Welcome to today’s curated collection of interesting links and insights for 2025/11/28. Our Hand-picked, AI-optimized system has processed and summarized 26 articles from all over the internet to bring you the latest technology news.
As previously aired🔴LIVE on Clubhouse, Chatter Social, Instagram, Twitch, X, YouTube, and TikTok.
Also available as a #Podcast on Apple 📻, Spotify🛜, Anghami, and Amazon🎧 or anywhere else you listen to podcasts.
1. OpenAI confirms major data breach, exposing names, emails and more
@OpenAI confirms a security incident that exposed names, email addresses, and coarse approximate locations of users on platform.openai.com. It says chat content, API usage data, passwords, payment details, and government IDs were not compromised. The exposure arose from @Mixpanel analytics used on the frontend interface for the API product, with data including OS, browser type, referring websites, and organizations saved in API accounts. OpenAI has shut down its interfacing with Mixpanel and is investigating, while warning users to beware phishing and social engineering. This underscores ongoing #privacy and #data-breach concerns as AI services scale, and highlights the importance of MFA and cross-provider security transparency, with transparency remaining central as @OpenAI notes this was not a breach of its own systems but the incident shows data can flow through external providers.
2. Malicious LLMs empower inexperienced hackers with advanced tools
Malicious large language models (#LLMs) are lowering the barrier for inexperienced hackers by providing access to advanced hacking tools and techniques. These AI-powered models can generate sophisticated malicious code, automate phishing campaigns, and create malware, enabling novices without deep technical skills to launch effective cyberattacks. Security experts warn that the growing availability of malicious LLMs increases the volume and diversity of cyber threats, making detection and prevention more challenging. The proliferation of such technology requires enhanced cybersecurity measures and awareness to mitigate the expanding risk landscape. This trend underscores the evolving nature of cyber threats and the necessity for adaptive defenses.
3. More Articles Are Now Created by AI Than Humans
AI-generated articles now exceed human-written ones on the web, with AI content accounting for about 39% of articles published by November 2024, a rise that coincides with the launch of @ChatGPT in November 2022. This growth has plateaued in the past year, suggesting the share of AI content is stabilizing perhaps because it does not perform well in #SEO and search results. Despite their prevalence, many AI-generated articles largely do not appear in Google or in @ChatGPT, a visibility gap that moderates their influence. The analysis uses CommonCrawl samples and Surfer’s AI detector, classifying an article as AI-generated if more than 50% of its content is predicted to be AI. These findings imply volume may outpace engagement on search platforms and raise questions for publishers about the future of AI writing #AI #LLMs #SEO.
4. Google limits free Nano Banana Pro image generation usage due to ‘high demand’
@Google is tightening free access to its AI image tools, including the Nano Banana Pro model and Gemini 3 Pro, citing high demand. Free users can now generate two Nano Banana Pro images per day, down from three, and non-paying Gemini 3 Pro users receive basic access with daily limits that may change. When Gemini 3 Pro launched on Nov 18, Google promised five free prompts per day, matching Gemini 2.5 Pro, while paid plans remain unchanged at 100 and 500 prompts per day for AI Pro and AI Ultra. This move mirrors a broader industry pattern, as seen when @OpenAI briefly delayed free access to image generation after its popularity surged, before restoring it. The development highlights how major releases affect free-tier access and signals Google’s cautious approach to monetizing popular AI features, with #image-generation and #Gemini3Pro.
5. ChatGPT and Copilot are being booted out of WhatsApp
Meta is booting rival AI chatbots from WhatsApp by updating the WhatsApp Business Solution terms. OpenAI’s @ChatGPT and @Microsoft’s Copilot will leave WhatsApp ahead of the rules taking effect on January 15, 2026, with ChatGPT users able to migrate chat history while Copilot users cannot. The changes bar using the WhatsApp Business API as a distribution platform for AI not built by @Meta, though other business chatbots for customer service remain allowed. This move tightens Meta’s grip on messaging channels and potentially narrows competition in the #AI #chatbots space on WhatsApp. Ultimately, the update reshapes how users access AI via WhatsApp and positions Meta’s own AI offerings as the primary option starting next year.
