<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Fathom News</title>
    <description>The latest articles on DEV Community by Fathom News (@fathomnews).</description>
    <link>https://dev.to/fathomnews</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4211764%2Fa4321cc1-ddf3-48df-baad-4c9b27e96819.png</url>
      <title>DEV Community: Fathom News</title>
      <link>https://dev.to/fathomnews</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/fathomnews"/>
    <language>en</language>
    <item>
      <title>Nvidia Puts National Security In Board's Hands Days After Huang Quits</title>
      <dc:creator>Fathom News</dc:creator>
      <pubDate>Tue, 17 Nov 2026 11:00:35 +0000</pubDate>
      <link>https://dev.to/fathomnews/nvidia-puts-national-security-in-boards-hands-days-after-huang-quits-4pad</link>
      <guid>https://dev.to/fathomnews/nvidia-puts-national-security-in-boards-hands-days-after-huang-quits-4pad</guid>
      <description>&lt;p&gt;&lt;strong&gt;Nvidia&lt;/strong&gt; has set up a new Nvidia national security committee on its board, putting export controls and China-related trade risk under direct board watch just days after &lt;strong&gt;Jensen Huang&lt;/strong&gt; quit as CEO.&lt;/p&gt;

&lt;h2&gt;
  
  
  Nvidia National Security Committee: Key Takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Nvidia's board created the National Security and Export Control Oversight Committee, a new standing committee&lt;/li&gt;
&lt;li&gt;It will oversee national security policies, export control compliance and global trade risk management&lt;/li&gt;
&lt;li&gt;Nvidia's National Security Team and Global Trade Compliance Team will report to the committee and to Interim CEO Colette Kress&lt;/li&gt;
&lt;li&gt;Director Suzanne Nora Johnson chairs it, with Stephen C. Neal and John O. Dabiri as members&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Nvidia &lt;a href="https://nvidianews.nvidia.com/news/nvidia-board-of-directors-establishes-national-security-and-export-control-oversight-committee" rel="noopener noreferrer"&gt;announced&lt;/a&gt; the move on Tuesday. The committee will meet at least quarterly, including with management and in executive session, and will hire independent outside experts in law, national security and trade compliance.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why Nvidia Is Doing This Now
&lt;/h2&gt;

&lt;p&gt;The timing says a lot. Nvidia announced its CEO transition on Nov. 12, when Huang stepped down and finance chief &lt;strong&gt;Colette Kress&lt;/strong&gt; took over as interim CEO. Two days before that, the independent members of Nvidia's board had issued their own statement.&lt;/p&gt;

&lt;p&gt;Now the board is pulling the company's most sensitive issue, who gets to buy and use its AI chips, up to its own level. "NVIDIA recognizes that advanced silicon technologies and the emergence of artificial and superintelligence present unprecedented challenges for national security," said &lt;strong&gt;Harvey Jones&lt;/strong&gt;, chairman of Nvidia's board.&lt;/p&gt;

&lt;p&gt;Jones added that Nvidia is "committed to working constructively with government and industry partners" on a framework that safeguards national security while supporting innovation. The board said the committee shows its commitment to engage more with government stakeholders.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Also Read:&lt;/strong&gt; &lt;a href="https://www.fathom.news/nvidia-stock-huang-resigns/" rel="noopener noreferrer"&gt;Nvidia Crashes 12% As Jensen Huang Quits, And Wall Street Slashes Targets To $200&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Who Sits On The Committee
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Role&lt;/th&gt;
&lt;th&gt;Who&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Chair&lt;/td&gt;
&lt;td&gt;Suzanne Nora Johnson&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Member&lt;/td&gt;
&lt;td&gt;Stephen C. Neal&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Member&lt;/td&gt;
&lt;td&gt;John O. Dabiri&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Reports to it&lt;/td&gt;
&lt;td&gt;National Security Team and Global Trade Compliance Team&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Meets&lt;/td&gt;
&lt;td&gt;At least quarterly&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;The committee will also oversee Nvidia's work to strengthen its internal security and trade compliance teams, which handle day-to-day checks on where the company's chips end up.&lt;/p&gt;

&lt;h2&gt;
  
  
  What To Watch Next
&lt;/h2&gt;

&lt;p&gt;Pressure on chip exports is rising fast. On Monday the U.S. Commerce Department issued a rule requiring licenses for data centers anywhere that rent large amounts of AI computing power to Chinese companies, and for advanced chips going to Malaysia, Singapore, Thailand, Indonesia and Brazil.&lt;/p&gt;

&lt;p&gt;The new committee will be the place where Nvidia's board answers for how it handles rules like that. The test is whether it changes who Nvidia sells to, and whether the board names a permanent CEO who can rebuild trust with Washington.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Read Next:&lt;/strong&gt; &lt;a href="https://www.fathom.news/ai-chip-remote-access-rule/" rel="noopener noreferrer"&gt;U.S. Shuts China’s Cloud Back Door To American AI Chips&lt;/a&gt;&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://www.fathom.news/nvidia-national-security-committee/" rel="noopener noreferrer"&gt;Fathom News&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>news</category>
      <category>technology</category>
      <category>machinelearning</category>
    </item>
    <item>
      <title>Ex-Palantir Team's Thread AI Gets A Fast Lane Into Microsoft's Customers</title>
      <dc:creator>Fathom News</dc:creator>
      <pubDate>Tue, 17 Nov 2026 08:31:51 +0000</pubDate>
      <link>https://dev.to/fathomnews/ex-palantir-teams-thread-ai-gets-a-fast-lane-into-microsofts-customers-3jmb</link>
      <guid>https://dev.to/fathomnews/ex-palantir-teams-thread-ai-gets-a-fast-lane-into-microsofts-customers-3jmb</guid>
      <description>&lt;p&gt;&lt;strong&gt;Thread AI&lt;/strong&gt; has joined Microsoft for Startups Pegasus and listed Lemma on Microsoft Marketplace, a move that could widen access to controlled AI workflows for enterprise teams in regulated industries.&lt;/p&gt;

&lt;h2&gt;
  
  
  Key Takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Thread AI was accepted into the invite-only Microsoft for Startups Pegasus Program&lt;/li&gt;
&lt;li&gt;Its AI orchestration platform Lemma is now available on Microsoft Marketplace&lt;/li&gt;
&lt;li&gt;Azure customers can purchase Lemma directly through Microsoft Marketplace&lt;/li&gt;
&lt;li&gt;Thread AI says Lemma is built to help enterprises build, govern, and run controlled AI workflows and agents&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Thread AI &lt;a href="https://www.prnewswire.com/news-releases/thread-ai-joins-the-microsoft-for-startups-pegasus-program-to-deliver-its-ai-workflow-orchestration-platform-on-microsoft-marketplace-302940646.html" rel="noopener noreferrer"&gt;announced&lt;/a&gt; Nov. 16 that it had been accepted into Microsoft for Startups Pegasus, an invite-only program that connects select companies to Microsoft's global network of enterprise customers.&lt;/p&gt;

&lt;p&gt;As part of that move, Lemma is now sold through Microsoft Marketplace. Thread AI said the platform lets organizations build, govern, and run controlled AI workflows and agents across critical operations in regulated industries.&lt;/p&gt;

&lt;h2&gt;
  
  
  What Lemma Is Meant To Do
&lt;/h2&gt;

