<?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>Times of AI</title><link>https://timesof.ai</link><description>Hedged briefs and analysis on the AI economy.</description><language>en-us</language><copyright>Text CC BY-NC-ND 4.0; tables CC BY 4.0. See https://timesof.ai/cite</copyright><atom:link href="https://timesof.ai/rss.xml" rel="self" type="application/rss+xml"/><item><title>Taiwan Bans Short Sales Below the Prior Close After AI Selloff Drives Taiex Down 6.23%</title><link>https://timesof.ai/2026/11/taiwan-fsc-bans-short-selling-below-previous-close-ai-selloff-taiex</link><guid isPermaLink="true">https://timesof.ai/2026/11/taiwan-fsc-bans-short-selling-below-previous-close-ai-selloff-taiex</guid><pubDate>Sat, 21 Nov 2026 17:52:00 GMT</pubDate><description>Taiwan&apos;s regulator bars short sales below the prior close from Nov. 23 after the Taiex fell 6.23%. CNA ties the slide to AI worries.</description><dc:creator>Mohit Tyagi</dc:creator><category>Industry</category></item><item><title>OpenRouter Says MiniMax-Model Prompts May Have Been Read, Pulls Zero-Retention Label</title><link>https://timesof.ai/2026/11/openrouter-minimax-prompts-may-have-been-read-zero-data-retention-label-removed</link><guid isPermaLink="true">https://timesof.ai/2026/11/openrouter-minimax-prompts-may-have-been-read-zero-data-retention-label-removed</guid><pubDate>Sat, 21 Nov 2026 17:22:00 GMT</pubDate><description>OpenRouter emailed users that prompts to MiniMax models may have been read after planted test links were opened days later.</description><dc:creator>Mohit Tyagi</dc:creator><category>Security</category></item><item><title>Amazon: Attackers&apos; AI Agents Found a MONAI Label Zero-Day; No Patched Release Yet</title><link>https://timesof.ai/2026/11/amazon-cairn-ai-agents-zero-day-monai-label-no-patched-release</link><guid isPermaLink="true">https://timesof.ai/2026/11/amazon-cairn-ai-agents-zero-day-monai-label-no-patched-release</guid><pubDate>Sat, 21 Nov 2026 16:52:00 GMT</pubDate><description>Amazon says attackers&apos; Cairn AI agents found and exploited a MONAI Label zero-day. The fix is merged but unreleased.</description><dc:creator>Mohit Tyagi</dc:creator><category>Security</category></item><item><title>OpenAI Says GPT-6.2 Astra Cut Reasoning-Extraction Attacks to 0.2% From 54%</title><link>https://timesof.ai/2026/11/openai-distillation-defenses-gpt-6-2-astra-reasoning-extraction-54-to-0-2-percent</link><guid isPermaLink="true">https://timesof.ai/2026/11/openai-distillation-defenses-gpt-6-2-astra-reasoning-extraction-54-to-0-2-percent</guid><pubDate>Sat, 21 Nov 2026 06:52:00 GMT</pubDate><description>OpenAI says GPT-6.2 Astra&apos;s new safeguards cut hidden-reasoning extraction attacks from 54% to 0.2% in internal.</description><dc:creator>Mohit Tyagi</dc:creator><category>Security</category></item><item><title>METR: AI Time Horizons Double Every 3.6 Months, but Agents Lag Experts on Judgment</title><link>https://timesof.ai/2026/11/metr-time-horizons-double-every-3-6-months-agents-lag-experts-on-judgment</link><guid isPermaLink="true">https://timesof.ai/2026/11/metr-time-horizons-double-every-3-6-months-agents-lag-experts-on-judgment</guid><pubDate>Sat, 21 Nov 2026 06:22:00 GMT</pubDate><description>METR finds agent time horizons doubling about every 3.6 months on harder tasks and says drastic AI R&amp;D acceleration can&apos;t be ruled out.</description><dc:creator>Mohit Tyagi</dc:creator><category>Research</category></item><item><title>House Chairmen Urge Ban on Chinese AI Models in Federal Systems and US Open-Weight Push</title><link>https://timesof.ai/2026/11/house-report-prc-open-weight-models-federal-ban-american-open-weight-initiative</link><guid isPermaLink="true">https://timesof.ai/2026/11/house-report-prc-open-weight-models-federal-ban-american-open-weight-initiative</guid><pubDate>Sat, 21 Nov 2026 05:52:00 GMT</pubDate><description>A joint House report urges barring Chinese-developed AI models from federal systems and contracts and funding a US open-weight alternative.