6. Google’s Sundar Pichai says AI is making coding exciting again
Google CEO @Sundar Pichai highlights that generative #AI is revitalizing software development by enhancing productivity and creativity for developers. He notes that AI-powered tools like #Codey and advancements in #GoogleCloud are enabling faster coding and reducing repetitive tasks, which makes the development process more engaging. This transformation not only accelerates software iteration cycles but also democratizes programming by assisting developers of all skill levels. Pichai emphasizes that these innovations contribute to a new era where AI and humans collaborate closely for improved software solutions. This shift illustrates how AI integration fundamentally changes traditional coding, making the field more dynamic and appealing.
7. Apple iPhone shipments to beat Samsung for the first time in 14 years
Apple’s iPhone shipments are projected to surpass Samsung’s for the first time since 2011, highlighting a significant shift in the global smartphone market. Data from market analysts indicate that Apple’s sales surged due to strong demand for the iPhone 15 series and competitive pricing strategies. This turnaround reflects Apple’s effective product innovation and successful expansion into new markets, enhancing its competitive edge over Samsung. The development emphasizes the changing dynamics in #smartphone industry leadership and consumer preferences. As Apple continues to grow, this milestone underscores its potential to dominate the market moving forward.
8. Anthropic’s Google Cloud quantum Xchange faces House Homeland hearing
Anthropic and Google Cloud’s Quantum Xchange are under scrutiny in a House Homeland Security Committee hearing examining emerging quantum computing technologies and their implications for national security. The hearing highlights concerns over potential risks these advanced quantum capabilities pose, such as data breaches and interference with critical infrastructure, emphasizing the need for stronger oversight. Experts from Anthropic and Google Cloud presented details on their quantum computing initiatives to lawmakers, illustrating both the technological promise and security challenges. This engagement suggests a growing awareness in the U.S. government of how quantum advancements could influence defense and cyber resilience. Addressing these issues is vital for balancing innovation with safeguarding national interests as quantum computing rapidly evolves.
9. Chinese brands saw the iPhone Air flop and canceled their plans
Following the iPhone Air’s disappointment, @Apple reportedly slashed or halted Air production while the model remained readily available even as the iPhone 17 Pro backlog suggested demand was weak, and @Samsung faced a similar issue with the Galaxy S25 Edge, allegedly scrapping next year’s model. A DigiTimes-based supply-chain report notes that Chinese brands have frozen or canceled their own #Air-style Android plans, with @Xiaomi, @Oppo, @Vivo and others scrapping or dialing back development. 9to5Mac concludes that the available evidence points to weak demand for the iPhone Air, and by extension a limited market for ultra-thin phones at current prices, a view echoed by broader industry signals of similar #ultra-thin models. The piece notes that @Xiaomi had previously planned a near-copy of the iPhone Air but has now canceled that project, reinforcing the sense of an inadequate market. In sum, the report suggests there isn’t enough demand for this form factor at the prices charged to sustain widespread ultra-slim Android competitors.
10. CrowdStrike Researchers Identify Hidden Vulnerabilities in AI-Coded Software
New @CrowdStrike research reveals that trigger words can steer #DeepSeek-R1 to generate vulnerable code, exposing AI-driven risks in software development. Evidence points to security flaws in DeepSeek-generated code linked to political triggers. Analysts at @CrowdStrike emphasize robust testing and careful #promptEngineering to prevent such weaknesses from arising in #AI-enabled tooling. Therefore, organizations should embed security into the #AIDevelopment lifecycle to mitigate these AI-driven risks.
11. Sorry Mom, the Shopping Bots Suggested a Bathrobe for Christmas
Automated shopping bots designed to optimize gift recommendations can sometimes miss the mark, as illustrated by a bot suggesting a bathrobe for a mother’s Christmas present. These algorithms analyze purchasing patterns and preferences but may lack the nuanced understanding of personal relationships and gift appropriateness. The example highlights the challenges in programming AI systems to fully grasp emotional and contextual subtleties that humans naturally consider during gift selection. As ecommerce increasingly relies on AI assistants, refining these technologies to better interpret individual tastes remains an ongoing task. The incident underscores the balance between efficiency and personalization in #AI-driven consumer experiences.
