184140514
submission
BrianFagioli writes:
The latest Digital Accessibility Index found that interior pages on websites contain more accessibility issues than homepages, and that AI search tools are increasingly sending users directly to those deeper pages instead of routing traffic through a siteâ(TM)s front door.
The report examined more than 165,000 pages across 6,100 websites in the US and Europe and found that product pages, support articles, account portals, and checkout flows often receive less accessibility attention despite being where users actually interact with businesses.
As AI search changes how people discover information online, organizations may need to rethink accessibility strategies that focus primarily on homepages.
184115026
submission
BrianFagioli writes:
Newegg has launched a conversational AI shopping experience that allows customers to describe the PC they want in natural language and refine the build through an ongoing chat session.
The system can adjust budgets, swap components, answer compatibility questions, and move shoppers from recommendations to checkout without leaving the conversation. Newegg says the assistant pulls from live pricing, inventory, deals, Open Box products, and combo offers, with future plans to integrate directly with the companyâ(TM)s PC Builder tools.
Newegg has also launched a ChatGPT app that brings its product catalog directly into ChatGPT conversations.
The move raises an interesting question: do PC enthusiasts actually want AI helping them shop? Many builders enjoy researching parts, comparing benchmarks, and obsessing over specifications themselves. For some buyers, AI guidance could make the process easier. For others, it may feel like outsourcing part of the hobby.
184115000
submission
BrianFagioli writes:
Newegg has launched a conversational AI shopping experience that allows customers to describe the PC they want in natural language and refine the build through an ongoing chat session.
The system can adjust budgets, swap components, answer compatibility questions, and move shoppers from recommendations to checkout without leaving the conversation. Newegg says the assistant pulls from live pricing, inventory, deals, Open Box products, and combo offers, with future plans to integrate directly with the companyâ(TM)s PC Builder tools.
Newegg has also launched a ChatGPT app that brings its product catalog directly into ChatGPT conversations.
The move raises an interesting question: do PC enthusiasts actually want AI helping them shop? Many builders enjoy researching parts, comparing benchmarks, and obsessing over specifications themselves. For some buyers, AI guidance could make the process easier. For others, it may feel like outsourcing part of the hobby.
184093264
submission
BrianFagioli writes:
The Linux Foundation has announced Akrites, a new initiative to coordinate vulnerability disclosure and remediation for critical open source software as AI dramatically speeds up vulnerability discovery. Founding members include AWS, Google, Microsoft, OpenAI, Red Hat, NVIDIA, IBM, Cisco, JPMorganChase, and others. Akrites will provide a shared Security Incident Response Team (SIRT), a standardized coordinated vulnerability disclosure process, and act as a âoemaintainer of last resortâ for abandoned but widely used packages. The goal is to reduce duplicate reports, avoid conflicting patches, and help upstream maintainers address vulnerabilities before they can be exploited. As AI makes it easier to find security flaws, can a coordinated industry effort help protect open source, or does it risk giving large corporations too much influence over the ecosystem?
184093226
submission
BrianFagioli writes:
The Linux Foundation has announced Akrites, a new initiative aimed at coordinating vulnerability disclosure and remediation for critical open source software as AI dramatically accelerates vulnerability discovery. Founding members include Amazon Web Services, Anthropic, Cisco, Google, IBM, Microsoft, NVIDIA, OpenAI, Red Hat, JPMorganChase, and others.
Rather than multiple companies independently reporting the same vulnerabilities to maintainers, Akrites proposes a shared Security Incident Response Team (SIRT) and a standardized coordinated vulnerability disclosure process. The initiative also says it will act as a âoemaintainer of last resortâ for abandoned but widely used open source projects.
While the effort has broad industry backing, its long-term success will likely depend on whether independent maintainers view it as a helpful partner rather than another layer of corporate oversight.
184040674
submission
BrianFagioli writes:
Samsung has announced UFS 5.0, a new mobile storage standard that it says delivers sequential read speeds of up to 10.8GB/s and write speeds of up to 9.5GB/s. According to the company, the new storage solution is designed to support increasingly demanding on-device AI workloads, where smartphones and other portable devices process large language models locally rather than relying entirely on cloud services.
Samsung also says UFS 5.0 improves power efficiency by more than 40 percent compared to its previous UFS 4.1 solution while reducing the physical package size by 16.7 percent. The company believes the faster, more efficient storage will help future smartphones, wearables, and XR devices deliver quicker AI responses and lower latency.
Mass production is scheduled to begin in the fourth quarter, with capacities reaching up to 1TB.
184037596
submission
BrianFagioli writes:
Florida International University researchers have developed a technique called JaiLIP (Jailbreaking with Loss-guided Image Perturbation) that uses subtle image modifications to bypass AI safety guardrails. Unlike traditional jailbreaks that rely on carefully crafted prompts, the attack works through images that appear normal to human viewers.