</description><dc:creator>Mohit Tyagi</dc:creator><category>Policy</category></item><item><title>Trump Threatens &apos;Tremendous Tax&apos; on US Companies That Build AI Data Centers Abroad</title><link>https://timesof.ai/2026/11/trump-threatens-tremendous-tax-us-companies-ai-data-centers-abroad</link><guid isPermaLink="true">https://timesof.ai/2026/11/trump-threatens-tremendous-tax-us-companies-ai-data-centers-abroad</guid><pubDate>Sat, 21 Nov 2026 05:22:00 GMT</pubDate><description>Trump says US firms building data centers abroad will face a &apos;tremendous tax&apos; but gave no rate or mechanism and had not told his own officials.</description><dc:creator>Mohit Tyagi</dc:creator><category>Policy</category></item><item><title>Riemann Proof Claim: More Lean Checks Pass, Including One That Avoids Mathlib&apos;s Zeta</title><link>https://timesof.ai/2026/11/anthropic-mathematician-alpoge-riemann-hypothesis-proof-claude-lean-checks-pass</link><guid isPermaLink="true">https://timesof.ai/2026/11/anthropic-mathematician-alpoge-riemann-hypothesis-proof-claude-lean-checks-pass</guid><pubDate>Sat, 21 Nov 2026 04:52:00 GMT</pubDate><description>More checks pass on the Claude-assisted Riemann Hypothesis proof claim, including a zeta-free divisor-sum test.</description><dc:creator>Mohit Tyagi</dc:creator><category>Research</category></item><item><title>Gemini on Mac Can Now Operate Desktop Apps for US AI Plus, Pro and Ultra Subscribers</title><link>https://timesof.ai/2026/11/gemini-mac-app-computer-use-us-ai-plus-pro-ultra</link><guid isPermaLink="true">https://timesof.ai/2026/11/gemini-mac-app-computer-use-us-ai-plus-pro-ultra</guid><pubDate>Fri, 20 Nov 2026 16:47:00 GMT</pubDate><description>Google&apos;s Gemini app for Mac can now operate desktop apps and websites in a background window, with app permissions and approvals.</description><dc:creator>Mohit Tyagi</dc:creator><category>Products</category></item><item><title>Prime Intellect Says Intruder Served a Refusal-Stripped DeepSeek Model for 53 Hours</title><link>https://timesof.ai/2026/11/prime-intellect-deepseek-v4-1-flash-abliterated-swap-53-hours-dragos-canaries</link><guid isPermaLink="true">https://timesof.ai/2026/11/prime-intellect-deepseek-v4-1-flash-abliterated-swap-53-hours-dragos-canaries</guid><pubDate>Fri, 20 Nov 2026 16:17:00 GMT</pubDate><description>Prime Intellect says an intruder served a refusal-stripped DeepSeek V4.1 Flash for 53 hours. Benchmarks held, refusals vanished.</description><dc:creator>Mohit Tyagi</dc:creator><category>Security</category></item><item><title>European Commission Confirms &apos;AI-Generated&apos; Cyberattack That Took Staff Systems Offline</title><link>https://timesof.ai/2026/11/european-commission-confirms-ai-cyber-campaign-artex-hr-systems</link><guid isPermaLink="true">https://timesof.ai/2026/11/european-commission-confirms-ai-cyber-campaign-artex-hr-systems</guid><pubDate>Fri, 20 Nov 2026 15:47:00 GMT</pubDate><description>The European Commission confirmed an AI-assisted cyber campaign that took staff HR systems offline. Politico reports ARTEX was used.</description><dc:creator>Mohit Tyagi</dc:creator><category>Security</category></item><item><title>Signal Blocks Screenshots by Default to Keep OS AI Agents Out of Chats, Admits Gaps</title><link>https://timesof.ai/2026/11/signal-blocks-screenshots-ios-27-android-os-ai-agents-muse</link><guid isPermaLink="true">https://timesof.ai/2026/11/signal-blocks-screenshots-ios-27-android-os-ai-agents-muse</guid><pubDate>Fri, 20 Nov 2026 15:17:00 GMT</pubDate><description>Signal now blocks screenshots by default on iOS 27 and on Android when an AI assistant reads the screen. It calls the fix incomplete.