Olhar Digital News presented the main events of last Thursday (July 9).
Olhar Digital News presented the main events of last Thursday (July 9).
The National Health Surveillance Agency (Anvisa) issued a new regulatory instruction that requires manufacturers to update the composition of Covid-19 vaccines used in Brazil. Immunizers must now be monovalent and adapted to combat the latest circulating strains of coronavirus.
The European Union may approve new restrictions or even ban minors from accessing social media before September. Pressure on the European Commission has increased after Australia banned access to digital platforms for individuals under 16 at the end of 2025.
OpenAI introduced the GPT-5.6 family this Thursday. This set includes three models with different power and cost characteristics: Sol, the most powerful, designed for complex tasks; Terra, balanced for daily use; and Luna, the fastest and most affordable.
Meta plans to begin its own artificial intelligence chip production in September of this year. A device called 'Iris' is part of a four-generation project and was developed internally to reduce dependence on suppliers such as Nvidia and AMD.
Planetary defense remains one of the main technological challenges of modern times. NASA's Dart mission has already demonstrated the possibility of deflecting a small asteroid. Currently, another concept for preventing giant asteroids from colliding with Earth is being developed in China.
Olhar Digital News will be published Monday through Friday on their social networks.
OpenAI plans to make its latest version of ChatGPT publicly available this Thursday, the 9th. The announcement was made by the startup itself on the X platform, detailing the launch of the GPT-5.6 model, along with the Terra and Luna models.
GPT-5.6 is presented as the developer's most advanced model. Previously, the presentation of this technology had been postponed due to requests from the United States government, motivated by growing White House concerns regarding national security and the misuse of sophisticated artificial intelligence systems. However, according to Axios, the Trump administration gave the green light for the large-scale launch.
Initially, access was restricted to a select group of partners whose data was shared with authorities. Three distinct models will be introduced: Sol, which is the most powerful; Terra, designed to be more accessible; and Luna, focused on offering high capability at a low cost.
In a preview released on the OpenAI blog, the company specified that the Terra model performs comparably to GPT-5.5 but costs half as much, while Luna ensures strong functionality at the lowest price. GPT-5.6 Sol incorporates the most robust security suite ever developed by the company. Protections against high-risk uses, sensitive cyber requests, and repeated misuse were reinforced after weeks of rigorous testing to identify and mitigate vulnerabilities against real attacks.
OpenAI also classified GPT-5.6 Sol as its 'strongest model to date,' highlighting its 'enhanced agentic capabilities in programming, biology, and cybersecurity.'
In the United States and China, there is an intense technological dispute where AI plays a central role. Advanced artificial intelligence models have the potential to accelerate cyberattacks in decades-old sectors with highly interconnected infrastructures. For this reason, the White House actively monitors new AI models to detect potential threats, fearing that the technology could be used for military purposes by other nations, such as China and Russia.
In China, the governmental approach remains similar: the government is in dialogue with major industry companies to limit foreign access to the cutting-edge models developed in the country, as reported by Reuters.
It is relevant to mention that in June, President Donald Trump signed an executive order establishing a voluntary benchmark, allowing the federal government to evaluate new AI systems before they are launched. Furthermore, the previous week, Anthropic's Fable 5 and Mythos 5 models became available again after the United States government suspended export controls applied to the startup. This change occurred less than three weeks after the restrictions were imposed for national security reasons.
The initial restrictions, set on June 12, forced Anthropic to deactivate both models for all users, as there was no way to verify the nationality of each individual in real time. After implementing new safeguards, the company obtained permission to resume access. Mythos 5 was gradually released to national and international partners, while Fable 5, intended for the general public, has been available since the 1st.
Following a similar logic, OpenAI chose not to make GPT-5.6 available to all users simultaneously.