&lt;p&gt;Thread AI describes itself as an AI infrastructure company founded by Palantir's former heads of AI product and engineering.&lt;/p&gt;

&lt;p&gt;The company said Lemma is a composable infrastructure and workflow orchestration platform that lets enterprises deploy AI into core operations. It said the platform provides the foundational layer for agentic processes to run at scale with control, governance, and reliability assurances.&lt;/p&gt;

&lt;p&gt;For buyers on Azure, the company said Lemma offers a single platform with built-in guardrails for building and deploying AI-powered automated workflows and agents across AI models, databases, systems of record, and applications.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Also Read:&lt;/strong&gt; &lt;a href="https://www.fathom.news/microsoft-and-amazon-modelop-expands-ai/" rel="noopener noreferrer"&gt;Software Firm Gives Companies 1 Rulebook For AI Across Microsoft And Amazon&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  What Microsoft Pegasus Adds
&lt;/h2&gt;

&lt;p&gt;Microsoft's Tom Davis, general manager for Microsoft for Startups, said Pegasus gives select startups access to Azure AI models, enterprise-grade security and privacy, and go-to-market support.&lt;/p&gt;

&lt;p&gt;Thread AI CEO Angela McNeal said many enterprises now have access to models but lack the infrastructure needed to implement automated workflows at scale or trust AI systems to act reliably enough to ship a product. She said the Microsoft collaboration expands discoverability and access to Lemma.&lt;/p&gt;

&lt;p&gt;Azure customers can now purchase Lemma directly through Microsoft Marketplace, giving Thread AI a route into Microsoft's enterprise ecosystem.&lt;/p&gt;

&lt;h2&gt;
  
  
  What Comes Next
&lt;/h2&gt;

&lt;p&gt;Thread AI said its participation in Pegasus will pair marketplace access with Microsoft's technical and go-to-market collaboration.&lt;/p&gt;

&lt;p&gt;The company did not give a rollout schedule beyond saying Lemma is now available on Microsoft Marketplace.&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;The Deal&lt;/th&gt;
&lt;th&gt;Details&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Startup&lt;/td&gt;
&lt;td&gt;Thread AI, founded by Palantir's former heads of AI product and engineering&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Product&lt;/td&gt;
&lt;td&gt;Lemma, a platform to build, govern and run AI workflows and agents&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Program&lt;/td&gt;
&lt;td&gt;Microsoft for Startups Pegasus, invite-only&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Where to buy&lt;/td&gt;
&lt;td&gt;Microsoft Marketplace, for Azure customers&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Target buyers&lt;/td&gt;
&lt;td&gt;Enterprises in regulated industries&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;&lt;strong&gt;Read Next:&lt;/strong&gt; &lt;a href="https://www.fathom.news/lumen-and-microsoft-bank-data-security/" rel="noopener noreferrer"&gt;Microsoft Teams With Lumen To Keep Bank Data Off Web&lt;/a&gt;&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://www.fathom.news/thread-ai-microsoft-pegasus/" rel="noopener noreferrer"&gt;Fathom News&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>news</category>
      <category>technology</category>
      <category>machinelearning</category>
    </item>
    <item>
      <title>U.S. Shuts China's Cloud Back Door To American AI Chips</title>
      <dc:creator>Fathom News</dc:creator>
      <pubDate>Tue, 17 Nov 2026 08:30:32 +0000</pubDate>
      <link>https://dev.to/fathomnews/us-shuts-chinas-cloud-back-door-to-american-ai-chips-4o0i</link>
      <guid>https://dev.to/fathomnews/us-shuts-chinas-cloud-back-door-to-american-ai-chips-4o0i</guid>
      <description>&lt;p&gt;The U.S. has moved to shut a cloud back door that could let Chinese companies tap American AI chips, with a new Commerce Department rule that makes overseas data centers get a license before renting large amounts of computing power to them.&lt;/p&gt;

&lt;h2&gt;
  
  
  Key Takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Overseas data centers that supply big volumes of AI computing power to companies based in China, Macau or other arms-embargoed countries now need a U.S. license&lt;/li&gt;
&lt;li&gt;Advanced AI chips shipped to Indonesia, Malaysia, Singapore, Thailand and Brazil will also need a license, unless they go to U.S.-headquartered companies&lt;/li&gt;
&lt;li&gt;Data centers already serving those customers have 90 days to get a license or face future export restrictions&lt;/li&gt;
&lt;li&gt;The rule formally scraps the Biden-era AI Diffusion Rule, which was never enforced&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The Commerce Department's Bureau of Industry and Security &lt;a href="https://www.bis.gov/press-release/bis-issues-rule-addressing-remote-access-american-super-intelligence-technology" rel="noopener noreferrer"&gt;released&lt;/a&gt; the interim final rule on Monday. It said the rule was written by and at the direction of the Treasury Department and the National Security Council, and is meant to stop existing chip controls being dodged "through remote computing arrangements involving data centers in third countries."&lt;/p&gt;

&lt;h2&gt;
  
  
  How The Back Door Worked
&lt;/h2&gt;

&lt;p&gt;The U.S. already bans shipping its most advanced AI chips to companies headquartered in China and other countries of concern. But a company does not need to own the chips to use them. It can rent the computing power from a data center somewhere else and train its AI models over the internet.&lt;/p&gt;

&lt;p&gt;The new rule targets exactly that. A license is now needed to send advanced AI chips to any data center, anywhere in the world, that will give companies headquartered in Macau or Country Group D:5, which includes China, access to more than 12,000,000 "aggregate TPP" of computing power. That counts data centers linked together by networks, not just those in one building. Data centers, chip exporters and foreign chip factories must check who really owns and controls their customers.&lt;/p&gt;

&lt;p&gt;There is one carve-out. Computing used for "commercial recommender systems," the kind online stores and social media apps use to suggest products and posts, does not need a license.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Also Read:&lt;/strong&gt; &lt;a href="https://www.fathom.news/ai-compute-grabs-72-chips-servers/" rel="noopener noreferrer"&gt;U.S. Grabs 72% Of AI Chips And Servers As China Gets 10%&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Five Countries Get New Checks
&lt;/h2&gt;

&lt;p&gt;The rule also adds a license requirement for advanced AI chips going to Indonesia, Malaysia, Singapore, Thailand and Brazil, which the department said carry a "higher-than-normal risk of diversion." Chips bound for companies headquartered in the U.S. are exempt if they meet reporting rules.&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;What Changes&lt;/th&gt;
&lt;th&gt;Who It Hits&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;License for renting big AI computing power&lt;/td&gt;
&lt;td&gt;Data centers anywhere serving firms based in China, Macau and other D:5 countries&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;License for advanced AI chips&lt;/td&gt;
&lt;td&gt;Shipments to Indonesia, Malaysia, Singapore, Thailand and Brazil&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;90-day deadline&lt;/td&gt;
&lt;td&gt;Data centers already serving those customers&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Tighter checks&lt;/td&gt;
&lt;td&gt;Chip factories and packagers making advanced chips&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Scrapped&lt;/td&gt;
&lt;td&gt;The Biden-era AI Diffusion Rule&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;The department also warned that countries which tolerate evasion could face import restrictions, new chip export limits, a downgrade in their export control status, and Entity List designations for the people involved.&lt;/p&gt;

&lt;h2&gt;
  
