Meridiano

Add Article

Articles tech

Showing articles 271 - 285 of 665 total (matching filters).

• 
  tech 5

  AI may be creating instead of destroying jobs for now, ECB blog argues

  (Tech-Economic Times / 2026-03-04 12:05)

  A European Central Bank blog post suggests that, contrary to common fears, firms in the euro zone significantly using AI are more likely to hire additional staff in the near term. It also finds that companies planning AI investments have positive expectations for employment growth, with no expected hiring pause due to AI investment in the coming year.

  However, the authors caution that the longer-term outlook could change as AI begins to more deeply transform production processes, noting that other surveys predicting job losses often look further ahead.

  The main topics covered are the short-term employment effects of AI adoption, contrasting views on AI's impact on jobs, and the difference between near-term and long-term forecasts.

• 
  tech 5

  China poised to expand global access to AI despite rivalry with the US: Wang Jian

  (Tech - South China Morning Post / 2026-03-04 12:01)

  A prominent Chinese AI scholar states China has the resources and responsibility to expand global access to artificial intelligence. Wang Jian argues this contrasts with other nations, like the U.S., which he says are increasingly acting in their own interests.

  He emphasizes the goal of turning AI into an "international public product," a concept promoted by President Xi Jinping. Wang positions China as uniquely capable of advancing global technology while adhering to this ideal.

  The main topics covered are China's role in global AI accessibility, the U.S.-China tech rivalry, and the concept of AI as a public good.

• 
  tech 4

  Nvidia driver 595.71 reportedly limits overclocks on some GeForce GPUs, but not all — troubled driver release seems to stifle voltages on RTX 40- and 50-series cards

  (Latest from Tom's Hardware / 2026-03-04 12:00)

  Nvidia's latest 595.71 graphics driver is reportedly imposing artificial voltage limits on some RTX 40- and 50-series GPUs, reducing overclocking headroom by up to 200MHz. The issue appears to be a bug, as it inconsistently affects certain models and is not mentioned in official patch notes.

  User reports and demonstrations show significant performance drops for overclocked cards, though some models, like specific Gigabyte and PNY variants, remain unaffected. Nvidia has not yet officially acknowledged the problem, leaving open the possibility of a future corrective driver update.

  The main topics covered are a problematic Nvidia driver update, its impact on GPU overclocking and voltage, inconsistent effects across different hardware models, and the lack of an official response from Nvidia.

• 
  tech 5

  Who needs data centers in space when they can float offshore?

  (TechCrunch / 2026-03-04 12:00)

  The severe power crunch for AI data centers is prompting innovative location proposals, including offshore. Startup Aikido plans to submerge a demonstration data center off Norway this year, powered by a floating wind turbine, with a larger U.K. deployment targeted for 2028.

  This offshore approach aims to solve multiple challenges: direct proximity to consistent wind power, simplified cooling using cold seawater, and avoidance of community opposition. However, it introduces new engineering hurdles due to the harsh, corrosive marine environment and the need to secure equipment against movement.

  The concept is not entirely new, as Microsoft previously ran a successful two-year submerged data center trial but ultimately discontinued the project in 2024.

  Main Topics: AI data center power demand, offshore/submerged data centers, renewable energy integration, engineering challenges, previous industry experiments.

• 
  tech 6

  OpenAI looking at contract with NATO, source says

  (Tech-Economic Times / 2026-03-04 11:58)

  OpenAI is considering a contract to deploy its AI technology on NATO's unclassified networks, clarifying an earlier misstatement about classified networks. This follows OpenAI's recent deal with the Pentagon, which came after the U.S. government was directed to stop working with rival Anthropic.

  Anthropic's removal stemmed from a standoff over its opposition to the Pentagon using its AI for mass domestic surveillance or autonomous weapons. In its Pentagon deal, OpenAI included a provision that its systems shall not be intentionally used for domestic surveillance of U.S. persons, with the Pentagon affirming intelligence agencies like the NSA would not use the services.

  OpenAI CEO Sam Altman acknowledged the Pentagon deal was a complex decision with significant negative short-term public relations consequences for the company.

  Main Topics: OpenAI's potential NATO contract; OpenAI's Pentagon deal and its ethical safeguards; Anthropic's removal and its stance on military AI use; The public relations impact of military contracts for AI companies.