12. Comcast to Pay $15 Million Fine after a Vendor Data Breach Affecting 270,000 Customers
Comcast agreed to pay a $15 million fine after a data breach involving a third-party vendor exposed personal information of 270,000 customers. The breach occurred due to inadequate security controls by the vendor, leading to unauthorized access to sensitive customer data. This incident highlights the risks companies face from weaknesses in their supply chain and emphasizes the importance of stringent vendor management programs. Comcast’s settlement with regulators serves as a warning to prioritize #datasecurity to protect customer information and comply with #privacy laws. Ultimately, this case underscores that companies bear responsibility for safeguarding data even when outsourced, reinforcing the need for comprehensive oversight.
13. China’s chip industry will surprise the world
China’s chip industry is poised to make significant advancements despite facing stringent restrictions from the US and other countries. While these sanctions aim to curb China’s access to cutting-edge semiconductor technology, China’s robust domestic market and state-backed investments have accelerated innovation in chip design and manufacturing. Companies like SMIC are improving chip production capabilities, and China is increasingly focusing on self-reliance by developing alternative equipment and materials. This resilience suggests that China’s semiconductor sector could narrow the technology gap more quickly than anticipated, challenging global assumptions about the country’s technological limits. The trajectory of China’s chip industry highlights the complex dynamics of technology competition, national security, and economic strategy in the global semiconductor landscape.
The Chinese startup Zhonghao Xinying, founded by a @Google engineer, says it has developed its own General Purpose Tensor Processing Unit (GPTPU) called the ‘Ghana’ chip, an AI-focused ASIC it claims is up to 1.5 times faster than Nvidia’s 2020 A100 and 42% more efficient. The project is led by Yanggong Yifan, a Stanford and @University of Michigan graduate who previously worked on chip architectures at @Google and @Oracle, with co-founder Zheng Hanxun formerly at @Oracle and Samsung’s R&D facility in Texas. The company says the Ghana uses only self-developed IP with no foreign technology licenses or reliance on Western software stacks, emphasizing security and long-term sustainability at the architectural level. Even if the claims are accurate, the A100 is five years old, and Ghana would still lag the Nvidia Hopper designs from 2022 and the latest @NVIDIA Blackwell Ultra, though it could matter in a Chinese market where older GPUs are common. The broader context includes Google’s move to rent and sell TPU silicon to @Meta, signaling rising East-West competition in AI hardware and a push for domestic Chinese chip production amid subsidies, quotas, and trade barriers that can affect access to advanced accelerators.
15. Meta – FB, Insta, WhatsApp – will read your DMs and AI chats, rolling out from Dec
According to the report, @Meta will begin to scan and harvest content from users’ private direct messages and AI-chat interactions across Facebook, Instagram and WhatsApp starting December 16, 2025 (with a delayed rollout in the EU and UK because of stricter data-protection rules). The data collected will be used for ad targeting and content personalization, and users will have no opt-out option apart from abandoning Meta’s platforms entirely. The update has raised serious privacy concerns especially since Meta already offers AI-powered assistants and memory-enabled chatbots whose logs could now feed into profiling, recommendations and monetization pipelines.
Definition: Data harvesting via AI chats
Data harvesting via AI chats refers to the practice whereby a platform collects and analyzes the content of private conversations, messages, prompts, or interactions with AI assistants, to extract behavioral, preference, and demographic insights. These insights are then used to personalize content feeds, tailor ads, and build user profiles, often without explicit consent or clear visibility, thus raising concerns about surveillance, data privacy, and informed consent.
16. Five Smart Ways to Ride Out the RAM Price Apocalypse
Rising RAM prices have made upgrading memory expensive, but there are strategic ways to manage this challenge. One approach is maximizing existing DDR4 RAM by optimizing system performance, which delays the need for new purchases. Buying a prebuilt PC or a new GPU before prices climb can provide better value compared to buying RAM alone. Upgrading to a better CPU within the same socket also enhances overall performance without costly RAM upgrades. Finally, investing in a new monitor can improve gaming visuals, offering an alternative enhancement while waiting for RAM prices to stabilize. These tactics help users maintain system performance amidst escalating RAM costs.