  
  What To Watch Next
&lt;/h2&gt;

&lt;p&gt;The rule takes effect when it is published in the Federal Register, and the public then has 60 days to comment. The 90-day clock for data centers already serving Chinese customers starts on that same date.&lt;/p&gt;

&lt;p&gt;The biggest test is Southeast Asia. Malaysia, Singapore and Thailand have been racing to build AI data centers, and every advanced chip headed there now needs Washington's sign-off unless the buyer is American. How fast licenses are granted will decide whether that building boom keeps its pace or slows down.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Read Next:&lt;/strong&gt; &lt;a href="https://www.fathom.news/why-nvidia-and-amd-are-fighting/" rel="noopener noreferrer"&gt;Why Nvidia And AMD Fight Over Inference As Data Center Hits $89B&lt;/a&gt;&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://www.fathom.news/ai-chip-remote-access-rule/" rel="noopener noreferrer"&gt;Fathom News&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>news</category>
      <category>policy</category>
      <category>technology</category>
    </item>
    <item>
      <title>AMD's Lux Supercomputer Opens At Oak Ridge As America's First AI Factory</title>
      <dc:creator>Fathom News</dc:creator>
      <pubDate>Tue, 17 Nov 2026 07:00:35 +0000</pubDate>
      <link>https://dev.to/fathomnews/amds-lux-supercomputer-opens-at-oak-ridge-as-americas-first-ai-factory-5aoc</link>
      <guid>https://dev.to/fathomnews/amds-lux-supercomputer-opens-at-oak-ridge-as-americas-first-ai-factory-5aoc</guid>
      <description>&lt;p&gt;&lt;strong&gt;AMD&lt;/strong&gt;'s Lux supercomputer is now open to scientists at Oak Ridge National Laboratory, becoming the first dedicated U.S. AI factory for the government's Genesis Mission and landing at No. 14 on the new list of the world's fastest machines.&lt;/p&gt;

&lt;h2&gt;
  
  
  Key Takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Lux is now open to scientific users at Oak Ridge National Laboratory in Tennessee&lt;/li&gt;
&lt;li&gt;It ranks 14th on the latest TOP500 list of the world's most powerful supercomputers&lt;/li&gt;
&lt;li&gt;The system runs on AMD EPYC processors and AMD Instinct MI355X GPUs and was built by HPE&lt;/li&gt;
&lt;li&gt;First projects include fusion power plant design, new materials and water forecasts for the Tennessee Valley&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;AMD &lt;a href="https://ir.amd.com/news-events/press-releases/detail/1307/amd-powered-lux-supercomputer-opens-to-scientific-users-at-oak-ridge-becoming-first-ai-factory-for-does-genesis-mission" rel="noopener noreferrer"&gt;announced&lt;/a&gt; the opening on Monday. Lux was built through a public-private partnership between AMD and the Department of Energy, with &lt;strong&gt;HPE&lt;/strong&gt; putting the system together.&lt;/p&gt;

&lt;h2&gt;
  
  
  What Lux Will Work On
&lt;/h2&gt;

&lt;p&gt;Lux mixes AI with the classic number-crunching simulations supercomputers are known for. Its first Genesis Mission projects include AI-assisted design of fusion power plants, an AI "foundation model factory" for new materials and advanced manufacturing, and forecasts of water supply, river flow and river temperature in the Tennessee Valley Authority region.&lt;/p&gt;

&lt;p&gt;"Lux represents the first major addition to DOE's computing infrastructure in support of the Genesis Mission," said &lt;strong&gt;Georgia Tourassi&lt;/strong&gt;, who leads computing at Oak Ridge.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Also Read:&lt;/strong&gt; &lt;a href="https://www.fathom.news/openai-tensorwave-deal/" rel="noopener noreferrer"&gt;OpenAI Hands TensorWave A 457MW Deal To Run On AMD Chips&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Why It Matters For AMD
&lt;/h2&gt;

&lt;p&gt;Oak Ridge is already home to Frontier, the AMD-powered machine that was the first supercomputer to break the exascale barrier. Lux adds capacity built just for AI-driven science, and AMD supplies nearly every part of it, from processors and graphics chips to networking and its ROCm software.&lt;/p&gt;

&lt;p&gt;That gives AMD a showcase in the race to power government AI, a market where Nvidia has long led. "Lux shows what is possible when national laboratories, government and industry work together," said &lt;strong&gt;Thomas Zacharia&lt;/strong&gt;, AMD's senior vice president for global public sector and strategic partnerships.&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Lux At A Glance&lt;/th&gt;
&lt;th&gt;Details&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Where&lt;/td&gt;
&lt;td&gt;Oak Ridge National Laboratory&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;TOP500 rank&lt;/td&gt;
&lt;td&gt;No. 14&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Chips&lt;/td&gt;
&lt;td&gt;AMD EPYC CPUs, AMD Instinct MI355X GPUs&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Built by&lt;/td&gt;
&lt;td&gt;HPE&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Mission&lt;/td&gt;
&lt;td&gt;First dedicated U.S. AI factory for Genesis&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Next up&lt;/td&gt;
&lt;td&gt;Discovery supercomputer&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  What To Watch Next
&lt;/h2&gt;

&lt;p&gt;AMD said the upcoming Discovery supercomputer at Oak Ridge will tie AI and supercomputing even closer together, though it gave no date. It is also working with the lab on the American Science Cloud, a planned national platform to share AI and supercomputing power across government labs and universities.&lt;/p&gt;

&lt;p&gt;Researchers who want time on Lux can apply through the Oak Ridge Leadership Computing Facility. How quickly that queue fills will show how hungry U.S. science is for AI computing power.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Read Next:&lt;/strong&gt; &lt;a href="https://www.fathom.news/why-nvidia-and-amd-are-fighting/" rel="noopener noreferrer"&gt;Why Nvidia And AMD Fight Over Inference As Data Center Hits $89B&lt;/a&gt;&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://www.fathom.news/amd-lux-supercomputer/" rel="noopener noreferrer"&gt;Fathom News&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>news</category>
      <category>technology</category>
      <category>machinelearning</category>
    </item>
    <item>
      <title>Google's Gemini Notebook Can Now Turn Your Documents Into A Website</title>
      <dc:creator>Fathom News</dc:creator>
      <pubDate>Tue, 17 Nov 2026 02:00:33 +0000</pubDate>
      <link>https://dev.to/fathomnews/googles-gemini-notebook-can-now-turn-your-documents-into-a-website-4nga</link>
      <guid>https://dev.to/fathomnews/googles-gemini-notebook-can-now-turn-your-documents-into-a-website-4nga</guid>
      <description>&lt;p&gt;&lt;strong&gt;Google&lt;/strong&gt;'s Gemini Notebook can now turn a pile of documents into a working website, such as a dashboard or a visual explainer, from one short request typed into chat.&lt;/p&gt;

&lt;h2&gt;
  
  
  Key Takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Gemini Notebook can now turn your sources into interactive sites like dashboards and visual explainers&lt;/li&gt;
&lt;li&gt;Users type "make me a site" in the Notebook chat and describe what they want&lt;/li&gt;
&lt;li&gt;The finished site opens in a new tab, where it can be viewed or shared&lt;/li&gt;
&lt;li&gt;Google has not said who can use it yet or what limits apply&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Google &lt;a href="https://x.com/Google/status/2122449668613579018" rel="noopener noreferrer"&gt;announced&lt;/a&gt; the feature on X on Monday, saying Gemini Notebook "can now turn your sources into interactive sites like dashboards and visual explainers."&lt;/p&gt;

