Meridiano

Add Article

Articles tech

Showing articles 481 - 495 of 810 total (matching filters).

• 
  tech 6

  Large genome model: Open source AI trained on trillions of bases

  (Ars Technica / 2026-03-04 22:14)

  The AI system Evo, originally trained on bacterial genomes to predict protein sequences, has been significantly upgraded to Evo 2. This new, open-source model was trained on trillions of DNA base pairs across all domains of life, including complex eukaryotes.

  As a result, Evo 2 has learned to identify intricate genomic features in eukaryotes, such as regulatory DNA and splice sites, which are challenging to analyze due to their weak and scattered patterns. This advancement overcomes the initial limitation of applying the method to more complex genome structures.

  The main topics covered are the evolution of the Evo AI system, the structural differences between bacterial and eukaryotic genomes, and the new capabilities of Evo 2 in understanding complex genetic features.

• 
  tech 5

  Here’s how Google describes its fee-reducing Apps Experience and Games Level Up programs

  (The Verge / 2026-03-04 22:31)

  Google is reducing its standard app store fees and introducing new programs that incentivize developers to expand beyond phones. Following a US antitrust ruling, it will generally charge 20% for in-app purchases and 10% for subscriptions, with further discounts available.

  Through its "Apps Experience" and "Games Level Up" programs, developers can save an additional 5% by optimizing apps for additional form factors like tablets, PCs, TVs, and XR headsets. The "Games Level Up" program specifically requires games to meet quality, stability, and integration standards across these devices by set deadlines starting in 2026.

  The changes are part of binding commitments from Google's legal settlement with Epic Games. The company aims to encourage a broader, higher-quality app ecosystem across multiple Android-powered devices.

  Main Topics: Google Play fee reduction; antitrust compliance; developer incentives for multi-form factor apps; "Games Level Up" program requirements.

• 
  tech 5

  Space Command chief throws cold water on the question of UAPs in space

  (Ars Technica / 2026-03-04 22:32)

  President Trump recently directed federal agencies to begin disclosing government files related to UAPs (unidentified anomalous phenomena), reigniting public interest. Official reports from NASA and the Office of the Director of National Intelligence have found no conclusive evidence linking UAPs to extraterrestrial origins, citing limited high-quality data.

  The government has released intriguing military videos of UAPs, but officials have not commented substantively on Trump's latest order or released imagery from the most sophisticated space-based sensors. The main topics covered are the political push for UAP disclosure, official investigative findings, and the limitations of available evidence.

• 
  tech 5

  Anthropic CEO Dario Amodei calls OpenAI’s messaging around military deal ‘straight up lies,’ report says

  (TechCrunch / 2026-03-04 22:40)

  Anthropic's CEO, Dario Amodei, criticized OpenAI's new defense contract as "safety theater," accusing Sam Altman of misleading the public. The criticism stems from Anthropic's own failed negotiations with the Department of Defense, where it refused a deal over concerns about potential use for mass surveillance or autonomous weapons.

  The main topics covered are the ethical dispute between Anthropic and OpenAI over military AI contracts, the specific contractual concerns regarding "lawful use," and the public and employee reactions to these deals.

  OpenAI secured the DoD agreement by stating it includes explicit protections against those same prohibited uses, though critics note laws can change. Public sentiment appears to side with Anthropic, evidenced by a spike in ChatGPT uninstalls and Amodei's claim that his company is now viewed as "the heroes."

• 
  tech 6

  TerraPower gets OK to start construction of its first nuclear plant

  (Ars Technica / 2026-03-04 22:54)

  The US Nuclear Regulatory Commission has issued its first construction approval in nearly a decade, allowing TerraPower to begin work on a new reactor site in Kemmerer, Wyoming. The company, backed by Bill Gates, is developing a novel sodium-cooled fast-neutron reactor called Natrium, which incorporates a salt-based energy storage system.

  This design allows the plant to operate flexibly alongside renewable energy sources and can temporarily boost its electrical output. While the construction permit is a major milestone, it does not guarantee future approval to operate the reactor.

  The main topics covered are the regulatory approval for construction, the novel design features of the TerraPower Natrium reactor (including sodium cooling and integrated storage), and the strategic rationale for its flexibility.

