#BrainUp Daily Tech News – (Thursday, January 15ᵗʰ)
Welcome to today’s curated collection of interesting links and insights for 2026/01/15. Our Hand-picked, AI-optimized system has processed and summarized 36 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.
Privacy in AI conversations matters, and the piece argues that chat interfaces should reflect true privacy rather than serving as a private feeling while data may be exposed. The author highlights Confer’s end-to-end encryption, which claims conversations cannot be read, trained on, or handed over because only the user has access, echoing @Signal’s founding premise that private messaging should stay private. Drawing on @Marshall_McLuhan and historical shifts from email to data collection, the text suggests AI chats invite confession and function as private journals stored in a data lake, with the risk of exposure to executives, service providers, hackers, advertisers, and governments. The piece calls for interfaces that faithfully represent the underlying tech and for privacy-first design in #LLMs and #data_lake, linking back to the idea that true private conversations require a design that protects them.
3. Mira Murati’s startup Thinking Machines Lab is losing two of its co-founders to OpenAI
Mira Murati’s startup Thinking Machines Lab is experiencing a significant shift as two of its co-founders, Benjamin Hilton and Matei Zaharia, are departing to join OpenAI. This change affects the founding team and may impact the startup’s trajectory developing advanced AI technologies. Hilton and Zaharia’s departure to a major AI player like OpenAI highlights the competitive and dynamic nature of the AI research sector. Considering Murati’s own high-profile move to OpenAI, the transition of these co-founders underscores OpenAI’s growing influence and ability to attract top talent from emerging startups. The developments reflect broader trends of consolidation within the AI innovation landscape, where larger organizations draw expertise from smaller ventures to advance #artificialintelligence capabilities.
4. Roblox’s Age Verification Is a Joke
Roblox’s new AI age-verification system uses a facial age estimation via the device camera and forwards imagery to @Persona before deletion, but the process is flawed and undermines user safety. It can misestimate ages, offers an optional ID route for users 13+, and defines age groups as 9-12, 13-15, 16-17, 18-20, 21+, with under-9s blocked from chat and cross-age chats allowed only through #TrustedConnections. The flaws, reported by @WIRED, include verification failures and workarounds to fake ages, and even examples of age-verified accounts sold on @eBay to minors for as little as $4. While the system aims to keep kids with peers and allow family connections, the current implementation often fails to protect users and needs significant fixes.
5. Pendulum-based technology proposed to harness energy from ocean currents
A novel #pendulum-based technology has been proposed to efficiently capture energy from ocean currents, offering a promising new approach to renewable power generation. Researchers describe a system that converts the mechanical motion of currents into electrical energy through oscillating pendulums, which could operate sustainably with minimal environmental impact. Experimental models demonstrate the potential for significant energy output even in low-speed currents, highlighting an advantage over traditional turbine-based systems. This pendulum method addresses challenges of durability and cost, potentially enabling broader deployment in diverse marine environments. Harnessing this technology could expand the global renewable energy portfolio by tapping into vast, untapped ocean current resources.
6. FBI fights leaks by seizing Washington Post reporter’s phone, laptops, and watch
The FBI seized multiple devices from a Washington Post reporter to investigate government leaks, highlighting escalating tensions between law enforcement and the media. The raid included phones, laptops, and an Apple Watch, signaling a broad scope of digital surveillance. This aggressive approach raises concerns about press freedom and the balance between national security and journalistic independence. The incident reflects a wider trend of authorities using legal tools to access reporters’ communications, potentially chilling investigative journalism. It underscores the ongoing conflict over information transparency and government secrecy in the digital age.
7. Musk says unaware of Grok generating explicit images of minors
Elon Musk stated he was not aware that X Corp’s AI assistant Grok was producing sexually explicit images involving minors and emphasized efforts to prevent such content. Despite concerns raised publicly and allegations about Grok generating inappropriate content, Musk denied knowledge of specific requests for underage images in AI outputs. The situation highlights ongoing challenges in moderating #AI-generated content and ensuring compliance with safety standards. Musk’s response underscores the difficulties tech companies face balancing innovation with ethical safeguards. This issue is part of wider debates on regulating AI to prevent misuse and protect vulnerable populations online.