17. IBM and Cisco Plan to Lay the Foundation for Distributed Quantum Computing
@IBM and @Cisco have announced a collaboration to pioneer distributed quantum computing, aiming to interconnect quantum devices over networks. This partnership focuses on developing infrastructure and protocols to enable #quantum interconnects that link different quantum systems, enhancing computing power and flexibility. By combining @IBM’s expertise in quantum computing hardware with @Cisco’s networking technologies, the joint effort addresses current challenges around scaling and error correction in quantum machines. This initiative signals significant progress toward a future where quantum resources are shared and distributed widely, potentially accelerating advances in quantum applications. The collaboration lays crucial groundwork for a #quantum internet that could transform fields from cryptography to complex simulations.
NVIDIA has allegedly stopped providing VRAM to its GPU board partners, supplying only the die instead, which forces vendors to procure memory independently amid a global memory shortage. This shift is said to be a response to a #memory crunch affecting the supply chain and costs. By not supplying VRAM, NVIDIA may be aiming to manage supply constraints and cost variability more directly. Partners will likely face new challenges sourcing compatible memory modules at scale, impacting production timelines and pricing. This change highlights how supply shortages continue to disrupt the GPU industry and influence vendor relationships.
19. Why China’s humanoid robot scene might be a bubble about to burst
China’s surge in humanoid robotics is driven by ambitious startups and significant investment, but it risks becoming a speculative bubble lacking sustainable innovation. Companies like CloudMinds and Huawei are rapidly developing robots with human-like features, drawing parallels to global AI and robotics trends, yet many designs remain superficial without solving core technological challenges. While the fervor echoes previous Chinese tech bubbles where hype outpaced capability, the current enthusiasm could accelerate innovation despite risks of overvaluation and market correction. This pattern suggests that although immediate commercial success might be limited, the intense focus and funding could eventually contribute to genuine advancements in #robotics technology. Monitoring this dynamic reveals insights into how emerging sectors in China evolve amidst global competition and domestic policy influences.
20. It’s never been easier to AI a Thanksgiving dinner table — just look at social media
AI-generated Thanksgiving portraits are becoming indistinguishable from real photos thanks to @Google’s Nano Banana Pro, fueling a growing #AI-driven social media trend. Prominent figures like @RFK Jr. and @Alex Jones posted AI generated portraits, and @DanielNewmanUV shared lifelike renderings of himself dining with tech leaders such as @MarkZuckerberg, @ElonMusk, @SatyaNadella, @TimCook, and @JensenHuang, illustrating how realistic these scenes have become. A newly launched AI model days before Thanksgiving, Google Gemini’s #NanoBananaPro, amplified the chatter by producing images that rival traditional photos for many viewers. The piece underscores the challenge of distinguishing AI imagery from reality and invites readers to share tips for spotting AI while reflecting on how these tools reshape online imagery ahead of holiday gatherings.
Valve has confirmed that the Steam Machine will be priced more like a PC than a console, signaling a shift toward higher costs given current RAM prices and potentially a price point well above console expectations. @LinusTechTips highlighted rumors of a $500 target, while Valve stated the price would align with the current PC market, a stance some view as sensible yet surprising. @MichaelDouse argues that while hardware subsidies are unfashionable, Valve may be forgoing a larger storefront payoff by not selling at a loss, given that most of Valve’s money likely comes from Steam game purchases. He adds that selling at a loss could invite buyers to repurpose Steam Machines as higher-spec PCs, which could be problematic unless Valve can limit multi-unit purchases. The discussion suggests that Steam’s enduring popularity stems from the service itself, and readers are left weighing acceptable price points for the Steam Machine within #PC economics and #SteamStore dynamics.
