Anthropic announced this Tuesday the expansion of Project Glasswing. The program now includes 150 additional organizations in more than 15 countries with access to the Mythos model. Essa initiative focuses on identifying vulnerabilities in critical software.
The measure expands the initial group of around 50 partners. Setores as energy, water, health, communications and hardware are reinforced. New organizations must meet strict security requirements before using the template.
Expansão targets underrepresented sectors in early stage
Project Glasswing seeks to strengthen the security of critical infrastructures. Anthropic highlighted that the first phase revealed more than 10,000 high or critical severity flaws. Parceiros as well as Apple, Nvidia, Microsoft, CrowdStrike and Palo Alto Networks participated in the initial tests.
- Energia and utilities
- Sistemas water and sanitation
- Hospitais and healthcare networks
- Operadoras telecommunications
- Hardware Fabricantes
Esses fields receive greater attention now. The model helps map and fix problems in widely used code.
Mythos demonstrates fault detection capabilities
Mythos excels at analyzing complex code. Ele identifies vulnerabilities and suggests fixes. The company estimates that a major cyberattack could affect more than 100 million people.
Especialistas already warned about the potential of the model. Hackers could exploit its functions to attack systems. Therefore, Anthropic has opted for controlled access since its launch in April.
The program has already generated concrete results. Initial partners have tested Mythos in their own repositories and open source projects. Muitos reported significant gains in patching holes.
Empresa advances with IPO and offering plans on União Europeia
Anthropic confidentially filed for an IPO with the SEC on Monday. The Mythos expansion comes one day after offering model access to União Europeia. Essas shares show the startup’s growth momentum.
The company’s official blog reinforces the long-term commitment. The objective is to make all software more secure in the face of the advancement of artificial intelligence. The company also plans to support open source projects with usage credits.
Global Parceiros assess impact on cybersecurity
The inclusion of organizations from different regions expands the reach of the program. Países allies and governments participate in discussions to define priorities. The focus remains on critical infrastructures and systems that affect millions of users.
Anthropic did not reveal the names of the new 150 organizations. The main criterion continues to be the level of security and relevance of the systems maintained by each company.
- Requisitos include internal process auditing
- Compromisso with responsible use of the model
- Relatórios journals about faults found
- Colaboração in patches for critical issues
Essa framework helps mitigate risk while expanding access.
Detalhes model technical and next steps
Mythos Preview operates with a focus on cybersecurity tasks. Ele analyzes code, detects flaws and proposes testable solutions. Diferente of general models, he received specific training for this type of application.
The company allocated up to US$100 million in usage credits for the project. Outros $4 million goes to open source security organizations. Esses features support the current phase of expansion.
Project Glasswing continues to evolve. Anthropic plans to work with governments and partners to include more entities. The program must prioritize systems that support essential day-to-day services.
