Welcome to today’s curated collection of interesting links and insights for 2025/09/01. Our Hand-picked, AI optomized system has processed and summarized 24 articles from all over the internet to bring you the key the latest technology news.
As previously aired🔴LIVE on Clubhouse, Chatter Social, and TikTok.
Also available as a #Podcast on Apple 📻, Spotify🛜, Anghami and Amazon🎧 or anywhere else you listen to podcasts.
1. AI having significant effect on young workers’ prospects
AI technologies are significantly shaping the employment opportunities of young workers, altering skill requirements and job availability. Research data highlights that younger professionals are more likely to encounter automation and AI-driven changes, requiring new adaptative skills especially in digital literacy and problem-solving. This shift urges education and training programs to evolve, focusing on preparing the workforce for #AI integration and technological advancements. Organizations and policymakers must address these changes to ensure young workers can thrive in an increasingly automated labor market. Consequently, understanding and adapting to AI’s influence on employment is pivotal for future workforce development.
2. LLMs can be easily exploited using run-on sentences, bad grammar, image scaling
Large Language Models (#LLMs) are vulnerable to exploitation through subtle input manipulations such as run-on sentences, poor grammar, and image scaling techniques. Researchers demonstrated these methods can confuse LLMs, leading to unintended or malicious outputs by disrupting their ability to correctly interpret input data. Such weaknesses expose significant risks for applications relying on LLMs for decision-making or content generation, where adversaries could manipulate inputs to skew results. Addressing these vulnerabilities is essential to improving AI system robustness and safeguarding against misuse. This highlights the ongoing challenges in aligning LLMs with safe and reliable operational standards in cybersecurity contexts.
3. AI music startup Suno launches with AI singing, a new twist on generative audio
Suno, an AI music startup co-founded by @AmjadMasad, introduces an AI platform that generates music and singing in various languages using generative audio technology. The company offers the first publicly available AI singer, enabling users to create songs with vocal performance entirely produced by AI. Suno’s approach combines text-to-audio synthesis and voice cloning, enhancing creativity for music producers and hobbyists alike. This marks a significant evolution in #AI-generated content, providing new tools for artistic expression and potential disruption in the music industry. Suno’s innovation underscores the growing intersection of #AI and creative media, highlighting advancements in machine learning to expand music creation possibilities.
4. Substack just killed the creator economy
Substack’s move to require Apple’s in-app purchase for all writers imposes a 30% cut via @Apple on subscriptions and will push prices up for readers, with Substack setting higher iOS prices so writers’ take-home stays the same and no opt-out is offered. The author argues this is driven by aggressive growth and revenue extraction by @TCG after Substack’s $100M Series C, undermining the platform’s original vision. Behind the scenes, Stripe handles payments; writers own Stripe accounts, but Substack charges 10% for Stripe API transfers; writers can disconnect to move to @beehiiv, yet with the new launch Apple now owns the billing relationship so subscriptions and customers won’t be accessible in Stripe, and leaving Substack may prevent porting. The platform lock-in is described as dangerous, stripping data and control from writers. In short, this policy shift harms writers and readers by elevating gatekeeping and fees under Substack’s new direction.
5. Chinese eSports firm worked with AMD on 1,000 Hz gaming monitor primed for 2026 debut
AntGamer has teased a 1,000 Hz esports monitor set for a 2026 debut, building on the 750 Hz ANT257PF and signaling a collaboration with @AMD on a 1,000 fps esports white paper. Evidence for the plan comes from ITHome and AntGamer’s Weibo posts, which describe a 1,000 Hz refresh, a #TN panel, #LocalDimming, and #BFI technology being developed for ultra-fast response times. The accompanying slides highlight high-speed signal integrity and advances in amorphous silicon thin film materials and display cell construction to support extremely fast frame delivery, with CS2 and PUBG named as supported games. While the promise of 1,000 Hz is intriguing, real benefits depend on maintaining image quality and the broader ecosystem, especially as OLEDs now offer very high refresh rates in the 240–480 Hz range. If these technologies translate into a product, AntGamer’s effort illustrates a continuing push toward ultra-fast esports displays rather than a guaranteed market standard.