&lt;h2&gt;
  
  
  How It Works
&lt;/h2&gt;

&lt;p&gt;Gemini Notebook is where people drop in their own material, such as documents, notes and links, and ask Gemini questions about it.&lt;/p&gt;

&lt;p&gt;The new option goes a step further. A user types "make me a site" in the Notebook chat, adds a short description of what they want, and Gemini builds an interactive page from those sources. Selecting "Open in a new tab" shows the finished site, ready to share with others.&lt;/p&gt;

&lt;p&gt;That could save students, teams and researchers hours of turning notes into slides or charts by hand. A class reading list could become a visual explainer, and a stack of reports could become a dashboard.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Also Read:&lt;/strong&gt; &lt;a href="https://www.fathom.news/google-playground-text-to-game-launch/" rel="noopener noreferrer"&gt;Google Launches Playground, Turns Text Prompts Into Playable Games: No Code Required&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Part Of A Bigger Gemini Push
&lt;/h2&gt;

&lt;p&gt;The update is one of a string of Gemini launches from Google in recent days. The company has said Gemini 4 Hydrogen is "available to everyone" across the Gemini app, Google Search and Google AI Studio, and it added Google Wallet to Gemini's connected apps so it can draw charts for cash flow and budgets.&lt;/p&gt;

&lt;p&gt;The pattern is clear. Google is pushing Gemini beyond answering questions and toward building finished things, from code to charts to whole web pages.&lt;/p&gt;

&lt;h2&gt;
  
  
  What To Watch Next
&lt;/h2&gt;

&lt;p&gt;Google has not said whether the site feature is open to all Notebook users or only paid plans, which file types work best, or how long shared sites stay online. Those answers will decide whether it becomes an everyday tool or a neat demo.&lt;/p&gt;

&lt;p&gt;The bigger test is trust. A site built by AI from your sources is only as good as Gemini's reading of them, so users will want an easy way to check that the charts match the documents.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Read Next:&lt;/strong&gt; &lt;a href="https://www.fathom.news/google-gemini-agent-enterprise-launch/" rel="noopener noreferrer"&gt;Google Debuts Universal Gemini Agent: ‘A Single Agent For Work’&lt;/a&gt;&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://www.fathom.news/gemini-notebook-sites/" rel="noopener noreferrer"&gt;Fathom News&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>news</category>
      <category>technology</category>
      <category>machinelearning</category>
    </item>
    <item>
      <title>OpenAI Hands TensorWave A 457MW Deal To Run On AMD Chips</title>
      <dc:creator>Fathom News</dc:creator>
      <pubDate>Tue, 17 Nov 2026 00:02:23 +0000</pubDate>
      <link>https://dev.to/fathomnews/openai-hands-tensorwave-a-457mw-deal-to-run-on-amd-chips-m11</link>
      <guid>https://dev.to/fathomnews/openai-hands-tensorwave-a-457mw-deal-to-run-on-amd-chips-m11</guid>
      <description>&lt;p&gt;Little-known cloud firm &lt;strong&gt;TensorWave&lt;/strong&gt; has landed a six-year deal to supply &lt;strong&gt;OpenAI&lt;/strong&gt; with 457 megawatts of AI computing power built on &lt;strong&gt;AMD&lt;/strong&gt; chips, the first named partner in OpenAI's giant bet on AMD.&lt;/p&gt;

&lt;h2&gt;
  
  
  Key Takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;TensorWave will deliver 457 megawatts of dedicated AI infrastructure for OpenAI across U.S. data centers&lt;/li&gt;
&lt;li&gt;The deal runs for six years and uses AMD Instinct MI455X GPUs in AMD Helios rack-scale systems&lt;/li&gt;
&lt;li&gt;TensorWave says it is the first publicly announced neocloud supporting OpenAI and AMD's strategic partnership&lt;/li&gt;
&lt;li&gt;The company has not shared the price, the data center sites or when the capacity comes online&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;TensorWave &lt;a href="https://x.com/tensorwave/status/2122332671955132454" rel="noopener noreferrer"&gt;announced&lt;/a&gt; the agreement on Monday, calling it "a landmark six-year agreement" and saying 457 megawatts "is just the beginning" for the company.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why 457 Megawatts Matters
&lt;/h2&gt;

&lt;p&gt;Megawatts are how the AI industry now measures computing deals, because the chips are limited by how much power a data center can feed them. A 457-megawatt commitment means hundreds of thousands of homes' worth of electricity going into one customer's AI work.&lt;/p&gt;

&lt;p&gt;All of it will run on AMD's newest Instinct MI455X graphics chips, packed into Helios racks that AMD sells as complete systems. That makes the deal a big test of whether AMD hardware can carry OpenAI's work at a scale long dominated by &lt;strong&gt;Nvidia&lt;/strong&gt;.&lt;/p&gt;

&lt;p&gt;"TensorWave was built to prove that AMD infrastructure could operate at the scale and reliability that the world's most demanding AI companies require," said &lt;strong&gt;Piotr Tomasik&lt;/strong&gt;, TensorWave's president and co-founder.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Also Read:&lt;/strong&gt; &lt;a href="https://www.fathom.news/why-nvidia-and-amd-are-fighting/" rel="noopener noreferrer"&gt;Why Nvidia And AMD Fight Over Inference As Data Center Hits $89B&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  The Bigger AMD Bet
&lt;/h2&gt;

&lt;p&gt;OpenAI and AMD struck a strategic partnership in October 2025 for up to 6 gigawatts of AMD chips. TensorWave is the first neocloud, a cloud company built just for AI, to say publicly that it is helping deliver that capacity.&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;The Deal&lt;/th&gt;
&lt;th&gt;Details&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Power&lt;/td&gt;
&lt;td&gt;457 megawatts&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Length&lt;/td&gt;
&lt;td&gt;Six years&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Chips&lt;/td&gt;
&lt;td&gt;AMD Instinct MI455X&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Racks&lt;/td&gt;
&lt;td&gt;AMD Helios&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Where&lt;/td&gt;
&lt;td&gt;U.S. data centers&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Price&lt;/td&gt;
&lt;td&gt;Not disclosed&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;"AI infrastructure is entering a new phase, where the ability to deliver hundreds of megawatts of advanced compute reliably and at speed will separate the companies that can support frontier AI from those that cannot," Tomasik said.&lt;/p&gt;

&lt;h2&gt;
  
  
  What To Watch Next
&lt;/h2&gt;

&lt;p&gt;TensorWave has not said which data centers will host the chips, when the first racks switch on or what OpenAI is paying. Those details will show how fast AMD's newest chips can actually reach OpenAI.&lt;/p&gt;