The researchers tested the technique against BLIP-2, a multimodal AI model, and found that manipulated images significantly increased the likelihood of harmful responses. According to the study, the approach outperformed previous image-based jailbreak methods and nearly doubled the number of unsafe outputs generated during testing.
The findings highlight a potential security risk for businesses deploying AI systems that process both images and text. While most discussions about AI safety focus on prompts, the research suggests that seemingly harmless images may also serve as an attack vector.
184030596
submission
BrianFagioli writes:
Cloudflare has announced a new initiative with Mozilla Firefox, Google Chrome, Microsoft Edge, and Shopify to develop a privacy-focused protocol called Private Access Control Tokens (PACT). The goal is to help websites distinguish legitimate users and authorized AI agents from abusive automated traffic without relying on CAPTCHAs, invasive tracking, or browser fingerprinting.
PACT would allow trusted services to issue anonymous tokens that browsers can present to other websites as proof that a human is involved, while avoiding the disclosure of personal identity information or browsing history. The companies plan to submit the protocol for standardization.
Cloudflare argues that existing anti-bot tools are becoming less effective as AI-powered agents become more common across the web.
184022758
submission
BrianFagioli writes:
While asking ChatGPT for advice on grilling a tomahawk steak, a curious statement appeared in the conversation. After learning that a handheld digital thermometer was being used, ChatGPT replied that it was âoeactually what I use too.â The statement was harmless, but it raised an interesting question: why would an AI imply personal experience when it has none? The resulting discussion explores how conversational AI models can sometimes use language that sounds personal despite having no real-world experiences, preferences, or possessions.
183986310
submission
BrianFagioli writes:
OpenAI introduced LifeSciBench, a new benchmark designed to evaluate AI systems on realistic life science research tasks rather than simple biology questions. While OpenAIâ(TM)s top-performing GPT-Rosalind model led the rankings, it achieved a pass rate of just 36.1 percent, failing nearly two-thirds of benchmark tasks. The company says the results highlight progress in scientific communication and evidence synthesis, but also reveal persistent weaknesses in artifact-heavy and design-oriented scientific work.
183963324
submission
BrianFagioli writes:
Security researchers are warning that attackers are increasingly targeting AI agents instead of the people using them. A new report from OrcaRouter argues that prompt injection is becoming the phishing attack of the AI era, allowing attackers to manipulate AI systems through emails, documents, websites, and other content they consume.
As AI agents gain access to email, source code, business documents, and other systems, researchers say organizations need to start protecting the agents themselves, not just the humans they serve.
183950788
submission
BrianFagioli writes:
SK hynix has shipped samples of its new 12-layer HBM4E memory to major customers. The company says the next-generation AI memory delivers speeds of up to 16Gbps per pin, more than 20 percent better power efficiency, and 17 percent improved heat resistance compared to HBM4. As AI workloads continue to grow, high-bandwidth memory is becoming increasingly important for reducing bottlenecks in training and inference systems.
183944636
submission
BrianFagioli writes:
Google, Microsoft, OpenAI, Arm, Mastercard, Siemens, and other companies have joined the newly launched Appia Foundation under the Linux Foundation. The project aims to create common specifications and assessment frameworks that organizations can use to demonstrate AI systems meet emerging safety, trust, and compliance requirements. According to the Linux Foundation, the framework is designed to allow conformity evidence to be reused across the AI supply chain, potentially reducing duplicate assessments and compliance costs. The announcement comes as governments around the world move toward enforcing AI regulations and organizations face increasing pressure to prove AI systems are trustworthy.
183944624
submission
BrianFagioli writes:
Google, Microsoft, OpenAI, Arm, Mastercard, Siemens, and other companies have joined the newly launched Appia Foundation under the Linux Foundation. The project aims to create common specifications and assessment frameworks that organizations can use to demonstrate AI systems meet emerging safety, trust, and compliance requirements. According to the Linux Foundation, the framework is designed to allow conformity evidence to be reused across the AI supply chain, potentially reducing duplicate assessments and compliance costs. The announcement comes as governments around the world move toward enforcing AI regulations and organizations face increasing pressure to prove AI systems are trustworthy.
183930944
submission
BrianFagioli writes:
Proton has introduced Easy Switch for Business, a migration tool designed to help organizations move email, calendars, contacts, domains, and user accounts from Google Workspace to Proton Mail with minimal disruption. The company says businesses can run Google Workspace and Proton in parallel during the transition, allowing IT teams to test the environment before fully switching over.
While Proton pitches the feature as a response to growing concerns about privacy, compliance, data sovereignty, and dependence on large US technology providers, Google Workspace remains deeply entrenched in many organizations. Proton currently supports migration of email, calendars, and contacts, though Google Drive migration is not yet available.