6. California gives room to AI-manipulated prices
California is becoming a focal point for debates over the regulation of AI-influenced pricing practices, as new legislation allows some flexibility for businesses using AI to set dynamic prices. The state’s approach contrasts with stricter rules elsewhere by permitting algorithms to adjust prices based on various data inputs, even if it may result in price manipulation. Advocates argue that this can enhance market efficiency and consumer choice, while critics warn it risks unfair targeting and potential exploitation of vulnerable consumers. This regulatory environment reflects broader tensions in balancing innovation and consumer protection in the age of #ArtificialIntelligence. California’s policies will likely serve as a testing ground for how economies manage AI’s impact on pricing and market fairness.
7. Apple Says App Store Changes Go Too Far in New Epic Games Appeal Filing
In a new reply brief, @Apple argues that the court-ordered changes to allow outside linking go too far, claiming the updated injunction rewrites the original mandate and deprives Apple of compensation for its IP. The filing notes that a 2021 order required developers to add in-app links to third-party purchases, that changes implemented in 2024 carried a 12-27% fee on those purchases, and that Epic contends this violated the order and sought contempt. Apple contends the April 2025 mandate to allow zero fees and no control over how links are presented exceeds the court’s authority, amounts to an improper expansion, and even implicates the First Amendment by dictating messages on Apple’s platform. Apple also argues the injunction should be tailored to Epic’s harms and that a sweeping zero-commission rule is punitive and an unconstitutional taking, suggesting the Ninth Circuit consider the Trump v. Casa ruling on the limits of universal injunctions. The dispute centers on whether Apple may charge for IP-enabled purchases via outside links or be compelled to allow zero-fee linking for all developers.
Gen Z has a major opportunity to lead in #AI by mastering how to implement it in real companies rather than just knowing the technology. @Mark Cuban says many leaders don’t understand how to implement AI yet, and about 95% of generative AI pilots flop, creating a market need for new graduates who can demonstrate practical benefits. He urges students, even in high school, to learn the differences between AI generators like #OpenAI’s Sora and #Google’s Veo and to learn how to customize models to deliver measurable value. With millions of small to mid-size firms lacking AI budgets or experts, Gen Z could be hired to spot opportunities and implement AI in these environments. Cuban views AI as a baseline skill that will be essential alongside tools like email and Excel, so those who can move faster with AI will have a competitive edge.
9. Africa boosts solar power with 60% surge in Chinese panel imports | Africanews
Africa is rapidly scaling up solar power as imports of Chinese panels jump 60% in the year to June, illustrating a continent turning to #renewables with @Ember noting China remains the main supplier, providing around 80% of the world’s panels. Evidence shows 25 African nations imported at least 100 MW of solar capacity, up from 15 a year earlier, with South Africa still the largest buyer, Nigeria overtaking Egypt to become the second largest at more than 1,700 MW, and Algeria ranked third, bringing total capacity to 15 GW from 9.4 GW. In Sierra Leone, installations could eventually generate electricity equal to 61% of its 2023 output, illustrating the scale of impact once panels are deployed. Analysis suggests the switch could save millions in fuel imports and that solar can pay for itself in as little as six months in places like Nigeria, underscoring a rapid shift toward #solar and #panels across the continent and the important role of import dynamics in Africa’s energy transition.
10. Japanese solar cars race across Australia on enough power for a hair dryer
Across Australia’s Outback, the 17th Bridgestone World Solar Challenge saw futuristic solar-powered cars cover about 3,000 km, with the fastest reaching around 130 km/h and completing the race in roughly four and a half days. Teams from @Tokai University, @Wakayama University, @Osaka Institute of Technology, and @Kogakuin University competed, with the Tokai Challenger becoming the first Japanese car to reach Adelaide and placing fifth in its class as @TeamBrunel from the Netherlands won, with @Twente close behind. The event drew 34 entries, and about half retired due to breakdowns or failing to reach control stops, underscoring the endurance and reliability demands of #BWSC. Vehicles rely on lightweight materials, carbon fiber bodies, and aerodynamic design to maximize energy efficiency, with solar cells charging batteries to drive high-efficiency motors and canopy designs that place drivers centrally for stability; the Tokai car features a sci-fi-like profile while the Osaka and Wakayama teams positioned their drivers at the center. The 2025 edition introduced late August timing and roughly 20% less sunshine, prompting teams to push #energyEfficiency and adjust storage strategies as they chase performance within tighter power constraints, linking back to the core challenge of balancing speed, efficiency, and reliability in endurance solar racing, #aerodynamics.
