DeepSeek's DSec Trains 3M Sandboxes Daily; Anthropic's Claude Used in Weapons

Omaha residents are speaking out as data centers expand rapidly across their community. Some welcome the new jobs and revenue, while others worry about traffic, noise, and the environmental impact. A CNN poll found that nearly 9 in 10 Americans fear AI could surpass human intelligence, adding urgency to the debate over how to regulate the industry.

Meanwhile, businesses are discovering that simply bolting AI onto existing processes delivers only modest gains. Costi Perricos, Deloitte's global generative AI leader, calls this "cappuccino ROI" — workers save small amounts of time here and there, but the savings never reach the bottom line. He advises companies to redesign processes with AI at the center, focusing on growth and new business models rather than just cutting jobs.

In the world of AI development, DeepSeek detailed a new training method called DeepSeek Elastic Compute (DSec) that can run about 3 million sandboxes per day by reallocating computing power only when agents are active. The company built its reputation on delivering advanced AI at lower cost than rivals, though it cautions that AI agents can be untrustworthy and may damage systems without careful monitoring.

Anthropic's Claude AI has made headlines in two separate developments. In one report released on September 10, 2026, Anthropic revealed that a Houthi group used Claude in a weapons-development program — they could not build a working weapon but tested a rocket and created an offline toolkit. Anthropic called for stronger AI regulation and industry-wide security guardrails. Separately, Claude discovered a CRISPR-like enzyme system by analyzing billions of genetic sequences with little human direction, marking the first result from Anthropic's effort to automate scientific discovery.

On the policy front, Senator Ben Ray Luján of New Mexico asked Senate leadership to create a special select committee on artificial intelligence, sending his letter on September 25, 2026, to Senate Majority Leader John Thune and Minority Leader Chuck Schumer. Pope Leo XIV also weighed in during his visit to Paris, warning that technology could erode human dignity and urging ethical education about AI. Rob McDole of Cedarville University argued that Christians should not fear technology but use it wisely, and that the university aims to become AI-resilient rather than AI-resistant.

Key Takeaways

  • A CNN poll shows nearly 9 in 10 Americans worry about AI surpassing human intelligence, fueling local debates over data center regulation in cities like Omaha.
  • Deloitte's Costi Perricos warns that adding AI to existing processes yields only "cappuccino ROI" — small time savings that never reach the bottom line.
  • DeepSeek introduced DeepSeek Elastic Compute (DSec), a training method running about 3 million sandboxes per day by reallocating compute power only when agents are active.
  • Anthropic's report details a Houthi group's use of Claude in a weapons-development program, calling for stronger AI regulation and global threat-intelligence sharing.
  • Claude AI independently discovered a CRISPR-like enzyme system by analyzing billions of genetic sequences, marking Anthropic's first automated scientific discovery.
  • Senator Ben Ray Luján urged Senate leadership to create a special select committee on AI, sending his request on September 25, 2026.
  • Pope Leo XIV warned during a Paris visit that AI could erode human dignity and create a "paradise of machines" without ethical education.
  • Rob McDole of Cedarville University advocates becoming AI-resilient rather than AI-resistant, urging wise and responsible use of the technology.
  • Omaha is proposing new rules to keep data centers in industrial areas and separate from residential neighborhoods.
  • Anthropic stresses that safe AI use requires global agreement on sharing threat intelligence and stronger industry-wide security guardrails.

Omaha residents debate data center rules as AI grows

Omaha residents are sharing their opinions on the rapid growth of data centers in their city. A CNN poll found that nearly 9 in 10 Americans worry about AI becoming smarter than humans. The debate over controlling the data center industry is becoming personal in Nebraska. Some residents welcome new jobs and revenue while others worry about traffic and noise.

Omaha faces data center growth concerns

A First Alert 6 investigation led to a proposed policy change about hiring homeschoolers in Fremont. Omaha residents are speaking out about the fast expansion of data centers in their community. Some people say the growth brings valuable jobs and money. Others say it also brings more traffic and noise pollution.

