Meta Removes News Scraping Terms; Nvidia's Profits Amid Debt Concerns

SpaceXAI appeared before the New York City Council Committee on Technology to discuss AI risks after missing an earlier hearing and being subpoenaed. Manhattan Councilwoman Carmen De La Rosa led the discussion, which focused on potential dangers posed by artificial intelligence technology. The company behind the Grok AI chatbot faced scrutiny from lawmakers concerned about its AI practices.

Meta updated its terms of service after backlash from publishers and legal experts. The company removed language that experts said allowed scraping of news websites to train AI models. The updated terms take effect Oct. 30, though Meta spokesperson Daniel Roberts said the changes do not alter existing crawling practices that follow robots.txt signals.

Steve Eisman warned that OpenAI's collapse could trigger an immediate US recession due to the massive capital flowing into AI from cloud providers, chip companies, and infrastructure firms. He described the AI boom as a dangerous concentration of risk, noting that cheaper open-weight models are creating a price war and companies are reinning in spending.

Nvidia faced pressure alongside other chip stocks after CoreWeave bond yields spiked. The data center company, which has $72 billion in liabilities and 51 data centers, saw its bond prices crash as investors lost confidence. Nvidia remains highly profitable with $192.88 billion in net income over the past 12 months and about 65% operating profit margins. Its Vera Rubin platform is in production at Microsoft Azure, Google Cloud, and Oracle.

Featherless AI showed that a five-developer team cut its monthly token bill from $150,000 to under $10,000 by running 116 billion tokens on a single node using 512 AMD MI325 GPUs. CEO Eugene Cheah recommended splitting API keys by workload, right-sizing models by task type, and using cheaper open models for repetitive tasks to reduce costs significantly.

AI hiring tools like HireVue face a paradox when it comes to preventing bias. Used by over 60% of the top 100 US companies and eight of the ten largest federal agencies, the tools use large language models to evaluate candidates. Experts warn that AI algorithms can make biased decisions look objective, while human decision-making itself can also be biased.

Meta AI CEO Alexandr Wang celebrated the launch of Muse Wonder and open-source Muse Spark, crediting the team at Meta Superintelligence Labs. The lab was formed just over a year ago with the goal of building a frontier AI model, and Wang said it has built a repeatable system for creating new frontier models despite initial doubts.

AI data centers are becoming a contentious issue in Florida's District 19 congressional race. A Fox News poll showed 71% of residents oppose data centers in their communities. Democrat David Jolly said he would impose a moratorium on data centers if elected governor, while Republican Jim Schwartzel argued AI data centers can protect water and create jobs.

Key Takeaways

  • SpaceXAI testified before NYC Council on AI risks after being subpoenaed for skipping an earlier hearing
  • Meta removed terms allowing news site scraping for AI training, effective Oct. 30
  • Steve Eisman warns OpenAI collapse could trigger US recession due to concentrated AI risk
  • Nvidia earned $192.88 billion in net income over 12 months despite CoreWeave debt concerns
  • CoreWeave, with $72 billion in liabilities and 51 data centers, saw bond prices crash amid AI debt boom fears
  • Featherless AI cut monthly token costs from $150K to $10K using 512 AMD MI325 GPUs
  • HireVue AI hiring tools are used by over 60% of top 100 US companies despite bias concerns
  • Meta AI launched Muse Wonder and open-source Muse Spark after just over a year of development
  • 71% of Florida District 19 residents oppose AI data centers in their communities
  • Nvidia's Vera Rubin platform is in production at Microsoft Azure, Google Cloud, and Oracle

SpaceX testifies at NYC Council on AI risks

SpaceXAI appeared before the New York City Council Committee on Technology to discuss AI dangers. The company, which created the Grok AI chatbot, skipped an earlier hearing and only testified after being subpoenaed. Manhattan Councilwoman Carmen De La Rosa led the discussion. The hearing focused on potential risks posed by artificial intelligence technology.

NYC council questions SpaceX on AI safety

The New York City Council held a meeting on Friday about AI safety measures. SpaceX attended the meeting after missing a previous session and being subpoenaed. CBS News New York reported on the grilling of the company regarding its AI practices. The hearing aimed to address concerns about artificial intelligence risks.

Meta removes terms allowing news site scraping

Meta removed language from its terms of service that experts said allowed scraping of news websites to train AI. The change followed backlash from publishers and legal experts. The Verge first reported that the old terms let Meta use content from websites for its products and features. Meta clarified that web crawlers still follow robots.txt signals from site owners.