• 
  tech 5

  Grammarly Is Offering ‘Expert’ AI Reviews From Your Favorite Authors—Dead or Alive

  (Feed: Artificial Intelligence Latest / 2026-03-04 22:56)

  Grammarly has expanded its AI-powered writing tool to include an "expert review" feature. This feature offers critiques from AI agents modeled on the writing styles of real authors and academics, both living and deceased, though without their endorsement or direct involvement.

  The company, now rebranded as Superhuman, states the feature leverages its underlying large language model to surface suggestions inspired by these experts' works. However, the practice of training AI on these individuals' bodies of work raises significant legal and ethical questions regarding copyright and the unauthorized use of their names and reputations.

  The main topics covered are the new AI features of Grammarly, specifically the "expert review" function, the company's rebranding, and the ethical and legal concerns surrounding the use of real individuals' work and personas without permission.

• 
  tech 5

  Tim Sweeney signed away his right to criticize Google until 2032

  (The Verge / 2026-03-04 23:52)

  Epic Games CEO Tim Sweeney, a famously vocal critic of Google and Apple, has agreed to a settlement with Google that severely restricts his ability to criticize the company. As part of the binding term sheet, Sweeney has signed away his right to disparage Google or advocate for changes to its app store policies until at least 2032.

  The agreement requires Sweeney and Epic to publicly state that Google's Android platform is "procompetitive" and to make good faith efforts to advocate for it as a model. While Epic can still target Apple through the Coalition for App Fairness, Sweeney is now effectively muzzled regarding Google and may even have to appear in court to defend the settlement.

  The main topics covered are the legal settlement between Epic and Google, the specific non-disparagement and advocacy clauses imposed on Tim Sweeney, and the effective end of his public criticism of Google's app store practices for nearly a decade.

• 
  tech 5

  House Iran War Powers Resolution Could Lose Support to Competing Bill by Pro-Israel Democrat

  (The Intercept / 2026-03-05 00:00)

  The U.S. Senate blocked a war powers resolution that would have required President Trump to seek congressional approval for continuing military strikes against Iran. The vote failed largely along party lines, with nearly all Republicans and one Democratic senator, John Fetterman, opposing the measure.

  A companion resolution faces a similarly difficult vote in the House of Representatives, where it is opposed by some pro-Israel Democrats who have introduced alternative legislation. This alternative would grant the President a 30-day window to continue operations, a timeframe advocates of the stricter resolution criticize as granting a "blank check."

  The debate centers on congressional war powers and the potential for a prolonged conflict, with Democrats warning of endless war and Republicans supporting the campaign to counter Iranian threats.

  Main Topics: U.S. Congressional war powers; Iran conflict; partisan political divisions; legislative procedures.

• 
  tech 5

  Seven tech giants signed Trump’s pledge to keep electricity costs from spiking around data centers

  (The Verge / 2026-03-05 00:17)

  Leaders from seven major tech companies met with President Donald Trump to sign a "Ratepayer Protection Pledge." This voluntary commitment aims to address concerns that data center expansion for AI will raise electricity costs by having companies pay for necessary grid upgrades and new power generation.

  The pledge is a response to bipartisan worries about soaring electricity demand and rising household bills, seeking to assure communities they will not bear the financial burden. It also includes provisions for companies to potentially contribute backup power to local grids during times of peak demand or emergency.

  The main topics covered are the tech industry's response to data center energy concerns, the specifics of a voluntary pledge to protect ratepayers from cost spikes, and the political context of the agreement.

• 
  tech 5

  With a little nudge from RBI, banks tap AI for bigger tasks

  (Tech-Economic Times / 2026-03-05 00:30)

  Financial services in India are moving beyond basic AI automation in call centers to adopt the technology for more complex functions like risk compliance, customer onboarding, and credit decisioning. This expansion follows regulatory guidelines from the Reserve Bank of India, which have provided a benchmark for AI adoption and increased sector-wide confidence.

  Key application areas now include customer servicing, loan sales, and collections, with several Indian AI startups deploying multilingual voice-bots for major banks and NBFCs. While cost reduction is a significant driver, with AI agents currently estimated to be cheaper than human agents, some industry executives note that substantial cost arbitrage has not yet been fully realized and may develop further over time.