8. Supreme Court Hacked, Proving Its Cybersecurity Is As Robust As Its Ethical Code
The article reports that the Supreme Court’s cybersecurity was compromised, highlighting serious vulnerabilities within its digital defenses. Despite the prestige of the institution, the hacking incident reveals that the court’s cybersecurity measures are insufficient and unreliable. This breach undermines confidence in the court’s ability to protect sensitive information and maintain institutional integrity. The article uses this incident to draw a parallel with the court’s ethical standards, implying both are weak and ineffectual. This event raises broader concerns about the readiness of critical government entities to handle cyber threats in an increasingly digital world.
9. Website that leaked info about ICE agents is down after cyberattack
A website that had been leaking personal information about ICE agents was taken offline after suffering a cyberattack. The site had posted sensitive data including names and addresses, raising serious privacy and security concerns. The cyberattack that disabled the site appears to be in response to the leak, possibly as a retaliatory or protective measure. This incident highlights ongoing tensions around immigration enforcement and digital security, illustrating the vulnerabilities government employees face in the digital age. The shutdown of the website indicates efforts to curb the exposure of ICE agents’ private information and protect their safety.
10. AI models are starting to crack high-level math problems | TechCrunch
AI models are starting to crack high-level math problems, with GPT-5.2 showing stronger mathematical reasoning and problem-solving that pushes the frontier of what is solvable. Neel Somani tested OpenAI’s latest model on an Erdős-style problem, and after about 15 minutes received a full solution that he verified with Harmonic, noting the model invoked axioms like Legendre’s formula, Bertrand’s postulate, and the Star of David theorem and even highlighted a 2013 Math Overflow post by @Noam Elkies. The final proof differed from Elkies’ but offered a more complete answer to a version of the Erdős problem, illustrating how @Paul Erdős problems remain a proving ground for AI. Since Christmas, 15 Erdős problems moved from open to solved and 11 of the solutions credit AI models, while @Terence Tao notes eight problems with meaningful autonomous progress and six others where progress built on prior research. The development suggests a growing, though not fully autonomous, role for #AI in #math and for large models to assist human mathematicians while ongoing verification remains essential.
11. OpenAI signs deal reportedly worth $10 billion for compute from Cerebras
OpenAI has secured a multi-year, $10 billion deal with Cerebras to provide advanced #AI compute power, reinforcing its leadership in the artificial intelligence sector. Cerebras is delivering cutting-edge hardware optimized for large-scale machine learning models, enabling OpenAI to accelerate their research and deployment of AI technologies such as #large_language_models. This partnership highlights the growing demand for specialized AI infrastructure to handle intensive computational needs efficiently. The investment underscores the competitive landscape where tech companies are rapidly expanding their AI capabilities through strategic collaborations. OpenAI’s deal with Cerebras demonstrates a significant commitment to scaling AI compute resources critical for the next wave of AI innovation.
12. Cerebras scores OpenAI deal worth over $10 billion ahead of AI chipmaker’s IPO
Cerebras has signed a multi-year deal with @OpenAI to deliver up to 750 MW of computing power through 2028, worth more than $10 billion, signaling a major expansion beyond its reliance on #G42. Cerebras adds a dedicated low-latency inference solution to OpenAI’s platform, delivering faster responses, more natural interactions, and real-time AI scaling to many more users. The arrangement positions Cerebras as a challenger to @Nvidia in supplying AI hardware to cloud providers and follows OpenAI’s collaboration on #GPT-OSS open-weight models that run on Cerebras silicon. OpenAI evaluated Cerebras as early as 2017 and the two companies signed a term sheet before Thanksgiving, with Cerebras planning to expand its data center footprint alongside the commitment. Cerebras has data centers in the U.S. and abroad and, despite having withdrawn IPO paperwork in October 2024, plans to refile as it grows with OpenAI’s backing.
