Meta's Superintelligence Lab released Muse Wonder and Muse Wonder Full, its first frontier-class models, claiming benchmark scores comparable to Fable 5.1 and GPT-6 Astra, including stronger results on coding tasks. Wonder will be priced at $4 per million input tokens and $20 per million output tokens via Meta's Model API, while Wonder Full is reserved for Muse Code subscribers and won't be offered through the API; both will power the Muse agent.
techcrunch.com
· 2026-10-12
Meta unveiled two new AI models, Muse Wonder and Muse Wonder Full, which the company says deliver frontier-level capabilities. The models will be accessible via an API and through Muse Code, a coding tool, and the top version will power features across Meta's platforms including its Muse personal assistant. Meta says the largest model outperforms Anthropic's Claude Fable 5.1, though it did not disclose training or architecture details.
wired.com
· 2026-10-12
Meta has released two new AI models, Muse Wonder and Muse Wonder Full, available immediately to Muse Code subscribers and accessible via the Meta Model API. The company says internal benchmark testing shows the models are competitive with leading systems from OpenAI and Anthropic, though it did not compare them to Google's newest model. Meta plans to eventually use Wonder to power its Muse AI agent, starting with subscribers, though no timeline was given.
theverge.com
· 2026-10-12
Microsoft CEO Satya Nadella posted on X that AI systems should no longer be treated as opaque 'black boxes' whose outputs are simply trusted or rejected. He called for containment mechanisms, including an 'emergency brake' that lets an authorized person pause or shut down a model mid-task, plus incident disclosure, independent audits and verifiable data. Nadella said more advanced models will need standardized, more advanced containment technologies.
theverge.com
· 2026-10-10
Speaking at the United Nations in late September, President Trump said the term 'artificial' made AI sound fake and announced his administration would start referring to the technology as 'superintelligence,' or 'SI.' The remark has reignited a long-running debate over how AI should be labeled and understood.
fastcompany.com
· 2026-10-10
ICANN has begun accepting applications for new top-level domains (TLDs), with 1,615 submissions from 481 applicants. Notably, several tech companies, including Meta and OpenAI, applied for domains like .agent and .agi, reflecting the growing influence of artificial intelligence. Applications also include generic domains such as .geek and .anime, alongside numerous branded TLDs like .facebook and .zillow.
theverge.com
· 2026-10-07
On September 29, President Donald Trump issued an executive order stating that humanity has entered a new era of 'Super Intelligence,' urging federal agencies to replace 'artificial intelligence' with the term 'Super Intelligence' and its abbreviation 'SI.' The order claims current systems surpass simple automation and imitation, though the term 'superintelligence' has long been used in academic circles to describe highly advanced, cognitive systems exceeding human capabilities in most domains. Researchers note that the precise definition of superintelligence remains unsettled, especially regarding what constitutes human-level intelligence.
nature.com
· 2026-10-07
Elon Musk announced that SpaceXAI will be renamed SpaceXSI to align with President Trump’s recent executive order, which rebranded 'Artificial Intelligence' as 'Super Intelligence.' The order was signed on September 29, 2026, and aims to better reflect the technology's potential. Musk confirmed the change via an X post on September 4, shortly after Trump launched the 'Super Intelligence Force' to oversee AI policy and promote American dominance in the field.
tomshardware.com
· 2026-10-06
Analysis of existing AI capabilities finds that today's models, not some future superintelligence, are already lowering barriers to fraud, cyberattacks, manipulation campaigns, and weapons development. The concern centers on present-day misuse rather than speculative future risks like AI takeover scenarios.
fastcompany.com
· 2026-10-05
In a Guardian op-ed, researchers Timnit Gebru and Emily M. Bender cite an unverified CNN report claiming U.S. officials nearly intercepted a Chinese vessel after a chatbot-generated intelligence report falsely stated it carried nuclear-weapons components. The plan was reportedly halted only after officials double-checked the AI-generated claim and found it erroneous.
tech.slashdot.org
· 2026-10-02
President Trump announced a 'morally binding' agreement called the Joint Commitment On Frontier Responsibilities, signed by executives including Elon Musk, Sam Altman's OpenAI cofounder Greg Brockman, Dario Amodei, Mark Zuckerberg, Sundar Pichai and Jensen Huang. The deal asks companies to maintain internal safety controls and work with external auditors, and Trump said a board would oversee compliance, though no members were named. Trump also signed an executive order renaming artificial intelligence as 'superintelligence' and reaffirmed his backing of data center expansion.
techspot.com
· 2026-09-30
Palisade Research, a nonprofit studying AI capabilities, published a series of video interviews with current and former researchers from OpenAI, Google DeepMind and Anthropic on a site called frominside.ai. In the videos, figures including Geoffrey Irving and Neel Nanda estimate the odds of AI causing human extinction at roughly 50% and at least 10% respectively, while former OpenAI researcher Daniel Kokotajlo describes superintelligent AI as potentially 'god-like powerful' and uncontrollable.
theverge.com
· 2026-09-29
GoKawiil