&lt;p&gt;The bigger question is who comes next. If more neoclouds sign on to OpenAI's AMD plan, it would show that the AI boom can run on more than one chipmaker, a shift that would matter for every company paying for AI computing.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Read Next:&lt;/strong&gt; &lt;a href="https://www.fathom.news/memory-chip-squeeze-ai-openai/" rel="noopener noreferrer"&gt;Forget GPUs, A Memory Chip Squeeze Could Choke AI Even After OpenAI’s $250B Splurge&lt;/a&gt;&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://www.fathom.news/openai-tensorwave-deal/" rel="noopener noreferrer"&gt;Fathom News&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>news</category>
      <category>technology</category>
      <category>machinelearning</category>
    </item>
    <item>
      <title>Robot Boss Adcock's Hark Card Pays 5% Cash Back With A Catch</title>
      <dc:creator>Fathom News</dc:creator>
      <pubDate>Tue, 17 Nov 2026 00:01:44 +0000</pubDate>
      <link>https://dev.to/fathomnews/robot-boss-adcocks-hark-card-pays-5-cash-back-with-a-catch-42p3</link>
      <guid>https://dev.to/fathomnews/robot-boss-adcocks-hark-card-pays-5-cash-back-with-a-catch-42p3</guid>
      <description>&lt;p&gt;&lt;strong&gt;Brett Adcock&lt;/strong&gt;, the founder behind humanoid robot maker &lt;strong&gt;Figure&lt;/strong&gt;, is turning his AI assistant &lt;strong&gt;Hark&lt;/strong&gt; into a money business with a new card that pays 5% cash back, but only if you pay Hark a monthly fee first.&lt;/p&gt;

&lt;h2&gt;
  
  
  Key Takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Hark Card is a digital card with 5% cash back on nearly every purchase made through Hark&lt;/li&gt;
&lt;li&gt;It is limited to 10,000 U.S. members, who can ask Hark for the card or join a waitlist&lt;/li&gt;
&lt;li&gt;The 5% only applies on Hark Pro² or higher, which costs $20 a month and caps rewards at $1,000 of spending&lt;/li&gt;
&lt;li&gt;Hark's AI will also hunt for deals, handle returns, catch accidental charges and cancel unwanted subscriptions&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Adcock &lt;a href="https://x.com/adcock_brett/status/2122406282045092235" rel="noopener noreferrer"&gt;unveiled&lt;/a&gt; the card on X on Monday, calling it "the first card that saves you money." The post drew more than 54,000 views in its first hour.&lt;/p&gt;

&lt;h2&gt;
  
  
  How The Card Works
&lt;/h2&gt;

&lt;p&gt;Members do not swipe Hark Card at a till. They ask the Hark assistant to buy things online, from groceries and food delivery to clothes and flights, and the card pays. Adcock said Hark will also find deals, search for lower prices, handle returns, catch accidental charges and cancel subscriptions people no longer need.&lt;/p&gt;

&lt;p&gt;The card creates a new card number for every merchant, so one hacked store cannot expose the rest of a member's spending. There is no annual fee and no credit check. Adcock added in a reply that Hark itself never sees or stores the card details, which go straight to the card processor and bank.&lt;/p&gt;

&lt;p&gt;Hark is not a bank. The prepaid Visa card is issued by &lt;strong&gt;Lead Bank&lt;/strong&gt;, &lt;strong&gt;Stripe&lt;/strong&gt; handles the money movement and customer funds are held at &lt;strong&gt;Fifth Third Bank&lt;/strong&gt;.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Also Read:&lt;/strong&gt; &lt;a href="https://www.fathom.news/synchrony-openai-agentic-commerce/" rel="noopener noreferrer"&gt;AI Is Coming For Your Checkout As Store Card Giant Synchrony Signs Up OpenAI&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  The Catch In The 5%
&lt;/h2&gt;

&lt;p&gt;The headline rate sits behind Hark's paid plans. Each tier caps how much spending earns the 5% each month, and rewards reset on the first of the month.&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Plan&lt;/th&gt;
&lt;th&gt;Monthly Fee&lt;/th&gt;
&lt;th&gt;Spending That Earns 5%&lt;/th&gt;
&lt;th&gt;Most Cash Back A Month&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Pro²&lt;/td&gt;
&lt;td&gt;$20&lt;/td&gt;
&lt;td&gt;Up to $1,000&lt;/td&gt;
&lt;td&gt;$50&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Pro³&lt;/td&gt;
&lt;td&gt;$100&lt;/td&gt;
&lt;td&gt;Up to $5,000&lt;/td&gt;
&lt;td&gt;$250&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Pro⁴&lt;/td&gt;
&lt;td&gt;$200&lt;/td&gt;
&lt;td&gt;Up to $10,000&lt;/td&gt;
&lt;td&gt;$500&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;That math means a Pro² member has to spend $400 a month through Hark just to win back the $20 fee. The deal only pays off for people who are ready to let an AI handle a big share of their shopping.&lt;/p&gt;

&lt;h2&gt;
  
  
  What To Watch Next
&lt;/h2&gt;

&lt;p&gt;Hark has moved fast since Adcock launched its Pro plan on Oct. 6, offering a free paid plan to the first 100,000 sign-ups. Since then it has added shopping inside iMessage, location-based help and a $1 million free pizza giveaway ordered through the app.&lt;/p&gt;

&lt;p&gt;The card is the clearest sign yet that AI assistants want to hold the wallet, not just give advice. The test is whether 10,000 members fill up and whether Hark opens the card to everyone. Adcock has not said when, or if, that will happen.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Read Next:&lt;/strong&gt; &lt;a href="https://www.fathom.news/base-card-spend-jumps-fourteenfold-at-checkouts/" rel="noopener noreferrer"&gt;Coinbase Base Card Spending Jumps 14x As Crypto Hits Checkouts&lt;/a&gt;&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://www.fathom.news/hark-card/" rel="noopener noreferrer"&gt;Fathom News&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>news</category>
      <category>technology</category>
      <category>machinelearning</category>
    </item>
    <item>
      <title>OpenAI Says New Free Tool Can Flag AI Made Abuse Images</title>
      <dc:creator>Fathom News</dc:creator>
      <pubDate>Tue, 17 Nov 2026 00:01:00 +0000</pubDate>
      <link>https://dev.to/fathomnews/openai-says-new-free-tool-can-flag-ai-made-abuse-images-2chc</link>
      <guid>https://dev.to/fathomnews/openai-says-new-free-tool-can-flag-ai-made-abuse-images-2chc</guid>
      <description>&lt;p&gt;&lt;strong&gt;OpenAI&lt;/strong&gt; is giving vetted organizations a free tool to detect child sexual abuse material, including AI-generated images, and is adding ChatGPT warnings and developer guidance.&lt;/p&gt;

&lt;h2&gt;
  
  
  Key Takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;The new tool is free for vetted organizations&lt;/li&gt;
&lt;li&gt;It helps detect child sexual abuse material, including AI-generated images&lt;/li&gt;
&lt;li&gt;The classifier was developed with assistance from the National Center for Missing &amp;amp; Exploited Children&lt;/li&gt;
&lt;li&gt;OpenAI is working with the Tech Coalition and ROOST to reach more organizations&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;OpenAI &lt;a href="https://x.com/OpenAINewsroom/status/2122375519853559954" rel="noopener noreferrer"&gt;announced&lt;/a&gt; the release of the classifier in a post. The company said it is also adding ChatGPT warnings, publishing guidance for developers and supporting child safety investigators.&lt;/p&gt;

&lt;h2&gt;
  
  
  What The Classifier Does
&lt;/h2&gt;

&lt;p&gt;The tool is built to help detect child sexual abuse material. OpenAI said that includes AI-generated images.&lt;/p&gt;