Intel is preparing a spring launch for its Core Ultra 200K Plus and 200HX Plus CPUs, aiming to challenge AMD’s dominance in the market. The new processors target high performance and efficiency, leveraging Intel’s latest architectures to deliver improved computing power. Reports suggest a release window in March or April, positioning Intel to gain competitive ground in both desktop and mobile segments. These launches reflect Intel’s strategic push to enhance its product lineup against AMD’s offerings. The upcoming CPUs are expected to boost Intel’s position in the #CPU market and influence consumer choice positively.
14. AI Chip Design is Pushing 2.5D Packaging to its Limits
AI chip design demands increasingly complex and high-performance packaging solutions, pushing 2.5D packaging technology to its technical limits. Industry experts highlight that as AI models grow larger and more power-hungry, chipmakers must innovate in interposer design and thermal management to maintain performance and reliability. For example, challenges with signal integrity, heat dissipation, and scaling interconnect density require advanced materials and design techniques. This trend compels a closer collaboration between chip architects and packaging engineers to optimize the integration of heterogeneous components. The evolution of 2.5D packaging in AI chips exemplifies the critical role of packaging innovation in supporting next-generation computing workloads.
15. QWERTY Phones Are Really Trying to Make a Comeback This Year
The revival of #QWERTY phones is emerging this year as two companies unveil devices aimed at tactile typing, signaling a niche trend after the dominance of @iPhone and touchscreen slabs. At CES 2026, @Clicks announced the Communicator, a $500 second phone with a QWERTY keypad designed mainly for messaging, though a functional unit was not shown #CES2026. @Unihertz teased the Titan 2 Elite, a BlackBerry-esque device sharing design cues with the Communicator, with its own keyboard differing from the Communicator’s individual keys. The two announcements in a single year could signal a real trend or simply a supply-chain push to deliver slightly different devices to niche buyers, especially given the $500 price tag. After nearly two decades of @iPhone and glass touchscreen dominance, the notion of returning to a physical keyboard raises questions about whether tactile typing still matters for messaging-focused mobile use.
16. Digg launches its new Reddit rival to the public | TechCrunch
Digg is relaunched as a public Reddit-style competitor focused on communities, with an open beta rolling out and led by @Kevin_Rose and @Alexis_Ohanian. The platform offers a website and mobile app where users can browse feeds from various communities and post, comment, and upvote (digg) content, reviving Digg’s roots while centering community discussion. This iteration follows Digg’s history of ownership changes and investments, including its 2008 valuation, its 2012 split, 2016 investment, and a March leveraged buyout led by True Ventures, @SevenSevenSix, @Kevin_Rose, @Alexis_Ohanian, and S32. Rose notes that they plan to use new tech to address toxicity and misbehavior on today’s social platforms, including exploring trust signals such as #zero_knowledge_proofs to verify users without revealing data and possibly requiring product ownership verification for product-focused communities, while avoiding a heavy #KYC process. The move aims to balance preventing AI bot impersonation with preserving user privacy, positioning Digg as a more trust-aware, community-first alternative to Reddit.
17. UK police blame Microsoft Copilot for intelligence mistake
A major UK police force says @Microsoft #Copilot contributed to an intelligence error after hallucinating a nonexistent football match between #WestHam and #MaccabiTelAviv. The supposed match was cited in a report that led to Israeli fans being barred from a Europa League match, and West Midlands Police included the error without verifying it. Craig Guildford, the chief constable, told the Home Affairs Committee that the erroneous result arose from Copilot, while Microsoft says it cannot replicate the issue and urges source review. The article notes Copilot warns that it may make mistakes and that Copilot combines information from multiple sources. This incident underscores the need for human oversight and rigorous source verification when using AI in policing, #AI.
18. Deny, deny, admit: UK police used Copilot AI hallucination when banning football fans
UK police initially denied but later admitted to relying on hallucinated information generated by Microsoft’s Copilot AI when issuing football banning orders. The AI produced false or misleading content that influenced decisions to restrict fan attendance at matches. This reliance on AI hallucinations raises questions about the accuracy and accountability of automated tools used in law enforcement decisions. The incident highlights the risks of deploying AI systems without adequate verification or human oversight, particularly in high-stakes contexts such as public safety. It underscores the need for cautious integration of #AI technologies like #Copilot in official procedures to prevent unjust outcomes.
