Turtles AI Festina Lente - Your leading source of AI news News

Work and AI: The Silent Change

Automated tasks, simplified roles and growing inequalities are transforming the European labour market without official monitoring, while the impact of AI is already being felt in the present of millions of workers

Super-resolution video in one pass with sharp details and temporal coherence

The DLoRAL method combines two specialized modules in an optimized diffusion architecture, alternating between learning the coherence between frames and refining the spatial details for realistic and efficient results

Centaur: The Model That Mimics Human Thought on a Large Scale

AI-based system predicts human behavior in hundreds of psychological experiments, getting closer to the real dynamics of the mind thanks to training on millions of individual choices

Huawei Rejects Allegations of Plagiarism on AI Pangu Pro Model

Company denies similarities to Alibaba’s Qwen 2.5 after HonestAGI analysis, claims independent development and adherence to open source licenses

ASUS GX10: The mini-PC with NVIDIA Blackwell heart debuts on July 22

Equipped with the Grace Blackwell GB10 chip, 128GB of unified memory and 200GbE networking, the GX10 delivers AI power in a compact form factor for developers and computing professionals.

Cisco Duo IAM: Digital Identity Under Control, Security Beyond Passwords

The new platform unifies MFA, BLE, AI and access management to protect user identity from phishing, session theft and targeted attacks, with a proactive approach and a seamless user experience

Chatbots distort science: New AIs exaggerate results and reduce accuracy

An analysis of 4,900 abstracts shows that recent linguistic models over-amplify and over-simplify scientific data, especially in the medical field, with increasing risks for the correct interpretation of evidence

Graduates waiting: the work that disappears between algorithms and promises

In Italy and the United States, unemployment among new graduates is growing while AI replaces young people in entry-level roles. Digital skills required, but few are truly ready

AI Accelerates Cutting of White-collar Roles

CEOs of companies such as Ford, Amazon and JPMorgan admit the direct impact of AI: millions of white-collar jobs at risk, with effects already visible in hiring strategies and work reorganization

Invisible Prompts to Trick AI: Researchers in the Crosshairs

Some authors hide instructions in preprints to influence AI-generated feedback during peer review. Universities involved in eight countries, between self-censorship and justifications against “automated” reviewers

Zhipu AI Accelerates to Stock Market with New Funds, Launches Powerful Multi-Modal Model

With a strategic investment of RMB 1 billion and 25 million active users, Chinese startup introduces GLM‑4.1V‑Thinking and expands AI integration into enterprise devices and services

HODDL Launches First AI Platform for Patriotic Education

A new personalized education system integrates civic values, advanced technologies and career paths for students and adults of all ages

Project Rainier: AWS’s New AI Supercluster to Power Up Claude Models

Built on hundreds of thousands of Trainium2 chips, AWS’s modular system integrates ultra-high-speed networking and custom design for training at scale

Cloudflare Launches Pay per Crawl: Paid Access for AI Bots on Web Content

The new system allows publishers to monetize AI crawler requests by blocking automatic access to content unless authorized or paid for, with the support of major media and online platforms

Brussels Confirms AI Act Deadlines Despite Big Tech Pressure

The European Union rejects the request for postponement made by over one hundred companies and keeps the calendar unchanged: first binding measures from August 2025, full implementation by 2026

Meta Trials Proactive Chatbots to Restart Conversations and Strengthen Bonds with Users

With the “Omni” project, the company introduces digital assistants capable of memorizing interactions, starting chats autonomously and promoting more constant and personalized communication, respecting precise limits so as not to be invas

Intel Changes Course on AI: Acknowledging Past Mistakes, Focus on Jaguar Shores

Former CEO Gelsinger admits he underestimated the impact of AI. With Falcon Shores shelved, the new course led by Lip-Bu Tan relaunches on design and rack-scale systems to regain ground in the sector

Healthcare and AI: MENA Market Grows with Data, Investments and Personalization

Expanding clinical databases, infrastructure enhancements and tailored medicine drive healthcare AI adoption in the Middle East and North Africa, with annual growth expected to reach 35.8% through 2029