&lt;p&gt;It is not open to everyone. OpenAI said it is releasing the tool to vetted organizations, and it is free for them.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Also Read:&lt;/strong&gt; &lt;a href="https://www.fathom.news/child-abuse-images-phones/" rel="noopener noreferrer"&gt;UK Lawmakers Push Apple, Google To Block Child Abuse Images On All Phones&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  More Steps Beyond The Tool
&lt;/h2&gt;

&lt;p&gt;OpenAI said the classifier was developed with assistance from the National Center for Missing &amp;amp; Exploited Children.&lt;/p&gt;

&lt;p&gt;The company described three other moves. It is adding ChatGPT warnings, publishing guidance for developers and supporting child safety investigators.&lt;/p&gt;

&lt;h2&gt;
  
  
  What Comes Next
&lt;/h2&gt;

&lt;p&gt;OpenAI said it is working with the Tech Coalition and ROOST to help make the tool available to more organizations.&lt;/p&gt;

&lt;p&gt;The post does not say when that wider access will begin. It also does not say how organizations are vetted or when the new ChatGPT warnings and developer guidance will arrive.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Read Next:&lt;/strong&gt; &lt;a href="https://www.fathom.news/chatgpt-ai-content-watermark-eu/" rel="noopener noreferrer"&gt;EU Forces OpenAI’s ChatGPT To Mark AI Content, But Detection Gap Remains&lt;/a&gt;&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://www.fathom.news/openai-child-safety-tool/" rel="noopener noreferrer"&gt;Fathom News&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>news</category>
      <category>technology</category>
      <category>machinelearning</category>
    </item>
    <item>
      <title>Meta Poll Finds 94% Of Singapore Small Firms Need More AI Guidance</title>
      <dc:creator>Fathom News</dc:creator>
      <pubDate>Mon, 16 Nov 2026 09:00:37 +0000</pubDate>
      <link>https://dev.to/fathomnews/meta-poll-finds-94-of-singapore-small-firms-need-more-ai-guidance-3o70</link>
      <guid>https://dev.to/fathomnews/meta-poll-finds-94-of-singapore-small-firms-need-more-ai-guidance-3o70</guid>
      <description>&lt;p&gt;&lt;strong&gt;YouGov&lt;/strong&gt; found in &lt;strong&gt;Meta&lt;/strong&gt;-commissioned research that 94% of Singapore small business owners surveyed say small firms need more guidance to get the full benefits of AI.&lt;/p&gt;

&lt;h2&gt;
  
  
  Key Takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;94% of Singapore small business owners surveyed agree that small businesses need more AI guidance&lt;/li&gt;
&lt;li&gt;82% of surveyed owners whose business has used AI report a positive impact on productivity over the past 12 months&lt;/li&gt;
&lt;li&gt;More than 180 small businesses, heartland merchants and creator-entrepreneurs attended Meta Upskill Day 2026 in Singapore&lt;/li&gt;
&lt;li&gt;Meta also introduced Dream Render, an AI web app that turns a product description or photo into a first video draft&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Meta &lt;a href="https://about.fb.com/news/2026/11/94-of-singapore-small-business-owners-surveyed-say-more-ai-guidance-is-needed/" rel="noopener noreferrer"&gt;announced&lt;/a&gt; the findings on Nov. 15 at Meta Upskill Day 2026 in Singapore, part of its Small Business Growth Academy. Senior Minister of State &lt;strong&gt;Low Yen Ling&lt;/strong&gt; attended as Guest of Honor. The research was conducted by &lt;strong&gt;YouGov&lt;/strong&gt;.&lt;/p&gt;

&lt;p&gt;Among owners who report a positive productivity impact, 62% use the time saved through AI to improve the quality of existing work.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Also Read:&lt;/strong&gt; &lt;a href="https://www.fathom.news/help-goldman-sachs-bets-500m-on-ai/" rel="noopener noreferrer"&gt;Goldman Sachs Bets $500M On AI Help For Small Firms&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  What The Survey Found
&lt;/h2&gt;

&lt;p&gt;The Academy helps owners use Meta's digital and AI tools to find customers, improve productivity and grow. Participants got hands-on Business Agent training, demonstrations and talks with Meta experts.&lt;/p&gt;

&lt;p&gt;Customer replies and content creation stood out as areas of interest among the owners surveyed in Singapore.&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Finding&lt;/th&gt;
&lt;th&gt;Share&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Agree small businesses need more AI guidance&lt;/td&gt;
&lt;td&gt;94%&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Owners whose business used AI report better productivity over 12 months&lt;/td&gt;
&lt;td&gt;82%&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Owners with a productivity gain who use saved time to improve existing work&lt;/td&gt;
&lt;td&gt;62%&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Have used or would be interested in AI-powered customer responses&lt;/td&gt;
&lt;td&gt;52%&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Say AI guidance and training for marketing and content creation would be useful&lt;/td&gt;
&lt;td&gt;46%&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  Business Agent And Dream Render
&lt;/h2&gt;

&lt;p&gt;The Business Agent masterclass showed how businesses can use agents to handle customer enquiries through WhatsApp and Messenger. Singapore communication training business &lt;strong&gt;SpeakEasy Lab&lt;/strong&gt; uses it with a two-person team. According to the business, routine questions that once took around an hour or longer now get almost immediate responses.&lt;/p&gt;

&lt;p&gt;The founders estimate the tool saves their team 10 to 15 working hours each week. "Meta Business Agent helps us respond to customers quickly, even when we’re teaching or away," said &lt;strong&gt;Mervyn Ye&lt;/strong&gt;, Co-founder of SpeakEasy Lab.&lt;/p&gt;

&lt;p&gt;Meta is also introducing Dream Render, an AI web app for small businesses and creators who produce product videos regularly. It turns a short product description or photo into a first draft in minutes.&lt;/p&gt;

&lt;p&gt;Users can adjust individual shots, framing, pacing, voiceover and on-screen text. They preview and approve before the final video is produced. Users download and publish the video themselves, with a label showing it was made with AI.&lt;/p&gt;

&lt;p&gt;The tool runs in a browser. People can sign in with the Facebook profile or Instagram professional account they already use, with nothing to install. Singaporean startup &lt;strong&gt;Haldy&lt;/strong&gt;, which makes sugar-free turmeric mints, has been using it for its video needs.&lt;/p&gt;

&lt;h2&gt;
  
  
  How Far The Findings Go
&lt;/h2&gt;

&lt;p&gt;YouGov conducted the research online among 3,983 small business owners across nine markets. They were Australia, Singapore, Vietnam, Indonesia, Taiwan, India, Thailand, Japan and Hong Kong. Data was collected between Aug. 6 and Sept. 2, 2026.&lt;/p&gt;

&lt;p&gt;Meta said the findings give a cross-section of small businesses in each market, rather than being nationally representative. It also said reported adoption and impact differ across the nine markets. These differences may partly reflect how owners interpret the questions, as well as translation and local survey response norms.&lt;/p&gt;