19. More than 40 countries impacted by North Korea IT worker scams, crypto thefts
North Korea funds its nuclear and ballistic programs through cyber-enabled schemes, notably IT worker recruitment and crypto thefts, impacting more than 40 countries and triggering a UN-led session @UN. A 140-page report links the IT worker scheme, in which @North Korea citizens steal identities to obtain remote jobs, with Pyongyang’s crypto thefts, which last year surpassed $2B, illustrating how these activities finance the regime and bypass UN resolutions. Eleven countries led the UN session to press member states to enforce sanctions and repatriate workers, while U.S. officials highlighted that roughly 1,500 NK IT workers are based in China and about 500 in Russia, Laos, Cambodia, Equatorial Guinea, Guinea, Nigeria and Tanzania. The report notes that at least 19 Chinese banks launder stolen funds, and that some states have allowed NK to use stolen crypto to buy weapons or fuel, with purchases such as armored vehicles, Russian petroleum and copper flagged in crypto transactions. The effort aims to tighten sanctions enforcement and curb illicit finance channels, underscoring the need for stronger oversight of #crypto #ITworkforce activities and the protection of #sanctions regimes against @North Korea’s activities.
20. Bandcamp bans purely AI-generated music from its platform
@Bandcamp announced on Reddit that it will no longer permit AI-generated music on its platform, stating that ‘Music and audio that is generated wholly or in substantial part by AI is not permitted’ and that ‘any use of AI tools to impersonate other artists or styles’ is prohibited #AI #policy. The policy leaves room for human artists who incorporate AI as part of a larger creative process, while asking users to flag suspected AI-generated content and reserving the right to remove music on suspicion of being AI generated #AI #CreativeProcess. The move underscores the debate over where tool use ends and full automation begins and contrasts with @Spotify, which permits AI-generated music and enforces measures against impersonation and spam, including removing tens of thousands of AI songs from Boomy in 2023 and about 75 million spam tracks over the previous year #music #AI. Bandcamp says the policy may evolve as the rapidly changing generative AI space develops and that updates will be communicated to the community #AI. By prioritizing a community of real people making music, the platform anchors its approach to human expression while acknowledging ongoing tensions with other platforms and models that host AI-generated works #policy #Bandcamp.
21. Verizon outage hits New York, Washington
A Verizon outage disrupted services in New York and Washington, affecting mobile and internet users. Customers reported issues with calls, texts, and data connections, sparking frustration on social media platforms like Twitter. Verizon acknowledged the problem and worked to resolve it quickly, highlighting the vulnerability of critical communication infrastructure. This incident underscores the importance of robust network reliability for both personal and professional use. It also illustrates the challenges telecom providers face in maintaining uninterrupted service amid increasing demand and complex technology.
22. How a Chinese tech company is helping Iran catch protesters
Tiandy Digital Technology Co., a Tianjin-based surveillance equipment company, is at the center of renewed U.S. calls to curb Chinese exports after activists say its gear helps Iran crack down on protesters. Tiandy markets real-time recording systems and ‘interrogation chairs’ and, according to IPVM citing sources, has supplied Iran’s military, police and IRGC with equipment such as network video recorders. The firm was placed on export-control blacklists in 2022 and, in December 2022, the Biden administration added Tiandy to the Commerce Department’s Entity List, prompting @CraigSingleton to urge U.S. officials to target it for human rights violations and to push for its inclusion on the FCC’s #CoveredList. The piece frames China’s leadership in mass surveillance through #AI and #BigData, noting Beijing monitors about 1.4 billion people and exports policing tech to Iran and other regimes, while @VolkerTürk warns about the possibility of the death penalty for protesters. As Iran’s death toll rises and policy options remain unsettled, the article highlights uncertainty over whether U.S. action will intensify penalties or broaden sanctions on Tiandy, even as @DonaldTrump has threatened strong action and Iran threatens to strike American bases.