• 
  tech 5

  Micron sampling first 256GB SOCAMM2 memory packages to customers — 2TB of RAM per CPU is now in reach of datacenter players

  (Latest from Tom's Hardware / 2026-03-04 11:30)

  Micron has begun sampling the industry's first 256GB SOCAMM2 memory modules, a 33% density increase over the 192GB modules released six months prior. This advancement enables data center servers to support up to 2TB of RAM per CPU, significantly boosting memory capacity for AI workloads.

  Beyond increased density, the new modules offer substantially improved power efficiency compared to traditional RDIMMs and support liquid cooling. The technology also aims to enhance AI performance by allowing for larger model context windows, which can reduce response times.

  The main topics covered are a breakthrough in high-capacity server memory, its benefits for AI data center performance and efficiency, and the industry collaboration behind the SOCAMM2 form factor.

• 
  tech 5

  Donald Trump to meet Google, Meta, OpenAI on energy pledge ahead of midterms

  (Tech-Economic Times / 2026-03-04 11:15)

  President Trump is meeting with major tech companies to formalize a "Ratepayer Protection Pledge," aiming to prevent rising household electricity costs driven by energy-intensive AI data centers. The expected commitments include tech firms securing their own power supplies, funding grid upgrades, and entering special rate agreements to avoid straining public utilities.

  A critic notes that the policy's focus on fossil fuel generation may not alleviate grid pressure quickly, as these plants take longer to build than renewable alternatives. The initiative is also viewed as politically timed ahead of midterm elections, with observers watching to see if it yields concrete action or remains symbolic.

  The main topics covered are the White House tech meeting, the "Ratepayer Protection Pledge" details, political and economic motivations, and criticisms regarding implementation speed and energy source focus.

• 
  tech 5

  How Is Kalshi Not Gambling?

  (Backchannel Latest / 2026-03-04 11:00)

  Kalshi is a CFTC-regulated prediction market that allows users to trade on the outcomes of various events, from sports to elections to pop culture. Its founders argue it is distinct from gambling, a claim supported by federal regulation enabling it to operate in all 50 states, unlike online sports betting.

  The platform has grown rapidly, handling over $1 billion in weekly trades and achieving an $11 billion valuation, though it faces legal challenges from 19 states that dispute its non-gambling classification. Kalshi's markets have demonstrated predictive accuracy in some areas, such as elections and interest rates, outperforming traditional polls and experts.

  The company's CEO, Tarek Mansour, vigorously defends its social utility in aggregating information and sees it as analogous to financial instruments like stock trading, despite critics who view it as a form of betting.

• 
  tech 5

  Gemini API key thief racks up $82,314 in charges in just two days, victim 'facing bankruptcy' — affected devs call for basic guardrails against 'catastrophic usage anomalies'

  (Latest from Tom's Hardware / 2026-03-04 11:00)

  A software developer's stolen Gemini API key was used to generate over $82,000 in charges in 48 hours, threatening their business with bankruptcy. The victim has secured their account, but initial feedback from Google suggests the company may enforce payment, citing its Shared Responsibility Model for cloud security.

  Affected developers are criticizing Google for lacking basic guardrails, such as spending caps or anomaly freezes, to prevent such catastrophic usage spikes. The article notes that while consumer Gemini accounts have usage caps, business users must proactively set their own quotas and budget alerts within Google's cloud platform.

  The main topics covered are a significant financial incident due to API key theft, the ensuing dispute over liability with Google, and calls for improved protective measures from service providers.

• 
  tech 4

  Eight Sleep raises $50M at $1.5B valuation

  (TechCrunch / 2026-03-04 11:00)

  Sleep tech company Eight Sleep has raised $50 million in a strategic funding round led by Tether Investments, achieving a valuation of $1.5 billion. The company, which sells smart mattress accessories for sleep tracking and temperature adjustment, plans to use the capital for new products, global expansion, and clinical validation.

  Key initiatives include pursuing FDA approval for sleep apnea detection and developing a proactive, AI-driven agent to control bed settings and provide personalized sleep habit guidance. The company also aims to expand beyond consumer products into broader health technology.

  Main topics covered include the company's funding and valuation, its product roadmap and expansion plans, and its competitive positioning in the sleep tech market.