&lt;p&gt;Meta said that since its Singapore launch in 2021, Upskill with Meta has trained more than 500 small businesses and upskilled over 4,500 entrepreneurs.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Read Next:&lt;/strong&gt; &lt;a href="https://www.fathom.news/meta-funds-5-universities-for-science/" rel="noopener noreferrer"&gt;Meta Hands 5 US Universities AI Tools To Speed Medical Research&lt;/a&gt;&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://www.fathom.news/singapore-ai-guidance-small-business/" rel="noopener noreferrer"&gt;Fathom News&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>news</category>
      <category>technology</category>
      <category>machinelearning</category>
    </item>
    <item>
      <title>OpenAI's Agents API Lets Any App Rent Codex, Here's How It Works</title>
      <dc:creator>Fathom News</dc:creator>
      <pubDate>Mon, 16 Nov 2026 08:30:27 +0000</pubDate>
      <link>https://dev.to/fathomnews/openais-agents-api-lets-any-app-rent-codex-heres-how-it-works-2445</link>
      <guid>https://dev.to/fathomnews/openais-agents-api-lets-any-app-rent-codex-heres-how-it-works-2445</guid>
      <description>&lt;p&gt;&lt;strong&gt;OpenAI&lt;/strong&gt;'s Agents API, in public beta since Sept. 10, lets developers build cloud agents on its Codex harness, with OpenAI handling sessions, recovery and context compaction.&lt;/p&gt;

&lt;h2&gt;
  
  
  Key Takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;The Agents API gives an application access to the Codex harness through an OpenAI-managed API&lt;/li&gt;
&lt;li&gt;OpenAI manages sessions, orchestration, context compaction and recovery&lt;/li&gt;
&lt;li&gt;Agents can run code, edit files, connect to MCP servers and produce artifacts in a sandbox&lt;/li&gt;
&lt;li&gt;Model usage is billed at the selected model's API rates, and hosted sandboxes use standard container rates&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;OpenAI &lt;a href="https://developers.openai.com/api/docs/guides/agents-api/overview" rel="noopener noreferrer"&gt;published&lt;/a&gt; documentation describing the Agents API as a way to build durable cloud agents with a managed Codex harness. The application provides the tools and chooses its execution environment.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Also Read:&lt;/strong&gt; &lt;a href="https://www.fathom.news/openai-model-spec-sets-rules-stop-ai/" rel="noopener noreferrer"&gt;OpenAI Sets 5 Rules To Stop AI Agents Going Off Script&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  How A Session Works
&lt;/h2&gt;

&lt;p&gt;The docs say a session is a durable instance of an agent that works on tasks and responds to input. With an OpenAI-hosted session, the application sends input and receives events. OpenAI runs the agent and provisions and manages its sandbox.&lt;/p&gt;

&lt;p&gt;The docs list five steps. First, create a session. Second, configure the agent while OpenAI provisions its environment. Third, give it a task, which starts a turn of work once the environment is ready. Fourth, follow progress by streaming output or using webhooks to learn when the agent finishes or needs input. Fifth, send another task to the same session or guide the agent during its current turn.&lt;/p&gt;

&lt;h2&gt;
  
  
  The Four Core Concepts
&lt;/h2&gt;

&lt;p&gt;The Agents API is built around four main concepts, according to the documentation.&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Concept&lt;/th&gt;
&lt;th&gt;What the docs say&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Agent&lt;/td&gt;
&lt;td&gt;The model, instructions, tools and MCP servers available to the agent&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Environment&lt;/td&gt;
&lt;td&gt;An optional sandbox or computer where the agent accesses files, loads skills and runs commands&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Session&lt;/td&gt;
&lt;td&gt;A durable instance of an agent that works on tasks and responds to input&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Events and items&lt;/td&gt;
&lt;td&gt;The inputs sent to an agent and the output produced during a session&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  Sandboxes, Examples And Pricing
&lt;/h2&gt;

&lt;p&gt;Agents can operate in a sandbox where they can execute code, edit files, connect to MCP servers and produce artifacts. A code sample in the docs sets up a self-hosted environment with a workspace directory at /workspace. The same sample turns on multi-agent mode with a maximum of 4 concurrent subagents and uses the model gpt-6-astra. Requests in the curl example carry an OpenAI-Beta header set to agents=v1.&lt;/p&gt;

&lt;p&gt;The docs point to several complete examples. They include an incident response agent, a Slack bot, a data analyst that answers warehouse questions with read-only SQL, a GitHub issue investigator and a document reviewer. Another example compares release notes with subagents and combines their findings into one answer.&lt;/p&gt;

&lt;p&gt;On cost, the docs say model usage is billed at the selected model's API rates. OpenAI tools use their standard rates. OpenAI-hosted sandboxes use standard container rates.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Read Next:&lt;/strong&gt; &lt;a href="https://www.fathom.news/openai-agents-36-hours-third-moment/" rel="noopener noreferrer"&gt;OpenAI’s 10,000 Agents Run 36 Hours Straight With Zero Human Checks&lt;/a&gt;&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://www.fathom.news/openai-agents-api-codex-harness/" rel="noopener noreferrer"&gt;Fathom News&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>news</category>
      <category>machinelearning</category>
      <category>research</category>
    </item>
    <item>
      <title>Jensen Huang Reportedly Eyes Nvidia Comeback As Elon Musk Backs Him</title>
      <dc:creator>Fathom News</dc:creator>
      <pubDate>Mon, 16 Nov 2026 00:04:08 +0000</pubDate>
      <link>https://dev.to/fathomnews/jensen-huang-reportedly-eyes-nvidia-comeback-as-elon-musk-backs-him-2p6l</link>
      <guid>https://dev.to/fathomnews/jensen-huang-reportedly-eyes-nvidia-comeback-as-elon-musk-backs-him-2p6l</guid>
      <description>&lt;p&gt;&lt;strong&gt;Jensen Huang&lt;/strong&gt; is reportedly weighing a return as &lt;strong&gt;Nvidia&lt;/strong&gt;'s CEO just days after quitting, while tech leaders including &lt;strong&gt;Elon Musk&lt;/strong&gt; publicly defend him and a petition pushes the board to bring him back.&lt;/p&gt;

&lt;h2&gt;
  
  
  Key Takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Huang has discussed a comeback with allies, according to a Sunday report, but neither he nor Nvidia has commented&lt;/li&gt;
&lt;li&gt;Elon Musk, Tim Sweeney and Gavin Baker have backed Huang in their own posts on X&lt;/li&gt;
&lt;li&gt;A "Stand with Jensen" open letter asks Nvidia's board to reinstate him while an independent investigation runs&lt;/li&gt;
&lt;li&gt;Nvidia reports quarterly results on Tuesday, Nov. 17, with CFO Colette Kress as interim CEO&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Musk &lt;a href="https://x.com/elonmusk/status/2121624209675665519" rel="noopener noreferrer"&gt;wrote&lt;/a&gt; on X on Nov. 14 that "if Jensen gave me his word about compute delivery or I asked for components to be accelerated, he always came through." Nvidia said on Nov. 12 that Huang had resigned as CEO and left its board with immediate effect, and that the board had opened an investigation.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Also Read:&lt;/strong&gt; &lt;a href="https://www.fathom.news/nvidia-stock-huang-resigns/" rel="noopener noreferrer"&gt;Nvidia Crashes 12% As Jensen Huang Quits, And Wall Street Slashes Targets To $200&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  What The Comeback Talk Says
&lt;/h2&gt;

&lt;p&gt;A report on Sunday said Huang has been talking with allies about returning to the job he held for 33 years, and that supporters in Silicon Valley have set up a "war room" to lobby the board. None of this has been confirmed by Huang, Nvidia or its board.&lt;/p&gt;