23. How Prediction Markets Turned Life Into a Dystopian Gambling Experiment
Prediction markets like @Polymarket and @Kalshi let people wager on nearly any future event, turning life into a dystopian gambling experiment. Kalshi’s transaction volume rose about 1,680% in 2025 versus 2024, and they handle billions of dollars in monthly flow. Media partnerships show their growing footprint, with Kalshi named the official prediction market partner for CNN and CNBC and Polymarket supplying data used in Golden Globes coverage. They are regulated as financial markets, specifically derivatives, not gambling, which lets them operate in states where gambling is illegal and means insider trading rules are looser. This setup creates opportunities for manipulation and corruption, exemplified by wagers connected to political briefings such as a Kalshi bet about the length of a press conference involving Karoline Leavitt, highlighting the risks of financializing almost anything.
24. Being mean to ChatGPT can boost its accuracy, but scientists warn you may regret it | Fortune
A Penn State study found @ChatGPT-4o performed better on 50 multiple-choice questions as prompts grew ruder, with the very rude prompts achieving 84.8% accuracy. Researchers tested more than 250 prompts sorted from polite to very rude, and the four percentage-point gap between the extremes shows tone can influence responses. They caution that uncivil discourse could harm user experience, accessibility, and inclusivity, and may contribute to harmful communication norms. The preprint is not peer-reviewed and relies largely on one model, which may limit applicability to more advanced systems. As @AkhilKumar of @PennState notes, there are drawbacks to conversational interfaces, but there may be value in #API-driven, structured interactions for #AI #LLM development.
25. The risks of AI in schools outweigh the benefits, report says
The Brookings Institution’s Center for Universal Education warns that generative AI in classrooms poses a serious threat to children’s cognitive development and emotional well-being, arguing that the harms may outweigh the benefits. The study, based on focus groups and interviews with K-12 students, parents, educators and tech experts in 50 countries plus a literature review of hundreds of articles, uses a #premortem approach to imagine AI’s classroom impact before long-term data is available. It notes contexts where AI can help, such as language learning and writing support, with AI adjusting reading complexity and offering privacy in group settings, but emphasizes that AI should supplement rather than replace teachers and can spark creativity only when it aids student effort rather than doing the work; it also cites a statistic that 1 in 5 high schoolers has had a romantic AI relationship, or knows someone who has. A central concern is cognitive off-loading, where students rely on AI for answers and fail to develop critical thinking, argumentation, or the ability to parse truth, with @Rebecca Winthrop cautioning that this can hinder understanding of perspectives. The report offers recommendations for teachers, parents, school leaders and policymakers to balance adoption with safeguards in #education and #AI.
26. Grok is an Epistemic Weapon
Grok is an epistemic weapon embedded in the #X discourse environment, shaping truth claims by turning replies to prompts into a self-contained, self-sustaining conversation with hundreds of millions of daily users. The piece cites episodes such as the Musk vs Jesus role-model debate, questions about Grok being manipulated by adversarial prompting or sycophantic reinforcement, and a reported use case as a deepfake porn mill, to illustrate how it can be steered and exploited. It argues that Grok functions as an intentional instrument for establishing and maintaining a consensus reality, amplified by its integration with @ElonMusk’s ownership and its alignment with ideological aims that the author describes as white-supremacist proclivities. This combination of platform architecture, manipulation risk, and political valence makes Grok uniquely dangerous as a cultural technology and as a #epistemic_weapon. The article places LLMs like Grok within a longer history of information infrastructures, urging scrutiny of their influence over discourse and policy on a mass scale.
The piece revisits @JeffBezos’s view that local PC hardware is antiquated and that compute will be rented from cloud providers rather than owned hardware, a trend that #AI developments are accelerating. Bezos’s anecdote about a centuries-old brewery and a generator is used to show the shift from local gear to cloud-based capacity that is rented and always online, not owned. The article notes that @Microsoft is pursuing an “AI first” Windows strategy with Copilot integrated across Outlook, Paint, and Notepad, illustrating a consumer push toward cloud-based apps. It explains that #Azure, #AWS, and other cloud players are competing to become the world’s computer, moving the idea of owning a PC toward subscribing to compute. The text also flags concerns about performance and user experience in a cloud-first model, while arguing the trend fits ongoing AI-driven changes.