• 
  tech 5

  Brokerages split on future of IT as AI automates nuts and bolts

  (Tech-Economic Times / 2026-03-05 00:30)

  Analysts are divided on AI's long-term impact on IT services firms, with some forecasting revenue deflation and earnings downgrades for core services like application management, while others believe AI will expand the market by creating new work alongside automation.

  The debate centers on execution, with performance hinging on how companies manage the transition. Key uncertainties include the pace of AI integration into enterprise systems and the balance between productivity gains and new revenue streams.

  Major topics covered include the divergent analyst forecasts for the sector, the specific business areas at risk from AI automation, and the critical role of IT firms in implementing and integrating AI within complex enterprise environments.

• 
  tech 5

  Human still in driver's seat: Ford chief AI officer Franziska Bell

  (Tech-Economic Times / 2026-03-05 00:30)

  Ford Motor Co. is deploying artificial intelligence across its operations, from vehicle design to factory quality assurance, to augment human workers and dramatically speed up processes. The company's strategy emphasizes "human AI teams," using AI as a tool to enhance human expertise rather than replace it, with key applications including generating design variations and accelerating aerodynamic simulations.

  AI is being integrated into multiple functions, including enabling designers to create hundreds of vehicle variations from text prompts and reducing simulation times from hours to seconds. The technology is also used for quality inspections via computer vision and for creating unified customer support interfaces.

  The main topics covered are Ford's AI integration strategy, specific use cases in design and engineering, and the role of its teams in India.

• 
  tech 5

  With a little nudge from RBI, banks tap AI for bigger tasks

  (Tech-Economic Times / 2026-03-05 00:30)

  Financial services companies in India are expanding their use of AI beyond basic call center automation to more complex functions like risk compliance, customer onboarding, and credit decisioning. This shift follows regulatory guidelines from the Reserve Bank of India, which have provided a benchmark for AI adoption and increased sector-wide confidence in the technology.

  Key application areas now include customer servicing, loan sales, and collections, with several AI startups deploying multilingual voice-bots for major banks and financial institutions. While customer engagement remains the primary use, the scope is widening, though human agents still handle complex escalations.

  A significant driver for adoption is cost, with AI agent costs currently estimated at half that of human agents by some startups, though other industry executives argue the cost arbitrage in India is not yet fully realized. The overall trend points toward increasing integration of AI across various banking workflows as technology costs are expected to decline further.

• 
  tech 5

  Jensen Huang says Nvidia is pulling back from OpenAI and Anthropic, but his explanation raises more questions than it answers

  (TechCrunch / 2026-03-05 01:08)

  Nvidia CEO Jensen Huang stated the company's recent investments in OpenAI and Anthropic will likely be its last in both firms, as the opportunity to invest will close once they go public later this year. Nvidia's spokesman emphasized the company's investments are strategically focused on expanding its ecosystem, a goal these stakes have met.

  Other dynamics may explain the pullback, including the circular nature of such deals where investments are often reciprocated with chip purchases. Furthermore, Nvidia's relationships with both AI companies have become complex, with Anthropic facing a federal blacklist and publicly criticizing chip sales to China, while OpenAI secured a Pentagon deal.

  The result is that Nvidia now holds stakes in two companies that are moving in divergent strategic and political directions, potentially affecting their broader ecosystem of customers and partners.

• 
  tech 4

  A new video from the White House mixes Call of Duty footage with actual video of Iran strikes

  (The Verge / 2026-03-05 01:09)

  The White House released a video depicting real U.S. military strikes on Iran, edited in the style of a "Call of Duty" highlight reel, complete with game footage and dramatic music. The video's introduction specifically used animation resembling a nuclear strike killstreak from "Call of Duty: Modern Warfare III."

  This incident is part of a pattern where the administration uses popular video game imagery and references to promote its agenda on social media. Previous examples include the Department of Homeland Security using Pokémon and "Halo" themes in posts about immigration enforcement.

  The main topics covered are the White House's use of video game aesthetics in official military footage, the specific "Call of Duty" references, and the broader trend of government agencies employing gaming culture for propaganda.