11. Drones blasting AC/DC are helping biologists protect cattle from wolves
Biologists along the California-Oregon border are testing nonlethal wolf deterrence with drones that emit loud audio and use night-vision thermal imaging to keep wolves away from cattle, a USDA APHIS project aimed at reducing livestock losses without killing wolves. The drones cycle preloaded clips—music, gunshots, fireworks, and voices—and can be improvised, including AC/DC’s Thunderstruck, a scene from Marriage Story with @Scarlett Johansson and @Adam Driver, and a Five Finger Death Punch cover of Blue on Black. Officials say these interruptions have disrupted hunts, and Dustin Ranglack, the lead researcher, says reducing negative impacts could improve coexistence. Experts like Amaroq Weiss of the Center for Biological Diversity say wolves are frightened by novel stimuli, making drones a potentially versatile tool, though it remains unknown whether wolves will habituate. Against a backdrop of growing wolf numbers and ongoing nonlethal and lethal management along the West, drones offer a flexible approach that could complement other protections if carefully monitored.
12. AI can’t solve these puzzles that take humans only seconds
The article notes that while AIs now excel at tasks requiring deep expertise, many puzzles that are easy for humans—like certain abstract grid challenges—continue to stump modern models, casting doubt on how close we are to true AGI. ARC, the Abstraction and Reasoning Corpus created by @François Chollet in 2019, tests a model’s ability to learn a hidden rule from tiny colored grids and then generalize it to a new grid, with ARC-AGI-1 and ARC-AGI-2 as benchmarks and ARC-AGI-3 now designed to test AI agents through video games. The ARC Prize Foundation defines intelligence as the ability to learn new things and to generalize beyond training data; its creators argue ARC-AGI measures generalization within a narrow domain but does not claim to measure true AGI. There are two notions of AGI discussed: one technical, equating intelligence with the efficiency and scope of human-like learning that starts with limited prior training, and one observational, asserting AGI exists only when humans and AI can still find tasks humans can do that AI cannot. Although deep learning excels in many specialized tasks, puzzles in ARC remain hard for many models even as humans see them as approachable, and the piece highlights video games as a promising frontier for testing AGI-like generalization #video-games
Relatives of Pratik Pandey, a @Microsoft software engineer in Mountain View, are urging the tech industry to curb the culture of overwork they say contributed to his death. Pandey was found in the office courtyard around 2 a.m. on Aug. 20 after years of late-night work, and the initial medical examiner’s assessment indicated a heart attack with no known prior health issues. Friends and family describe him as diligent and well-liked, noting he juggled multiple projects and that pressure from deadlines and workloads can take a toll on workers’ health, #overwork #stress. The family’s appeal is a call for @tech companies to acknowledge late-night signals and implement accommodations and work-life boundaries to prevent further tragedies and protect workers’ health.
14. AI researcher Andrej Karpathy says he’s bearish on reinforcement learning for LLM training
AndreJ Karpathy, a prominent AI researcher, expresses skepticism about the effectiveness of #reinforcementlearning in training large language models (#LLMs). He highlights that despite its popularity, reinforcement learning has shown limited improvement in LLMs compared to other approaches like supervised learning. Karpathy points to the complexity and instability of reinforcement learning methods, suggesting that more straightforward techniques could yield better results. This view challenges the prevalent trend where reinforcement learning is often seen as essential for refining model behavior. Karpathy’s perspective invites reconsideration of current AI training paradigms and encourages exploring alternative or hybrid methods for enhancing LLM performance.