• 
  tech 6

  Tennessee Wants to Let Schools Ban Immigrant Kids, Threatening to “End Public Education as We Know It”

  (The Intercept / 2026-03-04 10:52)

  Tennessee Republicans are advancing legislation that would challenge the Supreme Court's Plyler v. Doe decision, which guarantees a free public K-12 education to children regardless of immigration status. The bills, part of a broader national effort promoted by the Heritage Foundation, could force schools to deny enrollment to undocumented students or report their status.

  If enacted, these laws would likely provoke a Supreme Court challenge to overturn Plyler. Educators warn such a move would fundamentally damage public education, creating an "uneducated underclass" and forcing school staff to act as immigration informants.

  Similar bills have been introduced in five other states, signaling a coordinated national push. The Tennessee legislation is scheduled for committee hearings, with one bill requiring schools to report on students' immigration status contingent on both the passage of the enrollment bill and the overturning of Plyler.

  Main Topics: Tennessee immigration education bills, challenge to Plyler v. Doe, national effort led by the Heritage Foundation, potential consequences for public education and undocumented children.

• 
  tech 5

  Fintech lender Moneyview files DRHP with Sebi for IPO, seeks to raise Rs 1,500 crore via fresh issue

  (Tech-Economic Times / 2026-03-04 10:38)

  Fintech startup Moneyview has filed for an IPO to raise Rs 1,500 crore through fresh shares, with existing investors like Accel also offering shares for sale. The company plans to use the proceeds to strengthen its in-house NBFC and support loan disbursals through partner lenders.

  The company, which started with consumer loans, has diversified into a full-stack financial platform offering investments, insurance, and UPI payments. It reported a net profit of Rs 209 crore for the first nine months of fiscal 2026 and has been profitable for three consecutive years.

  Moneyview serves 125 million users, with 10 million revenue-generating customers, and partners with 42 financial institutions. It has also improved its asset quality, reducing its annualised bad loan rate to 6.9%.

  Main Topics: IPO filing and details, financial performance and diversification, user base and business model, asset quality and future growth plans.

• 
  tech 5

  Google launches Gemini 3.1 Flash-Lite, its “fastest and most cost-efficient” AI model

  (Tech-Economic Times / 2026-03-04 10:01)

  Google has launched Gemini 3.1 Flash-Lite, a new AI model it describes as its fastest and most cost-efficient in the Gemini 3 series. It is now available in preview for developers and enterprises via Google AI Studio and Vertex AI, priced at $0.25 per million input tokens.

  The model is designed for scalable tasks like translation and content moderation, as well as more complex workloads requiring reasoning, such as generating user interfaces. It features "thinking levels" to allow control over computational effort per task.

  Early testers praise its efficiency and capability to handle complex inputs. Benchmark results show it outperforms previous models in its tier, achieving high scores on reasoning and multimodal understanding tests.

• 
  tech 6

  What It’s Like to Have a Brain Implant for 5 Years

  (Science Latest / 2026-03-04 10:00)

  Rodney Gorham has successfully used a Synchron brain-computer interface implant for five years, a milestone for the company. The implant, called the Stentrode, allows him to control digital devices with his thoughts, providing critical assistance as he lives with ALS.

  Synchron is preparing for a larger pivotal trial needed for regulatory approval, working with the FDA to define clinical endpoints. The technology relies on decoding algorithms to translate neural signals into actions, with Gorham's long-term feedback being instrumental in its development.

  The main topics covered are: the milestone of long-term BCI use, the technical function of Synchron's Stentrode implant, the path to regulatory approval, and the role of user feedback in evolving the technology's capabilities.

• 
  tech 5

  Lux Capital warns portfolio startups to brace for market risks

  (Tech-Economic Times / 2026-03-04 09:54)

  Venture capital firm Lux Capital has warned its portfolio startups to prepare for potential economic challenges, citing concerning signals like falling bond yields alongside high stock markets. The firm advised founders to extend their cash runway, review debt covenants, and closely analyze costs, especially for companies affected by tariffs.

  Lux Capital co-founder Josh Wolfe also expressed concern about an AI bubble, stating that fewer than ten AI startups truly matter, and suggested companies should go public while market enthusiasm is high. Despite these warnings, Wolfe remains broadly optimistic about startup investing, emphasizing that difficult times can still produce major successes, but urged founders to rigorously stress-test their assumptions.

  The main topics covered are: a venture capital firm's economic warning and advice to startups, concerns about an AI investment bubble, and the paradoxical optimism about long-term startup investment despite near-term risks.