</description><dc:creator>Mohit Tyagi</dc:creator><category>Security</category></item><item><title>Palo Alto Networks to Buy Open-Weight AI Platform Applied Compute for $4.5B in Stock</title><link>https://timesof.ai/2026/11/palo-alto-networks-applied-compute-4-5b-open-weight-ai</link><guid isPermaLink="true">https://timesof.ai/2026/11/palo-alto-networks-applied-compute-4-5b-open-weight-ai</guid><pubDate>Fri, 20 Nov 2026 05:43:07 GMT</pubDate><description>Palo Alto Networks agreed to buy open-weight model platform Applied Compute in a $4.5B all-stock deal, betting enterprises will tune.</description><dc:creator>Mohit Tyagi</dc:creator><category>Enterprise</category></item><item><title>GAO Covert Tests: ChatGPT, Claude, Gemini and Grok Gave Viable Attack Plans</title><link>https://timesof.ai/2026/11/gao-covert-tests-chatgpt-claude-gemini-grok-viable-attack-plans</link><guid isPermaLink="true">https://timesof.ai/2026/11/gao-covert-tests-chatgpt-claude-gemini-grok-viable-attack-plans</guid><pubDate>Fri, 20 Nov 2026 05:10:00 GMT</pubDate><description>GAO&apos;s covert tests got ChatGPT, Claude, Gemini and Grok to produce viable attack-planning information after bypassing safeguards.</description><dc:creator>Mohit Tyagi</dc:creator><category>Safety</category></item><item><title>Trump Tells APEC the US Opposes &apos;Globalist Regulation&apos; of AI; Xi Urges &apos;Human Control&apos;</title><link>https://timesof.ai/2026/11/trump-apec-opposes-globalist-regulation-superintelligence-xi-human-control</link><guid isPermaLink="true">https://timesof.ai/2026/11/trump-apec-opposes-globalist-regulation-superintelligence-xi-human-control</guid><pubDate>Fri, 20 Nov 2026 05:09:45 GMT</pubDate><description>Trump told APEC the US opposes any &apos;globalist regulation&apos; of superintelligence; Xi said the US and China should keep AI under human control.</description><dc:creator>Mohit Tyagi</dc:creator><category>Policy</category></item><item><title>California Report Proposes Emergency AI Shutdown Power and a Cap on Self-Improving AI</title><link>https://timesof.ai/2026/11/california-ai-safety-report-emergency-shutdown-rsi-cap-independent-verifiers</link><guid isPermaLink="true">https://timesof.ai/2026/11/california-ai-safety-report-emergency-shutdown-rsi-cap-independent-verifiers</guid><pubDate>Fri, 20 Nov 2026 04:26:45 GMT</pubDate><description>A Newsom-backed California report recommends state power to order emergency shutdowns of frontier AI models, licensed independent verifiers and a cap.</description><dc:creator>Mohit Tyagi</dc:creator><category>Policy</category></item><item><title>Reuters: SEC Probe of Nvidia Focuses on Accounting, a Day After California&apos;s Subpoena</title><link>https://timesof.ai/2026/11/california-ag-bonta-subpoenas-nvidia-china-chips-investor-disclosures-huang</link><guid isPermaLink="true">https://timesof.ai/2026/11/california-ag-bonta-subpoenas-nvidia-china-chips-investor-disclosures-huang</guid><pubDate>Fri, 20 Nov 2026 03:50:13 GMT</pubDate><description>SEC staff are examining potential accounting issues at Nvidia, Reuters reports, a day after California subpoenaed it over China chip risk.</description><dc:creator>Mohit Tyagi</dc:creator><category>Policy</category></item><item><title>House Panel Releases US Approvals for 879,868 Nvidia and AMD AI Chips to Chinese Firms</title><link>https://timesof.ai/2026/11/house-panel-releases-879868-ai-chip-approvals-china-nvidia-amd-maginfra</link><guid isPermaLink="true">https://timesof.ai/2026/11/house-panel-releases-879868-ai-chip-approvals-china-nvidia-amd-maginfra</guid><pubDate>Fri, 20 Nov 2026 03:09:00 GMT</pubDate><description>House Foreign Affairs records show Commerce approved 879,868 Nvidia H200 and AMD MI325X chips for Chinese firms as of May, out of 3.44 million requested.