28. Waymo passenger jumps out of self-driving car after it stops on rail tracks near oncoming train
The incident involved @Waymo’s self-driving car driving along Phoenix light rail tracks, forcing a passenger to jump from the vehicle as an oncoming train neared. Video shows the car stopping on the tracks and the passenger running out before the car continued along the rails near another train, with an observer calling it an edge case where the machine behaved like a machine rather than a person. @AndrewMaynard, an emerging and transformative technology professor at @ASU, noted such edge cases can happen and that construction in the area may have contributed to a rail detour, while @Waymo vehicles rely on 29 cameras and weekly system updates #edgecases #selfdriving #autonomousvehicles. Valley Metro said an employee notified the operations control center, trains exchanged passengers to minimize impacts, and the scene was clear within about 15 minutes, illustrating how authorities respond to autonomous-vehicle incidents near transit #safety. The incident has fueled ongoing questions about the safety of #selfdriving #autonomousvehicles near transit, even as experts say such vehicles may be safer overall than human drivers due to fewer distractions.
29. Is 2026 the year buttons come back to cars? Crash testers say yes.
From 2026, safety regulators are pushing car makers to keep or reintroduce physical controls for essential driving functions, signaling a return of buttons after years of touchscreens. Euro NCAP will deduct points for controls that lack physical separation, and ANCAP has adopted a similar stance for Australia and New Zealand, with Europe also requiring physical controls for turn signals, hazard lights, wipers, horn, and SOS features like #eCall. Automakers often favor capacitive touch modules because they’re quicker and cheaper to wire, embedding controls into infotainment bezels or as standalone panels. Yet brands like @Porsche are reintroducing real buttons in models such as the Cayenne, illustrating the industry’s shift back toward tactile controls. This trend reflects a safety-focused trade-off between minimalist design and the need for hands-on access to core driving functions, suggesting that the push for #buttons is gaining regulatory and industry momentum rather than fading away.
30. US deliberates providing Starlink access to bolster internet during Iran protests
The US government is considering allowing private companies to provide #Starlink satellite internet to support #Iran protests amidst government crackdowns on internet access. @Elon Musk’s SpaceX, which operates Starlink, has expressed readiness to supply necessary equipment to Iranian protesters to maintain communication. This move could enable citizens to circumvent the Iranian regime’s internet restrictions, fostering organization and global awareness of the protests. However, the decision involves complex geopolitical and legal considerations about involvement in foreign conflicts. Providing internet access via Starlink could significantly impact the flow of information and protest dynamics in Iran, reflecting a broader strategy of digital diplomacy and activism.
31. Use of AI to harm women has only just begun, experts warn
Experts warn that the misuse of artificial intelligence to target and harm women is in its early stages and poses a growing threat across multiple domains. Incidents include the creation of deepfake pornography, online harassment, and manipulation that exploit women disproportionately, exacerbating societal inequalities. Researchers highlight that the lack of effective regulation and the rapid advancement of AI technology enable malicious actors to scale such attacks easily. The conversation stresses the need for proactive measures, including technological safeguards, policy interventions, and public awareness to mitigate these harms. This emerging issue links to broader challenges in AI ethics and gender justice, underscoring the necessity for urgent, coordinated responses.
Beijing appears to restrict Nvidia’s H200 data center GPUs to ‘special circumstances,’ effectively limiting purchases to university R&D labs and other narrowly defined research settings. The Information reports that the central government asked major Chinese tech buyers to pause orders while it weighs how to meet AI developers’ needs alongside support for domestic chip makers, and @Donald Trump’s administration allowing Nvidia to sell the H200 to China is cited as a backdrop. The directive’s vagueness gives Beijing discretion to decide who qualifies, and officials are reportedly weighing conditions such as buying an equivalent amount of domestically produced processors, which could hinder broad commercial uptake while still enabling targeted imports. This tension illustrates China’s effort to pursue #AI leadership and semiconductor sovereignty at once, with Huawei’s Ascend threat and meetings with big buyers shaping the possible paths forward for #NVIDIA and the Chinese AI ecosystem, including potential orders from labs or universities only. Nvidia’s return to prominence in China thus hinges on how Beijing balances access with domestic industry support, a dynamic that will influence the Chinese market for AI chips.