BrainProbe: AI Enters Psychiatry with Over 90% Accuracy

Developed in Taipei, the new AI platform analyzes MRI to identify early signs of schizophrenia and monitor brain aging, opening up more objective and personalized diagnoses

Gemini Gets a Makeover on iOS: New Look, More Features, and Deep Integration

Google Assistant updates interface with brand colors, adds real-time camera and screen support, integrates apps like Spotify and Maps, and introduces live subtitles and contextual visual search

Cisco Launches Secure, Intelligent Network for the AI ​​Era

New unified architecture with AI-powered management, high-performance devices and integrated security for enterprise, industrial and distributed environments

Grammarly Acquires Superhuman: AI Enters Email

With the integration of the famous email client, Grammarly accelerates on the productivity and AI front, bringing new smart agents for communications, tasks and calendar into its ecosystem

X Tests AI Chatbot to Write Community Notes: Experimental Phase Begins

AI-generated notes will be integrated on-demand, clearly labeled and human-verified, with the aim of making fact-checking on the platform faster and more transparent

US Senate Blocks State AI Law Ban: Big Tech Defeat

In a near-unanimous vote, senators eliminate a decade-long moratorium that would have prevented states from regulating AI. Bipartisan victory for local protections against deepfakes and disinformation risks

Leo XIV and the digital era: spiritual guide in the era of the algorithm

The new Pope recalls the social doctrine of the Church to face the challenges of AI, relaunching human dignity, work and ethical responsibility as pillars of the future

Two photons beat supercomputer: a new step for quantum machine learning

An integrated photonic circuit demonstrates real speed advantages, precision and efficiency compared to classic algorithms, opening the road to hybrid systems for learning based on kernel

NVIDIA GB300: New 72-GPU AI Servers to Ship in 2025

Foxconn, Other Taiwanese Makers Speed ​​Up Assembly of Blackwell Ultra Systems as Supply Chain Pressure Mounts and Alternatives to NVIDIA Chips Emerge

Meta Centralizes AI: Superintelligence Labs is Born

Led by Alexandr Wang and Nat Friedman, the new division unifies advanced AI research and development, with billions in investments, strategic acquisitions and an aggressive recruiting campaign

Baidu Opens Ernie 4.5 Code, Relaunches Challenge in Chinese AI

With ten models published on Hugging Face, the Beijing giant adopts the open source path to remain competitive in the AI ​​ecosystem, driven by pressure from agile startups and the growing centrality of transparency in the sector

Oasis: The AI ​​Model That Generates Interactive Worlds in Real Time

Created by Etched and Decart, Oasis processes every frame in response to user input, supported by an optimized transformer architecture and the new Sohu chip for high-definition, low-latency AI video

Microsoft Introduces AI System Capable of Outperforming Doctors in Complex Diagnoses

AI in Banking: “Digital Colleagues” Support Human Staff

BNY Mellon and JPMorgan deploy intelligent agents with autonomous, manager-supervised enterprise logins for operational tasks such as coding and validation

Meta AI Accesses Your Camera Roll: New Features to Edit Photos with AI

Facebook is testing a feature that uploads unpublished images to the cloud to offer creative suggestions via AI, analyzing faces, dates and objects. The option can only be activated with consent and deactivated in the settings

XVerse: Precise control of identities and attributes in multi-subject text-image models

Thanks to a token-specific modulation based on reference images, XVerse enables the consistent and customizable generation of complex scenes with multiple subjects, maintaining high visual quality and independence between features

Denmark, protection extended to digital identity: copyright on face, body and voice

A proposed law aims to tackle deepfakes by giving citizens the right to remove unauthorized content and obtain compensation, with exemptions for satire and fines for non-compliant platforms

Microsoft Slows Down on AI Chips: Braga to Arrive in 2026

Internal project suffers delays due to technical issues and team turnover, while projected performance fails to match NVIDIA’s Blackwell chips, accentuating dependence on external suppliers