Meta Releases Open-Weight Muse Spark Model for Coding, Agentic Tasks

The logo of Meta is seen at the entrance of the company's temporary stand ahead of the World Economic Forum (WEF) in Davos, Switzerland January 18, 2025. (Reuters)
The logo of Meta is seen at the entrance of the company's temporary stand ahead of the World Economic Forum (WEF) in Davos, Switzerland January 18, 2025. (Reuters)
TT

Meta Releases Open-Weight Muse Spark Model for Coding, Agentic Tasks

The logo of Meta is seen at the entrance of the company's temporary stand ahead of the World Economic Forum (WEF) in Davos, Switzerland January 18, 2025. (Reuters)
The logo of Meta is seen at the entrance of the company's temporary stand ahead of the World Economic Forum (WEF) in Davos, Switzerland January 18, 2025. (Reuters)

Meta ‌Platforms on Thursday released an open-weight version of its Muse Spark artificial intelligence model for coding and AI agents, expanding developers' ability to download and adapt the technology.

Open-weight models make trained parameters available for developers to modify, unlike proprietary models ‌that remain under ‌their creators' control. ‌They ⁠can offer businesses ⁠a lower-cost alternative to closed systems.

Meta said Muse Spark has 816 billion total parameters, with 42 billion active at a time, and ⁠can handle a context window ‌of up ‌to 1 million tokens.

The ‌model can run locally on ‌two Apple M5 Ultra Mac Studio computers and is also available through some API providers.

CEO ‌Mark Zuckerberg has argued that advanced AI should ⁠not be ⁠concentrated among a small number of companies.

While open-weight models can give companies greater control over their data, they can require significant computing resources and technical expertise to operate at scale.

In August, Meta released Muse Glimmer, a smaller open-weight model.



Taiwan Chip Giant TSMC Says Net Profit Hit Record in 3rd Quarter

FILE PHOTO: File Photo: The logo of Taiwan Semiconductor Manufacturing Company (TSMC) at the company's second-quarter earnings conference in Taipei, Taiwan, July 16, 2026. REUTERS/Ann Wang/File Photo/File Photo
FILE PHOTO: File Photo: The logo of Taiwan Semiconductor Manufacturing Company (TSMC) at the company's second-quarter earnings conference in Taipei, Taiwan, July 16, 2026. REUTERS/Ann Wang/File Photo/File Photo
TT

Taiwan Chip Giant TSMC Says Net Profit Hit Record in 3rd Quarter

FILE PHOTO: File Photo: The logo of Taiwan Semiconductor Manufacturing Company (TSMC) at the company's second-quarter earnings conference in Taipei, Taiwan, July 16, 2026. REUTERS/Ann Wang/File Photo/File Photo
FILE PHOTO: File Photo: The logo of Taiwan Semiconductor Manufacturing Company (TSMC) at the company's second-quarter earnings conference in Taipei, Taiwan, July 16, 2026. REUTERS/Ann Wang/File Photo/File Photo

Taiwanese chip giant TSMC said Thursday that net profit soared 68.7 percent to a record high in the third quarter, buoyed by massive global demand for AI hardware.

The result marked the sixth straight record quarterly net profit for the company, the world's biggest contract maker of microchips, including those used in Nvidia AI data center processors.

But with the AI boom only getting bigger, analysts warn that chipmakers and other tech firms reporting bumper financial results risk being "judged against perfection" by investors.

Meanwhile the company's planned expansion into the United States, including building chip plants in Arizona, is also likely to shape its future earnings.

TSMC's net profit for July to September surged year-on-year to NT$763.1 billion (US$23.6 billion), beating analyst estimates of NT$739.5 billion.

The result surpassed the company's previous quarterly record of NT$706.6 billion set in April-June 2026.

Quarterly revenue rose 50.9 percent to NT$1.5 trillion.

"Our business was supported by strong demand for our leading-edge process technologies," AFP quoted the company as saying in a statement.

Governments and tech giants are pouring eye-watering sums into building data centers that can train and run AI tools such as image generators and agents that can execute tasks.

But concerns about overstretched valuations have fueled fears of a market bubble, along with questions over when the massive investments into AI will reap returns.

TSMC's share price has ballooned nearly 60 percent this year, adding pressure on its earnings announcement as investors ask how long the good times can last.

TSMC builds chips for the likes of Nvidia, Apple, AMD and Broadcom -- meaning it "sits at the very center of the AI ecosystem", Zavier Wong, Market Analyst at eToro, told AFP ahead of the earnings announcement.

"Every leading AI chip still runs through its factories, so for investors, this week's results are about as close as it gets to a health check on the entire AI build-out."

Fears over AI's dangers have intensified in recent weeks, leading to calls for leading US labs to deliberately slow down their technological advancement.

"If the biggest buyers start to pace their spending, whether because of funding costs or AI labs slowing the development of new models, chipmakers would feel it quickly," Wong said.

