Dernière minute
ARحسام حسن: من أساطير اللاعبين إلى مدرب مثير للجدل.. هل يحقق المفاجأة أمام الأرجنتين؟ESIncendio forestal en Sierra de Grazalema: desalojan a más de cien personas y elevan la emergencia a nivel 1INTLNigeria Warns of 'All Options' as African Immigrant Safety Deteriorates in South AfricaRUВ Нижегородской области введут продажу бензина по четным и нечетным днямINDoda-Kishtwar Highway Blocked by Landslide Near Hydroelectric ProjectBRFérias Funesc oferece programação gratuita para crianças em 4 cidades da ParaíbaFRChristophe Gleizes: His family hopes for a pardon from the Algerian presidentDEMicrosoft baut weitere 4800 Stellen wegen KI abRUSudan Pursues 'Alternative Economy' Strategy to Counter SanctionsESAsesinan a balazos a cantante de regional mexicano Adán R. en CuliacánARحسام حسن: من أساطير اللاعبين إلى مدرب مثير للجدل.. هل يحقق المفاجأة أمام الأرجنتين؟ESIncendio forestal en Sierra de Grazalema: desalojan a más de cien personas y elevan la emergencia a nivel 1INTLNigeria Warns of 'All Options' as African Immigrant Safety Deteriorates in South AfricaRUВ Нижегородской области введут продажу бензина по четным и нечетным днямINDoda-Kishtwar Highway Blocked by Landslide Near Hydroelectric ProjectBRFérias Funesc oferece programação gratuita para crianças em 4 cidades da ParaíbaFRChristophe Gleizes: His family hopes for a pardon from the Algerian presidentDEMicrosoft baut weitere 4800 Stellen wegen KI abRUSudan Pursues 'Alternative Economy' Strategy to Counter SanctionsESAsesinan a balazos a cantante de regional mexicano Adán R. en Culiacán
Newsgather
BackAnthropic's Powerful Claude Mythos AI Not Released, Security Tool Launched Instead
Anthropic's Powerful Claude Mythos AI Not Released, Security Tool Launched Instead
DERNIÈRE MINUTE
NPR Business21.05.2026TechUnited States

Anthropic's Powerful Claude Mythos AI Not Released, Security Tool Launched Instead

AIAnthropicClaude Mythos

L'essentiel

  • Anthropic decided not to release its powerful AI model, Claude Mythos, to the public.
  • Instead, the company launched a security tool aimed at addressing software vulnerabilities.

Résumé généré par IA

Taille de police

Anthropic decided not to release its powerful AI model, Claude Mythos, to the public. Instead, the company launched a security tool aimed at addressing software vulnerabilities.

Read the full article on NPR Business

Sujets liés

AI
SujetsAnthropicClaude Mythossecurity toolsoftware vulnerabilities
This article was originally published by NPR Business.

Articles liés

Anthropic Removes Secret Tracker After Privacy Breach in Claude Code
En développement·1 sa önce

Anthropic Removes Secret Tracker After Privacy Breach in Claude Code

Anthropic removed a secret tracker from Claude Code after a researcher exposed it. The AI firm used "prompt steganography" to hide code that flagged Chinese users' timezone, proxy, and potential ties to Chinese AI labs accused of distillation attacks. Anthropic claims it was an experiment to prevent account abuse and distillation, but privacy advocates view it as a serious breach of trust.

Ars Technica
Plus sur ce sujetAI