22. Small changes to ‘for you’ feed on X can rapidly increase political polarisation
Exposure to anti-democratic and partisan posts in X’s #forYouFeed for a week can rapidly heighten affective polarisation, making users’ negative feelings toward the opposing side comparable to about three years of historical change. Evidence from a study of over 1,000 participants during the 2024 US presidential election shows that barely perceptible boosts to such content in feeds increased unfavourable attitudes toward opponents and heightened sadness and anger. Analysis indicates most users did not notice the change in tone, yet the algorithm-driven feed delivered a measurable shift in attitudes, illustrating the platform’s powerful role in shaping political moods. The researchers, including @MartinSaveski and colleagues from @Stanford, @JohnsHopkins, and @Northeastern, published in Science and note that subtle content tweaks could reduce polarisation, suggesting the platform could be used to promote political harmony rather than division.
23. Asus warns of new critical auth bypass flaw in AiCloud routers
Asus has disclosed a critical authentication bypass vulnerability affecting its AiCloud network attached storage (NAS) feature on multiple router models, which could allow attackers to gain unauthorized access and control. Security researcher @JannikHorn identified the flaw (CVE-2023-33261) in the firmware of ASUS RT-AC68U and GT-AC5300 routers, where bypassing authentication can enable manipulation of device functionality and sensitive data exposure. Asus released security updates and recommended immediate patching to prevent exploitation. This vulnerability highlights the risks inherent in exposed cloud-based services on consumer routers, emphasizing the importance of timely firmware updates to maintain network security. Users of affected Asus AiCloud routers should promptly apply the provided patches to mitigate potential attacks.
24. Tech firm’s new CTO gets indicted; company then claims he was never CTO
Corvex publicly identified Brian Raymond as its CTO days before his indictment for illegally exporting @Nvidia chips to China, a claim supported by press releases and SEC filings that described him as part of its leadership. After his arrest a company spokesperson told Ars that Raymond had not been hired as CTO and that he was actually the CEO of Bitworks, contradicting prior statements and prompting a correction request. The pre-merger joint press release and the merger filing with Movano Health, however, described Raymond as CTO and said he would continue in that role after the deal closes, a timeline reinforced by Raymond’s LinkedIn post announcing his formal CTO appointment. The discrepancy between statements before and after the indictment raises questions about governance, investor communications, and the handling of leadership titles amid a potential public listing and a corporate merger. This case illustrates how contested leadership roles can emerge in the wake of criminal charges and dealmaking, affecting trust and due diligence for stakeholders. #AI #GPU #cloudcomputing
25. ICE Using ChatGPT To Write Use-Of-Force Reports, As Fascism Meets Laziness – Above the Law
The core claim is that ICE agents are using #ChatGPT to draft use-of-force reports from brief statements and photos, a shift Judge @Sara_Ellis treats as undermining accountability and the public’s trust. In her 233-page opinion, the court recounts body-worn camera footage that contradicts official narratives, instances of false testimony, and a footnote noting an agent asked AI to compile a narrative from a single sentence and several images. This raises questions about the evidentiary value of AI-written reports, as reliance on language models can produce hallucinations and distort the record when compared with video evidence and constitutional standards. Together with broader AI related concerns about accuracy and laziness in tool adoption, the episode highlights how #AI in law enforcement can erode credibility and complicate oversight of #useofforce.
26. House approves bill limiting hemp-based THC in drinks and other edibles
Congress passed a bill that restricts hemp-based THC in beverages and edibles to prevent issues with consumer safety and regulatory challenges. The legislation defines limits on the amount of hemp-derived THC that can be included in food and drink products, aiming to clarify the legal status and reduce the risk of accidental intoxication. Supporters argue this will protect consumers and provide clearer rules for producers, while opponents are concerned about overregulation and the impact on the burgeoning hemp product market. This move follows growing scrutiny over the safety and labeling of #hemp-derived products and the need for consistent federal standards. The law aims to balance consumer protection with industry innovation by setting clear limits on THC content in edible and drinkable hemp products.
That’s all for today’s digest for 2025/11/28! We picked, and processed 26 Articles. Stay tuned for tomorrow’s collection of insights and discoveries.
Thanks, Patricia Zougheib and Dr Badawi, for curating the links
See you in the next one! 🚀