Meta updates terms over AI training concerns

Meta changed its commercial terms to remove provisions about using news content for AI training. The updated terms take effect on Oct. 30 and say automated crawling is for advertising and business products. Meta spokesperson Daniel Roberts said the changes do not alter existing crawling practices. Vice President Andy Stone said the update was meant to collect accurate product information for ads and recommendations.

Candidates debate AI data centers in Florida race

AI data centers are a major issue in Florida's District 19 congressional race. A Fox News poll showed 71% of residents oppose data centers in their communities. Democrat Victor Arias called it a local issue affecting Lee and Collier Counties. Republican Jim Schwartzel said AI data centers can protect water and create jobs. Democrat David Jolly said he would impose a moratorium on data centers if elected governor.

Steve Eisman Warns OpenAI Collapse Could Trigger US Recession

Steve Eisman says OpenAI has become so important to the US economy that its failure could cause an immediate recession. He pointed to the huge capital flowing into AI from cloud providers, chip companies, and infrastructure firms. If demand for AI services suddenly dropped, those suppliers would suffer serious harm. Eisman also warned that the AI industry faces a price war as cheaper open-weight models emerge and companies rein in spending. He described the AI boom as a dangerous concentration of risk.

Meta AI CEO Alexandr Wang Thanks Team for Muse Model Launch

Alexandr Wang, CEO of Meta AI, praised researchers and engineers after launching Muse Wonder and open-source Muse Spark. He said Meta Superintelligence Labs was formed just over a year ago with the bold goal of building a frontier AI model. Many people doubted the goal was possible, but the team worked with urgency and no set deadlines. Wang said the lab has built a repeatable system for creating new frontier models. He expressed pride in the team's rapid progress and scientific breakthroughs.

AI Hiring Tools Face Paradox of Preventing Bias

AI hiring tools like HireVue are used by over 60% of the top 100 US companies and eight of the ten largest federal agencies. Students such as Matt Segel have applied for internships through AI-conducted video interviews. HireVue uses a large language model to evaluate candidates based only on their answers, and a human recruiter makes the final decision. Experts warn that AI algorithms can make biased decisions look objective and scientific. They also note that human decision-making itself can be biased, creating a difficult paradox.

Nvidia Trades Below Usual Multiple Amid Customer Debt Concerns

Nvidia slipped along with other chip stocks after bond yields of CoreWeave, a cloud provider and Nvidia customer, spiked. Nvidia is both an investor in CoreWeave and a supplier of its chips. Despite this risk, Nvidia remains highly profitable, earning $192.88 billion in net income over the past 12 months with about 65% operating profit margins. In its fiscal second quarter of 2027, revenue reached $96.2 billion, up 106% year over year. The Vera Rubin platform is in production at Microsoft Azure, Google Cloud, and Oracle, and Nvidia is arranging financing with Apollo, BlackRock, and KKR.

Featherless AI Cuts Token Costs From $150K to $10K

Featherless AI CEO Eugene Cheah said a five-developer team reduced its monthly token bill from $150,000 to under $10,000. The team ran 116 billion tokens on a single node using 512 AMD MI325 GPUs. Cheah stressed the importance of splitting API keys by workload to track spending and right-sizing models by task type. He recommended using cheaper models for repetitive tasks and separating interactive traffic from CI/CD pipelines. He also highlighted open models that can be about 10 times cheaper in narrow domains.

A.I. Debt Boom Shows Cracks as CoreWeave Struggles

Trillions of dollars in borrowed money have fueled the A.I. boom, but doubts are growing about how easily companies can borrow more and pay it back. This week, investors lost confidence in CoreWeave, a data center company with $72 billion in liabilities and 51 data centers, causing its bond prices to crash. Investors also sold off bonds for other A.I. giants like Oracle, though the declines were less severe. Meanwhile, prices rose for financial instruments insuring corporate debt, including those tied to SoftBank and its partnership with OpenAI. The moves could lead to losses for bondholders and force companies to pay higher interest rates, raising the cost of their expensive A.I. projects.

Sources

NOTE:

This news brief was generated using AI technology (including, but not limited to, Google Gemini API, Llama, Grok, and Mistral) from aggregated news articles, with minimal to no human editing/review. It is provided for informational purposes only and may contain inaccuracies or biases. This is not financial, investment, or professional advice. If you have any questions or concerns, please verify all information with the linked original articles in the Sources section below.

AI Risks AI Regulation AI Ethics AI Hiring Tools AI Cost Optimization AI Data Centers AI Debt Boom AI Concentrated Risk AI Recession AI Scraping AI Model Development AI Open Source AI Large Language Models AI Bias AI Tools AI Companies AI Investments AI Infrastructure AI Cloud Providers AI Chip Companies

Comments

Loading...