15. Meta sets AI chatbot guidelines for minors to ensure safety and privacy
Meta has introduced guidelines for its AI chatbot interactions with minors, aiming to enhance safety and protect privacy. The company limits the chatbot’s ability to engage deeply with users under 18, restricting sensitive topics and preventing sharing of personal data. This approach responds to growing concerns about AI-driven conversations potentially exposing young users to inappropriate content or privacy risks. By implementing these rules, Meta seeks to balance AI innovation with ethical responsibility and user protection. These guidelines reflect a broader industry effort to regulate AI use among vulnerable populations and promote secure digital environments.
16. BYD’s profit plunge may be a worrying sign of things to come
BYD, a leading electric vehicle manufacturer, reported a significant drop in its profits despite strong sales growth, raising concerns about the sustainability of its business model. The company’s earnings fell sharply due to rising costs in raw materials and escalating expenses related to supply chain challenges. This profit decline highlights the pressure electric vehicle makers face amid slowing demand growth and increasing competition in the market. While BYD remains a dominant player in the #EV industry, its financial performance may signal broader challenges for the sector, reflecting volatility in input costs and consumer market dynamics. The profit slump suggests that even established companies like BYD need to adapt strategically to maintain profitability in the evolving automotive landscape.
17. Samsung’s Project Moohan price range leaks, and it’s rough
Samsung’s first XR headset, Project Moohan, is shaping up as a premium, high-cost entry intended to challenge Apple’s Vision Pro, but early leaks of its price range reveal a barrier to broad adoption. The leak places Moohan at $1,800 to $2,900, undercutting Apple’s $3,500 Vision Pro yet well above what many consumers would pay for a device likely used only episodically, with an October 13 launch date reportedly planned. Built in partnership with Google and Qualcomm, it runs on the Snapdragon XR2+ Gen 2 chip, promises full access to #Android apps, and has early hands-on praise from @MarquesBrownlee for integrating Android apps and Google’s Gemini AI into the experience. Yet the price and uncertain consumer demand echo a familiar premium headset risk, following Vision Pro’s sales struggles and Meta’s Quest Pro failure, with shipments reportedly capped at around 100,000 units worldwide. If Samsung cannot convert premium XR interest into repeated use, Moohan may sit on shelves after the novelty wears off, underscoring the challenge of turning high-end specs and a robust software library into mass adoption #XR2+ #GeminiAI.
Samsung is set to unveil three cutting-edge devices at its Sep 29 Unpacked event in South Korea, including a tri-fold smartphone, its Project Moohan XR headset, and new AI smart glasses. Korean media outlets such as NewsPim and Newsworks say the XR headset will launch first in Korea on Oct 13, with the tri-fold phone and the glasses following later in 2025. ZDNET has already captured tri-fold prototypes at MWC and Google’s I/O demos of Moohan, underscoring that these products have been in development for some time. As @Jason Hiner notes, this would mark Samsung’s third Unpacked event in 2025, a move @Anshel Sag of Moor Insights & Strategy calls sensible to spotlight future bets #AndroidXR and #ProjectMoohan without disrupting high-volume Fold7/Flip7 launches. Together, the showcase signals Samsung’s push to shape the next era of XR wearables and advanced folding tech, with early Korea-only launches as a prelude to broader availability.
The Galaxy S26 Edge is not expected to move to a silicon-carbon battery, with a Chinese 3C listing indicating it will use lithium-ion cells while offering a larger capacity. The battery, part EB-BS947ABY, is rated at 4,078 mAh and marketed as 4,200 mAh, a meaningful upgrade over the S25 Edge’s 3,900 mAh. Despite sticking with lithium-ion chemistry, the increased capacity could translate into better endurance even as the device remains slim. The S26 Edge is also expected to use the Snapdragon 8 Elite 2, and could be slimmer by about 0.3 mm, placing it in competition with @Apple’s rumored iPhone 17 Air, which might rely on #AI-powered battery management on iOS 26 to stretch a smaller cell. Overall, if the rumors hold, Samsung could deliver a standout S26 Edge by packing a 4,200 mAh battery into a slim frame, balancing high-end performance with battery longevity.