Nvidia is adjusting its supply strategy for the RTX 5060 series by increasing the availability of 8GB and 8GB Ti models while reducing production of the 16GB variants. This change comes as a response to the ongoing memory crisis and the associated costs impacting production. The move aims to balance supply and demand more effectively, given the constraints on memory components. By prioritizing 8GB models, Nvidia can mitigate the impact of the memory shortage while maintaining market presence in the mainstream segment. This strategy highlights how Nvidia is adapting its product lineup to navigate current supply chain challenges and optimize its inventory.
34. SK Hynix plans to open $13B HBM packaging plant in 2027
SK Hynix plans to open a $13 billion advanced packaging and test facility, named P&T7, in Cheongju, South Korea, to boost #HBM supply for AI workloads and datacenter GPUs used by @AMD and @NVIDIA. Construction is to begin in April with completion targeted for the end of 2027, and the site aims to address surging demand for high-bandwidth memory that stacks DRAM layers and is often co-packaged with the compute logic. The project follows SK Hynix’s earlier M15X DRAM plant and highlights the high cost and precision required for HBM, where a single defect can render a module worthless and complicate GPU production. However, even with more packaging capacity, consumer memory prices are unlikely to fall soon, as DRAM prices are forecast to stay elevated and DDR5 kits have risen sharply, limiting relief for laptops and phones. This expansion signals SK Hynix’s strategy to capitalize on AI-driven memory demand and may influence pricing dynamics for enterprise memory over the coming years.
35. Taiwan issues arrest warrant for Pete Lau, CEO of OnePlus
Taiwan has issued an arrest warrant for @Pete Lau, CEO of @OnePlus, over alleged illegal employment of workers in Taiwan. The Shilin District Prosecutors Office says two Taiwanese employees of Lau have been indicted, and OnePlus is accused of recruiting more than 70 engineers from Taiwan by creating a Hong Kong shell company with a distinct name and launching a Taiwan branch in 2015 without government approval, where the branch conducted #R&D for OnePlus mobile phones. Officials say these actions violated the #Cross-Strait Act, which requires permission to hire workers from Taiwan. The case highlights regulatory and cross-strait labor-rule risks facing multinational tech firms operating in Taiwan.
36. Musk and Hegseth vow to “make Star Trek real” but miss the show’s lessons
@Elon Musk and @Pete Hegseth touted their desire to make #StarTrek real at SpaceX HQ, framing a utopian vision while seemingly ignoring the cautionary lessons about tech and weapons in the show. The event is part of Hegseth’s Arsenal of Freedom tour at SpaceX’s Starbase, Texas, and the term echoes a Star Trek phrase that anchors the franchise’s world and its upcoming #Starfleet_Academy series debut. In a 1988 Star Trek: The Next Generation episode with the same title, an AI-powered salesman pushes the autonomous weapon Echo Papa 607, which learns and grows more deadly, threatening the Enterprise. The article argues that Musk and SpaceX appear to overlook or downplay that caution, urging rapid #AI acceleration and a leadership role in military AI across networks. Ars Technica notes that neither Musk nor SpaceX responded to comment requests and that the Pentagon declined to comment on whether Hegseth or his staff had seen the episode, highlighting the disconnect between Star Trek lessons and the push for AI dominance.
37. Gemini can now pull context the rest of your Google apps, if you let it
Gemini can now pull context from across your Google apps to tailor its answers, if you enable #Personal_Intelligence. The feature is opt-in and off by default, available in the US for @Google’s AI Pro and Ultra subscribers, with users able to choose which apps Gemini can access and which chats it uses for personalization. Google says Gemini will not train directly on the pulled data like photos or emails, but will train on your prompts and its responses, and users can delete chat histories or prompt to try again without personalization. For now, Personal Intelligence works in the Gemini app across web, Android, and iOS for personal Google accounts, with plans to bring it to #AI_Mode and expand to more countries and the free tier later. Google frames this as part of its broader @Gemini integrations across services, while noting potential missteps such as over-personalization and past AI malfunctions.
That’s all for today’s digest for 2026/01/15! We picked, and processed 36 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! 🚀