Omaha data center debate grows amid AI expansion

A First Alert 6 investigation sparked nationwide comments and led to a hiring policy change for homeschoolers in Fremont. Other local updates include a fall movie night at Tepe Park and a milestone for the I-69 Ohio River Crossing. A Cedar Rapids man pleaded guilty to starting a brush fire that damaged two properties. Omaha residents continue to discuss the impact of data center growth.

Omaha considers new data center regulations

Omaha is thinking about new rules for data centers that store computer servers and equipment. The proposed rules would keep data centers in industrial areas away from homes. A CNN poll shows nearly 9 in 10 Americans fear AI surpassing human intelligence. The city wants to balance industry growth with resident concerns about the environment and safety. This debate is happening across the country as the industry grows fast.

Cappuccino ROI keeps AI savings from profits

Businesses that add AI to every step of an existing process see small time savings that never reach the bottom line. Costi Perricos from Deloitte calls this cappuccino ROI because workers just save a little time here and there. Real gains come from redesigning processes with AI at the center instead of just adding it on top.

Deloitte expert says AI gains often miss the bottom line

Costi Perricos, Deloitte's global generative AI leader, says companies that just add AI to existing steps get small time savings he calls cappuccino ROI. Real gains come from redesigning processes with AI at the center. He advises focusing on growth and new business models instead of cutting office jobs. He also urges leaders to build AI fluency and support a learning culture.

DeepSeek reveals new method for training AI agents

DeepSeek detailed a new training method called DeepSeek Elastic Compute, or DSec, that can run about 3 million sandboxes per day. The method reallocates computing power only when agents are active, saving resources. DeepSeek notes that agents can be untrustworthy and may damage systems, so careful monitoring is needed. The company built its reputation on delivering advanced AI at lower cost than rivals.

Westmont professors host interactive AI and robotics talk

Engineering professors Dan Jensen and Kristin L. Wood will lead a free public lecture on AI and robotics on October 15 at Westmont Downtown. Attendees should bring a smartphone to participate. The talk will explain how AI works, explore creative tools, and share ten practical ways to engage with the technology. The event is sponsored by the Westmont Foundation.

Undark editor explores the spread of artificial intelligence

An Undark editorial describes AI as spreading like kudzu through culture, from classrooms to legislatures. The piece introduces a series called Brain Trust covering lab-grown brain organoids, engineered cognition, and other forms of AI beyond traditional software. Undark is also testing a limited print edition as a fundraising effort. The magazine is based in Cambridge, Massachusetts.

Anthropic report examines Houthi use of Claude AI

Anthropic released a report on September 10, 2026, about a Houthi group that used Claude in a weapons-development program. The group could not build a working weapon but tested a rocket and made an offline toolkit. The report calls for stronger AI regulation and industry-wide security guardrails. It warns that safe AI use needs global agreement on sharing threat intelligence.

Pope Leo warns France about AI creating a paradise of machines

Pope Leo XIV opened his four-day visit to Paris by calling for ethical education about AI. He warned that technology could erode human dignity if people do not learn moral discernment. He described a risk of a paradise of machines invading daily life. He urged people to acknowledge the moral good.

Anthropic AI discovers CRISPR-like enzyme system it does not fully understand

Anthropic's Claude AI found a new enzyme system that looks similar to CRISPR gene-editing technology. The AI analyzed billions of genetic sequences with little human help. The researchers did not tell the AI to look for specific patterns. It is the first result from Anthropic's effort to automate scientific discovery.

Rob McDole discusses faith and the future of AI at Cedarville

Rob McDole shared his views on AI and faith in a Cedarville Stories interview. He said Christians should not fear technology but use it wisely to serve others. He cautioned against treating AI systems as people. He also explained why Cedarville University aims to become AI-resilient rather than AI-resistant.

Senator Luján calls for Senate AI select committee

Senator Ben Ray Luján asked Senate leadership to create a special select committee on artificial intelligence. He sent a letter to Senate Majority Leader John Thune and Minority Leader Chuck Schumer. The call came on September 25, 2026. The senator represents New Mexico.

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 Expansion AI Regulation AI Impact AI Development AI Ethics AI Threat AI Automation AI Education AI Resilience AI Policy AI Industry AI Weapons AI Safety AI Economy AI Environment AI Jobs AI Regulation Debate AI Industry Growth AI Scientific Discovery AI Global Threat

Comments

Loading...