BrianFagioli writes:
Patreon is making a notable privacy change that could catch some users off guard. Beginning July 27, 2026, any new creator memberships joined on the platform will be publicly visible by default unless users manually change their privacy settings. Existing memberships will remain private unless users choose otherwise, but Patreon is clearly pushing toward a more social, discovery-focused platform with features like public creator memberships, mutual community connections, and more visible activity feeds.
The company says the goal is to help fans discover creators through shared interests and connections, but the move also changes the feel of the platform. Patreon historically operated more like a direct creator-to-supporter relationship than a social network. Now it increasingly resembles something closer to a creator-centric social graph. While Patreon is adding more granular privacy controls, critics may argue that many users will never notice the changes until their subscriptions become visible to others.
183445686
submission
BrianFagioli writes:
announced a new handheld gaming device called the Nitro Blaze Link, but unlike devices such as the Steam Deck or ASUS ROG Ally, this one is not trying to run games locally. Instead, Acer describes it as a “streaming-first” handheld designed to stream games from an existing gaming PC using Sunshine and Moonlight. The company says the device runs Debian Linux, includes a 7-inch WUXGA touchscreen, Wi-Fi 6, and weighs just 464 grams. Curiously, Acer never disclosed the processor powering the device, while the published specs list only 1GB RAM and 8GB eMMC storage.
The idea here seems pretty simple: instead of cramming increasingly power-hungry GPUs into portable gaming PCs, Acer is betting some gamers would rather have a lightweight Linux streaming terminal for couch gaming around the house. The Nitro Blaze Link is expected to launch in North America during Q4 2026, although Acer has not announced pricing yet.
183434552
submission
BrianFagioli writes:
LG Display says it has started mass production of what it calls the worldâ(TM)s first 240Hz RGB Stripe OLED panel, aiming to address one of the biggest complaints about OLED monitors: text clarity. Unlike many current OLED displays that use alternative subpixel layouts, the new panel uses a traditional RGB stripe arrangement that LG says improves readability for coding, spreadsheets, document editing, stock trading, and other desktop-heavy workloads. The 27-inch panel also combines a 160 PPI pixel density with support for switching between 4K at 240Hz and FHD at 480Hz using the companyâ(TM)s Dynamic Frequency & Resolution technology.
OLED monitors have traditionally been associated with gaming and media consumption, while many office users continued sticking with IPS LCD panels due to concerns over text rendering and burn-in. LG Display appears eager to change that perception by positioning OLED as a single display solution for both productivity and gaming. The company says it is beginning production alongside major monitor brands, although it did not name specific partners or products yet.
183429784
submission
BrianFagioli writes:
Microsoft has published a new research paper arguing that AI systems are not replacing human intelligence, but instead extending structures already rooted in human cognition and language. The paper claims large language models work because they absorb and remix patterns humans have embedded into writing and communication over generations, not because the systems possess true understanding or consciousness. Microsoft also points to hallucinations and reasoning failures as evidence that current AI still lacks real-world grounding and compositional reasoning comparable to humans.
The company additionally pushes back on fears of âoerogue AI,â arguing the larger risk comes from humans deploying flawed AI systems irresponsibly at scale. Critics, however, may see the paper as an attempt to calm public anxiety while the tech industry aggressively integrates AI into workplaces and software ecosystems. Microsoft repeatedly emphasizes the need for governance, safeguards, monitoring, and operational controls around AI systems, which also happens to align closely with its growing enterprise AI and Azure business.
183401050
submission
BrianFagioli writes:
A new academic benchmark called âoeAllFaithâ claims leading AI models from OpenAI, Anthropic, Google, Meta, and xAI show measurable religious bias and often avoid faith perspectives entirely when responding to ethical questions, grief, and personal struggles. Researchers from Baylor University, Notre Dame, Brigham Young University, and Yeshiva University say models frequently suggest therapists, family members, or teachers for guidance, while rarely recommending pastors, rabbis, imams, or other spiritual leaders, despite survey data showing many users expect religion to be included in these conversations.
The study also examined religious conversion prompts and found what researchers describe as repeatable favoritism toward some belief systems and negative bias toward others. According to the benchmark, Grok showed some of the strongest measurable biases, while Anthropic and Meta models were among the least biased. The consortium says the issue is likely unintentional, stemming from training data and moderation choices rather than deliberate discrimination. Still, the findings raise an uncomfortable question for the AI industry: if chatbots increasingly become humanityâ(TM)s source for emotional support and moral guidance, can they really claim neutrality while largely excluding religion from the discussion?
183387842
submission
BrianFagioli writes:
A new report from the Web3 Foundation claims Big Tech and AI companies could generate as much as $831,497 in inflation-linked lifetime value from a single American internet user. The report argues that modern internet platforms are monetizing far more than targeted ads, with everything from search queries and shopping habits to chatbot prompts, uploaded images, location history, and behavioral data feeding AI systems and recommendation engines. Companies including Google, Meta, Microsoft, Amazon, and Anthropic are specifically mentioned as examples of firms benefiting from large-scale personal data collection.
While the report comes from a Web3 advocacy organization and should be viewed with some skepticism, its core argument may resonate with privacy critics and anti-AI users alike: the internet stopped being âoefreeâ a long time ago. The paper argues that AI has made user data even more valuable because human-generated content is now being used not only for advertising, but also to train increasingly capable machine learning systems. Meanwhile, ordinary users see little transparency, control, or financial participation in the value created from their digital lives.
