Model

• El modelo de IA Mythos de Anthropic, que la empresa consideró demasiado peligroso para su lanzamiento público, presuntamente ha sido accedido por un tercero no autorizado en una brecha de seguridad significativa. • El incidente ha generado preocupaciones globales sobre la seguridad de la IA y los protocolos de seguridad que rodean a los sistemas de IA avanzados, particularmente aquellos modelos marcados como de alto riesgo por sus desarrolladores. • La brecha subraya las tensiones actuales entre las consideraciones de seguridad de la IA y las vulnerabilidades de ciberseguridad en la gestión de modelos de lenguaje de vanguardia.

• OpenAI announced a new AI model specifically engineered to enhance the performance and capabilities of other AI systems. • The development positions OpenAI as a leader in meta-AI tools that optimize existing models for better efficiency and functionality. • This breakthrough could accelerate AI advancement cycles, enabling faster iterations and broader applications across industries.

• The Trump administration pledges a crackdown on Chinese companies exploiting American-made AI models, focusing on national security risks. • Officials singled out China for foreign tech firms' unauthorized use of US AI technologies, vowing stricter enforcement measures. • The initiative aims to protect US intellectual property and prevent adversarial nations from leveraging American innovations in AI development.

• Meta released Llama 4 on April 20, 2026, a 2 trillion parameter multimodal AI model trained on 15 petabytes of US-centric data. • The model outperforms GPT-4o on benchmarks like MMLU (92% score) and supports vision-language tasks for enterprise apps. • Open-sourcing aims to spur US AI innovation, with downloads exceeding 500,000 in first hours via Hugging Face.

• Anthropic announced a new frontier AI model on April 16, 2026, capable of advanced performance that sent shockwaves through the tech industry. • Investors reacted negatively, resuming an AI-driven slump in software stocks as the model's capabilities raised concerns over market dynamics. • The release highlights ongoing US AI breakthroughs and their immediate impact on Wall Street valuations.

• Researchers published the most thorough study of cosmic expansion in the journal Astronomy & Astrophysics on April 10, combining decades of research to derive the most precise Hubble constant yet while confirming that tension between early and late-universe measurements persists. • The study, launched at an International Space Science Institute Breakthrough Workshop in Bern, Switzerland in March 2025, expanded the cosmic distance ladder into a comprehensive survey called the Local Distance Network, achieving a goal considered 'potentially unreachable' a decade ago. • Study co-author Richard Anderson from the University of Göttingen stated that the Hubble tension reveals that 'something's missing' in the current understanding of basic physics on cosmological scales.

• OpenAI Group PBC announced the launch of GPT-5.4-Cyber on April 14, 2026, a fine-tuned variant of its GPT-5.4 model specifically designed for defensive cybersecurity tasks. • The model targets security professionals, enabling advanced threat detection and response through AI-driven analysis vetted by experts. • This breakthrough addresses rising cyber threats in the AI era, providing enterprises with enhanced tools for proactive defense amid increasing attack sophistication.

• US Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell convened Wall Street leaders on April 7 to warn about Anthropic's new Mythos AI model, which can identify software vulnerabilities that have evaded decades of human review and millions of automated security tests. • Mythos is being released only to carefully selected partners for defensive security work, as Anthropic fears the tool could provide ransomware gangs and hostile governments with powerful weapons to steal data or disrupt critical infrastructure. • The model represents a significant leap in AI cybersecurity capabilities, with the potential to both protect systems through accelerated penetration testing and pose risks if misused by malicious actors.

Reports say Fed chair Jerome Powell among attenders at meeting in Washington The US Treasury secretary, Scott Bessent, summoned major American bank chiefs to a meeting in Washington this week amid concerns over the cyber risks posed by Anthropic’s latest AI model, according to reports.Bosses including the Federal Reserve chair, Jerome Powell, were said to have gathered at the Treasury headquarters for the meeting after the release of the Claude Mythos AI model that Anthropic says poses unprecedented cybersecurity risks. Continue reading...

AI company says purpose of its Claude Mythos model is to bolster defenses against hacking in common applicationsAnthropic on Tuesday said its yet-to-be-released artificial intelligence model called Claude Mythos has proven keenly adept at exposing software weaknesses.Mythos has laid bare thousands of vulnerabilities in commonly used applications for which no patch or fix exists, prompting the San Francisco-based AI startup to form an alliance with cybersecurity specialists to bolster defenses against hacking and withhold wide distribution. Continue reading...

• OpenAI, Anthropic, and Google's Alphabet began sharing information via the Frontier Model Forum to detect Chinese firms' adversarial distillation of US AI models, violating terms of service. • US officials estimate this unauthorized copying costs Silicon Valley labs billions in annual profits, with OpenAI accusing DeepSeek of free-riding on American innovations. • The rare cooperation highlights national security risks and competitive threats in the global AI race, as imitation models undercut prices and siphon customers.

• Google open-sourced Gemma 4 AI model under Apache 2.0 license, enabling broad customization and commercial applications. • The release advances accessible AI development following prior Gemma iterations. • Model strengthens Google's position in open source AI competition with rivals like Arcee.