&lt;p&gt;Nvidia's board is looking at internal and external candidates for a permanent CEO. Kress is running the company in the meantime, and Harvey C. Jones became the board's first-ever chair.&lt;/p&gt;

&lt;h2&gt;
  
  
  Who Is Backing Huang In Public
&lt;/h2&gt;

&lt;p&gt;Musk also hit back at critics of Huang, writing that one writer was "projecting his own mean spirit onto Jensen." &lt;strong&gt;Epic Games&lt;/strong&gt; CEO &lt;strong&gt;Tim Sweeney&lt;/strong&gt; posted "Save Jensen!" on Sunday.&lt;/p&gt;

&lt;p&gt;Investor &lt;strong&gt;Gavin Baker&lt;/strong&gt; of &lt;strong&gt;Atreides Management&lt;/strong&gt; wrote that he supports "the petition to immediately reinstate Jensen while awaiting the results of the independent investigation." He added that it is possible to support both Huang and Kress.&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Who&lt;/th&gt;
&lt;th&gt;What They Said On X&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Elon Musk&lt;/td&gt;
&lt;td&gt;Huang "always came through" on his word&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Tim Sweeney, Epic Games&lt;/td&gt;
&lt;td&gt;"Save Jensen!"&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Gavin Baker, Atreides Management&lt;/td&gt;
&lt;td&gt;Supports reinstating Huang while the investigation runs&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Nvidia&lt;/td&gt;
&lt;td&gt;No comment on a return&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  The Petition Pushing The Board
&lt;/h2&gt;

&lt;p&gt;The "Stand with Jensen" open letter describes itself as an independent effort by people who work or have worked at Nvidia and hold or have held its stock. It asks the board for a transparent account of the CEO change, and its supporters say it has passed 150,000 signatures. Nvidia has not responded to it publicly.&lt;/p&gt;

&lt;h2&gt;
  
  
  What Happens Next
&lt;/h2&gt;

&lt;p&gt;All eyes are on Tuesday's earnings report, the first since Huang left. Investors will be listening for any word from the board on the investigation, the CEO search or Huang himself.&lt;/p&gt;

&lt;p&gt;A return would need the board's approval, and the investigation that came with his exit is still open. Until Nvidia or Huang says something on the record, the comeback is talk, not a plan.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Read Next:&lt;/strong&gt; &lt;a href="https://www.fathom.news/nvidia-jensen-huang/" rel="noopener noreferrer"&gt;Nvidia Board Reportedly Moves To Oust Jensen Huang, Here’s Why Its AI Boom Can Survive Him&lt;/a&gt;&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://www.fathom.news/jensen-huang-nvidia-comeback/" rel="noopener noreferrer"&gt;Fathom News&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>news</category>
      <category>technology</category>
      <category>machinelearning</category>
    </item>
    <item>
      <title>Microsoft Playwright Now Helps Ai Agents Run Browsers And Tests</title>
      <dc:creator>Fathom News</dc:creator>
      <pubDate>Mon, 16 Nov 2026 00:03:26 +0000</pubDate>
      <link>https://dev.to/fathomnews/microsoft-playwright-now-helps-ai-agents-run-browsers-and-tests-m70</link>
      <guid>https://dev.to/fathomnews/microsoft-playwright-now-helps-ai-agents-run-browsers-and-tests-m70</guid>
      <description>&lt;p&gt;&lt;strong&gt;Microsoft&lt;/strong&gt; has an open source browser framework called Playwright that now doubles as a tool for AI agents, and its GitHub page shows 100.2k stars.&lt;/p&gt;

&lt;h2&gt;
  
  
  Key Takeaways
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Playwright drives Chromium, Firefox and WebKit with a single API&lt;/li&gt;
&lt;li&gt;The GitHub repository shows 100.2k stars, 6.9k forks and 18,491 commits&lt;/li&gt;
&lt;li&gt;A CLI for coding agents is described as more token-efficient than MCP&lt;/li&gt;
&lt;li&gt;The MCP server lets agents control a browser using structured accessibility snapshots&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The project's GitHub page &lt;a href="https://github.com/microsoft/playwright" rel="noopener noreferrer"&gt;said&lt;/a&gt; Playwright is a framework for web automation and testing. It works in tests, in scripts and as a tool for AI agents. The code is released under the Apache-2.0 license.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Also Read:&lt;/strong&gt; &lt;a href="https://www.fathom.news/gpt-6-1-terra-github-gives-all-copilot/" rel="noopener noreferrer"&gt;GitHub Gives All Copilot Users OpenAI Tool That Solves More Code&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Several Ways To Use Playwright
&lt;/h2&gt;

&lt;p&gt;The page lists several install paths. Playwright Test is a test runner for end-to-end testing. It runs tests in parallel with full browser isolation, auto-waiting and web-first assertions.&lt;/p&gt;

&lt;p&gt;Playwright Library is for scripts such as web scraping, PDF generation and screenshot capture. The VS Code extension adds test running, debugging and code generation inside the editor. Playwright is also available for Python, .NET and Java.&lt;/p&gt;

&lt;h2&gt;
  
  
  Built For AI Agents
&lt;/h2&gt;

&lt;p&gt;Playwright CLI is a command-line interface designed for coding agents, with Claude Code and Copilot named as examples. The page said it is more token-efficient than MCP because commands avoid loading large tool schemas and accessibility trees into the model context.&lt;/p&gt;

&lt;p&gt;The Playwright MCP server gives AI agents full browser control through the Model Context Protocol. The page said agents interact with pages using structured accessibility snapshots, with no vision models or screenshots required. Agents use element refs to click and type.&lt;/p&gt;

&lt;p&gt;The page said the MCP tools cover navigation, form filling, screenshots, network mocking and storage management. Setup works with clients such as VS Code, Cursor, Claude Desktop and Windsurf.&lt;/p&gt;

&lt;h2&gt;
  
  
  Browser Support Across Platforms
&lt;/h2&gt;

&lt;p&gt;The page said headless and headed execution works on all platforms. Chromium uses Chrome for Testing by default.&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Browser&lt;/th&gt;
&lt;th&gt;Version listed&lt;/th&gt;
&lt;th&gt;Platforms&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Chromium&lt;/td&gt;
&lt;td&gt;159.0.8154.6&lt;/td&gt;
&lt;td&gt;Linux, macOS, Windows&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;WebKit&lt;/td&gt;
&lt;td&gt;27.2&lt;/td&gt;
&lt;td&gt;Linux, macOS, Windows&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Firefox&lt;/td&gt;
&lt;td&gt;159.0&lt;/td&gt;
&lt;td&gt;Linux, macOS, Windows&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;&lt;strong&gt;Read Next:&lt;/strong&gt; &lt;a href="https://www.fathom.news/google-researchers-publish-method-to-stop-ai-agents-memorizing-their-own-tests/" rel="noopener noreferrer"&gt;Google Researchers Publish Method to Stop AI Agents Memorizing Their Own Tests&lt;/a&gt;&lt;/p&gt;




&lt;p&gt;&lt;em&gt;Originally published on &lt;a href="https://www.fathom.news/microsoft-playwright-100k-github-stars/" rel="noopener noreferrer"&gt;Fathom News&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>news</category>
      <category>technology</category>
      <category>machinelearning</category>
    </item>
  </channel>
</rss>