TSMC is not the only chip giant enjoying the benefits of the AI boom.

Last week, South Korea's Samsung Electronics forecast a more than 780 percent leap in third-quarter operating profit from a year earlier.

But Samsung's shares still fell because its forecast missed expectations.

"Record profits have become the new normal, and Nvidia has lived with this for years, judged against perfection," Wong said.

Often "investors skip straight past the headline numbers and pick through the guidance and the fine print for any warning signs.”


Mexico Gives Social Networks One Month to Remove Violent Content

FILE PHOTO: Mexico's President Claudia Sheinbaum speaks during her daily conference in Mexico City, Mexico, May 29, 2026. REUTERS/Quetzalli Nicte-Ha/File Photo
FILE PHOTO: Mexico's President Claudia Sheinbaum speaks during her daily conference in Mexico City, Mexico, May 29, 2026. REUTERS/Quetzalli Nicte-Ha/File Photo
TT

Mexico Gives Social Networks One Month to Remove Violent Content

FILE PHOTO: Mexico's President Claudia Sheinbaum speaks during her daily conference in Mexico City, Mexico, May 29, 2026. REUTERS/Quetzalli Nicte-Ha/File Photo
FILE PHOTO: Mexico's President Claudia Sheinbaum speaks during her daily conference in Mexico City, Mexico, May 29, 2026. REUTERS/Quetzalli Nicte-Ha/File Photo

Mexican President Claudia Sheinbaum on Wednesday ordered Meta, TikTok, X, Apple and Microsoft to remove content inciting violence on their social networks following a spate of deadly attacks in schools.

Sheinbaum gave the platforms a month to comply or face sanctions.

In the past two weeks, two people have been killed and several people injured in two separate school attacks.

The attacks have caused alarm in a country that lives in terror of brutal drug cartels but is a relative stranger to shootings or stabbings in schools.

On October 1, two teenage brothers armed with knives and fireworks killed a female assistant principal and injured three other people at a school in Coahuila state.

Then on Tuesday, a teacher was shot dead at a school in the southern state of Oaxaca, AFP reported.

Sheinbaum linked the violence to "behavior associated with communication on social media."

She said that if the tech companies did not "adjust their algorithms" to remove "all violent content linked to the accounts of children and adolescents," she would legislate to impose sanctions on them.

The decree issued by Sheinbaum on Wednesday orders social networks to prevent minors accessing "images and videos that lead to or encourage violence in a school setting," including "abuse, suicide, self-harm, cyberbullying and hate speech."

The networks must alert the authorities "if they identify content that endangers life."

In a discussion with TikTok and Instagram users, Sheinbaum denied that the new rules amounted to censorship.

"We're not banning social media," she said.


Eight US States Sue OpenAI, Claiming ChatGPT Harms Young Users

A keyboard is placed in front of a displayed OpenAI logo in this illustration taken February 21, 2023. (Reuters)
A keyboard is placed in front of a displayed OpenAI logo in this illustration taken February 21, 2023. (Reuters)
TT

Eight US States Sue OpenAI, Claiming ChatGPT Harms Young Users

A keyboard is placed in front of a displayed OpenAI logo in this illustration taken February 21, 2023. (Reuters)
A keyboard is placed in front of a displayed OpenAI logo in this illustration taken February 21, 2023. (Reuters)

A bipartisan group of eight states said on Wednesday they had sued OpenAI in courts across the country over harms they say ChatGPT has caused to their residents, alleging the chatbot has coached its users — including teenagers — to harm themselves or others.

Attorneys general from the states, which include Iowa, Kansas and Arizona, filed the lawsuits independently in state courts, claiming ChatGPT is designed ‌to be ‌agreeable and engaging to earn the trust ‌of ⁠users. But the ⁠chatbot reinforces delusions, encourages dangerous behavior and assists users in planning violence like mass shootings, the lawsuits claim.

The lawsuits represent the most significant coordinated state action against an artificial intelligence company, adding to a wave of legal challenges OpenAI faces as it ⁠prepares for a public listing. Florida's attorney ‌general filed the first ‌such state suit in June.

The lawsuits accuse the company of ‌violating state consumer protection laws by misleading users ‌about the safety of ChatGPT. The states also allege that the company created a public nuisance by endangering the health of their residents.

The lawsuits seek orders forcing OpenAI ‌to make changes to how ChatGPT works. Among other things, the states want ⁠the company barred ⁠from developing AI models that can improve themselves unless it can ensure those models are aligned with human values.

The states, which also include Arkansas, Idaho, Montana, Nebraska and West Virginia, are also seeking civil monetary penalties for OpenAI's violations of state law as well as damages.

A spokesperson for OpenAI did not immediately respond to a request for comment.

The company is also facing dozens of lawsuits by private plaintiffs, alleging ChatGPT helped users plan mass shootings or other violence.