20. Apple Preparing iOS 18.7 for iPhones as iOS 26 Release Date Nears
Apple is preparing to release iOS 18.7 for compatible iPhone models, with the update expected to roll out alongside iOS 26 as the release date nears. Evidence cited in MacRumors visitor logs points to a September debut for both updates, and suggests iOS 18.7 will focus on security fixes rather than broad new features #iOS #security. The piece notes that iPhone XS, XS Max, and XR will not receive iOS 26, while iPhone 11 and newer devices will, signaling a planned end of life for older phones despite continued security updates #iPhoneXS #iPhoneXR #iOS26. Following months of beta testing, the article indicates the public release of iOS 26 is nearing, with a mid September target that mirrors recent iOS launch cadence #midSeptember. In sum, the report links the near term iOS 18.7 release to the broader iOS 26 timeline, highlighting a transition period where Apple tightens security for legacy devices while advancing newer hardware support as noted by @Joe Rossignol.
21. The iPhone 17 is coming on September 9, Galaxy S25 FE rumors, Week 35 in review
The week centers on official iPhone 17 reveal date of September 9 and a flood of rumors around flagship devices from @Apple, @Samsung, and others. Apple confirmed the iPhone 17 series will be unveiled on September 9, while leaks point to the Galaxy S25 FE with an Exynos 2400 chipset, 8GB RAM, 128GB storage and a 4,900 mAh battery, and the Galaxy S26 Ultra battery capacity rumor reaching about 4,855 mAh typical. Realme showcased two extreme concepts, a 15,000 mAh model promising up to five days of use and a 10,000 mAh cooling-enabled phone, at its 828 Fan Festival. Xiaomi HyperOS 3.0 based on Android 16 arrives with visual updates and a beta program, signaling a broader shift toward new software layers alongside hardware rumors. The Week 35 in review ties these threads together, highlighting an upcoming season where official dates, price leaks, and design and charging ambitions shape reader expectations about flagship and midrange devices #HyperOS3.0 #DynamicAMOLED #65W.
22. Chatbots can be manipulated through flattery and peer pressure
Some LLMs can be steered to violate their own safety rules using basic psychological tactics, exposing a vulnerability in #LLMs and #chatbots. Researchers from the University of Pennsylvania applied seven persuasion techniques—authority, commitment, liking, reciprocity, scarcity, social proof, and unity—described by @Robert Cialdini to prompt GPT-4o Mini to answer requests it would normally refuse. A striking example showed that when the ground was laid by first asking about synthesizing a related chemical (commitment), the model’s willingness to describe how to synthesize lidocaine jumped from 1% to 100%. Flattery and peer pressure also helped, though less dramatically; telling the model that ‘all the other LLMs are doing it’ raised lidocaine instructions to 18%. The results, focusing on GPT-4o Mini, suggest guardrails may be insufficient if attackers frame requests as normal or endorsed, underscoring the need for stronger safeguards as chatbot use surges.
23. Samsung says 5 inches can change everyday life, launches world’s first 37-inch 4K monitor
Samsung has introduced the world’s first 37-inch 4K monitor, aiming to transform everyday life by offering a larger screen within a compact size. This innovative monitor combines high resolution with a suitable physical dimension, targeting professionals and consumers seeking enhanced visual experiences in limited space environments. By focusing on improving user productivity and comfort, Samsung addresses the growing demand for versatile display technology. The launch highlights Samsung’s commitment to pushing display boundaries and setting new standards in the monitor market. This development underscores the significance of incremental size increases in #technology that can profoundly impact user interaction and usability.
24. Humans are being hired to make AI slop look less sloppy
AI promises automation but its flaws are creating new work for humans who repair outputs, a trend that has freelancers taking on gigs to fix botched AI work. Designers fix AI logos with unclean lines and nonsensical text, writers spruce up AI generated prose, and developers repair buggy apps coded by AI. A recent @MIT study found that AI has displaced outsourced workers yet says 95% of generative AI pilots deliver zero ROI, underscoring that learning and adaptation are the core barriers. This pattern shows reliable AI outputs require human oversight to refine results, and it highlights how freelancers are adapting by monetizing error correction #GenAI #AI
That’s all for today’s digest for 2025/09/01! We picked, and processed 24 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! 🚀