• CMS announced the Accelerating State Pediatric Innovation Readiness and Effectiveness Model (ASPIRE), a 10-year voluntary program for up to five states. • ASPIRE targets Medicaid and CHIP enrollees up to age 21 with complex medical and behavioral needs through early identification, care coordination, and quality accountability. • A Notice of Funding Opportunity will be released later this year to solicit state participation.

• Moody's AI-based recession probability model has surged to 49%, just below the 50% threshold that historically signals economic recession within a year, amid mounting economic concerns. • The US labor market delivered a major disappointment with March data showing a loss of 92,000 jobs, far below expectations, while GDP growth projections have been revised downward from 1.4% to as low as 0.7%. • Recession forecasts from major institutions have escalated: Goldman Sachs raised odds to 30%, EY-Parthenon estimates risk at around 40%, and BlackRock CEO Larry Fink warned rising oil prices could trigger global recession.

Sara Ziff, founder of Model Alliance, said business leaders need to be hauled before House oversight committeeA top modeling industry activist has called for business leaders to be hauled before lawmakers in Washington to investigate what role modeling agencies may have played in the Jeffrey Epstein sex trafficking scandal.Sara Ziff is founder of the Model Alliance, a non-profit advocacy group calling for fair treatment, labor rights and safe working conditions for fashion industry workers. Continue reading...

Former talent agency boss had closer relationship with sex offender than thought, and supported him after 2009 arrestA female executive at the top of the modelling industry had a close friendship with Jeffrey Epstein and introduced him to women on the agency’s books, a Guardian investigation has found.Until last November, Faith Kates ran Next Management modelling and talent agency, which has represented the likes of Alexa Chung, Milla Jovovich and Billie Eilish, a position she held for decades as the founder of the business. She stepped down quietly just weeks before the first major Epstein files were released, saying she intended to focus on charity work.18 July 2009 10.18amI am and will always be your friend...Unconditionally...will always be there for you.5 September 2009 7.47pmThinking of you a lot and hoping you are finally enjoying some please [sic] and quiet..know you are always in my thoughts and prayers. You are a good friend my dear friend..5 September 2009 7.54pmthanks,, lets get back to work. Continue reading...

• DOE's SLAC National Accelerator Lab and collaborators created a generative AI model that rebuilds molecular structures from fragmentation patterns during explosions. • The breakthrough enhances analysis of molecular dynamics using accelerator data, advancing chemistry and materials science research. • Published within the last 8 hours, it promises novel insights into molecular behaviors under extreme conditions.

• Scientists at Sanford Burnham Prebys Medical Discovery Institute developed TLPath, a new computational tool that infers telomere length—repeating DNA sections at chromosome ends that shorten with age—by detecting structural changes in cells captured in routine biopsy images. • The model successfully predicted telomere length more accurately than using patient age alone and could identify telomere length differences between individuals of the same chronological age, potentially transforming large-scale aging research. • Published in Cell Reports Methods, the findings enable researchers to measure an important aging metric quickly and cheaply without costly traditional tests, opening new opportunities to study telomere biology and human aging.

Wall Street closed lower on Friday with all three major indexes ending the week in the red, while the Russell 2000 hit its lowest level of the year, despite strong corporate earnings and solid guidance reportedly supporting broader market trends. Meta Platforms shares fell nearly 4% following a report that the social media giant has postponed the release of its AI model Avocado to at least May, signaling potential delays in the company's AI product roadmap. Ulta Beauty shares slumped more than 14% as rising costs deepened margin concerns, with the company's new CEO flagging potential global cost impacts. The mixed market performance reflects investor uncertainty about technology sector growth prospects amid ongoing economic and competitive pressures.

Meta delayed rollout of its new AI model code-named Avocado after internal tests showed it lagging competitors in reasoning, coding, and writing, reported on March 13, 2026. The New York Times detailed the model's underperformance against benchmarks. This setback comes amid intense AI competition and highlights challenges in model development. Meta must refine the model to catch up in the race for advanced capabilities.

Meta delayed its new AI model, code-named Avocado, after internal tests showed it lagging competitors in reasoning, coding, and writing, reported on March 13, 2026. The setback impacts Meta's AI ambitions amid intense industry competition. Engineers are iterating to close performance gaps before public release. This highlights challenges in achieving frontier AI capabilities quickly.

Meta delayed the rollout of its Avocado AI model after internal tests revealed it underperformed competitors in reasoning, coding, and writing, reported on March 13, 2026. The New York Times cited concerns over the model's benchmarks against rivals. This setback impacts Meta's AI ambitions amid intensifying competition. Analysts expect revisions before any public release.

Scientists at Texas A&M University created the most challenging AI test to date, involving over 50 collaborators, to evaluate advanced language models on complex reasoning tasks. Initial results showed even top AI systems struggling significantly, highlighting gaps in current capabilities despite acing simpler benchmarks. The benchmark, detailed in a peer-reviewed paper, pushes boundaries in AI safety and robustness research. Future iterations will incorporate multimodal challenges to further stress-test emerging models.