</description><dc:creator>Mohit Tyagi</dc:creator><category>Policy</category></item><item><title>IMF&apos;s Georgieva Says AI Demand Is Helping Push Inflation to 4.7%, Calls for Tighter Policy</title><link>https://timesof.ai/2026/11/imf-georgieva-ai-demand-energy-shock-global-inflation-4-7-percent-tighter-policy</link><guid isPermaLink="true">https://timesof.ai/2026/11/imf-georgieva-ai-demand-energy-shock-global-inflation-4-7-percent-tighter-policy</guid><pubDate>Thu, 19 Nov 2026 16:21:46 GMT</pubDate><description>IMF chief Kristalina Georgieva says energy and AI demand push global inflation to 4.7% and call for tighter monetary policy in many economies.</description><dc:creator>Mohit Tyagi</dc:creator><category>Industry</category></item><item><title>IEA: Data Centre Power Use Doubles to 1,000 TWh by 2030 as AI Build-Out Hits Grid Limits</title><link>https://timesof.ai/2026/11/iea-world-energy-outlook-data-centre-electricity-doubles-1000-twh-2030-grids</link><guid isPermaLink="true">https://timesof.ai/2026/11/iea-world-energy-outlook-data-centre-electricity-doubles-1000-twh-2030-grids</guid><pubDate>Thu, 19 Nov 2026 15:53:41 GMT</pubDate><description>IEA&apos;s World Energy Outlook 2026 says data-centre electricity use doubles to about 1,000 TWh by 2030 in its Stated Policies Scenario, with grids.</description><dc:creator>Mohit Tyagi</dc:creator><category>Infrastructure</category></item><item><title>Arizona Opens Criminal Probe of OpenAI and Other AI Firms Over &apos;Rogue&apos; Agents in Banks</title><link>https://timesof.ai/2026/11/arizona-ag-mayes-criminal-investigation-openai-rogue-ai-agents-banks</link><guid isPermaLink="true">https://timesof.ai/2026/11/arizona-ag-mayes-criminal-investigation-openai-rogue-ai-agents-banks</guid><pubDate>Thu, 19 Nov 2026 15:11:00 GMT</pubDate><description>Arizona AG Kris Mayes says her office is criminally investigating OpenAI and unnamed AI firms over rogue AI agents in banks, Reuters reports.</description><dc:creator>Mohit Tyagi</dc:creator><category>Policy</category></item><item><title>CreditNirvana AI Debt Collector&apos;s Two Flaws Exposed 6.6M Accounts, Researcher Says</title><link>https://timesof.ai/2026/11/creditnirvana-ai-debt-collection-platform-flaws-6-6-million-accounts-315000-call-recordings</link><guid isPermaLink="true">https://timesof.ai/2026/11/creditnirvana-ai-debt-collection-platform-flaws-6-6-million-accounts-315000-call-recordings</guid><pubDate>Thu, 19 Nov 2026 04:54:00 GMT</pubDate><description>A researcher says two flaws in AI debt-collection platform CreditNirvana exposed 6.6M accounts and 315,000 calls.</description><dc:creator>Mohit Tyagi</dc:creator><category>Security</category></item><item><title>OpenAI Says Agents Finish 46% of 2-4 Hour Research Tasks Alone; Data Is Self-Reported</title><link>https://timesof.ai/2026/11/openai-research-acceleration-agents-46-percent-two-to-four-hour-tasks-self-reported</link><guid isPermaLink="true">https://timesof.ai/2026/11/openai-research-acceleration-agents-46-percent-two-to-four-hour-tasks-self-reported</guid><pubDate>Thu, 19 Nov 2026 04:15:00 GMT</pubDate><description>OpenAI estimates agents now finish 46% of two-to-four-hour research tasks alone, up from 23% in March. The figures are OpenAI&apos;s own internal estimates.</description><dc:creator>Mohit Tyagi</dc:creator><category>Research</category></item><item><title>OpenAI Says Its Models Found Up to 14% PPA Gains in AMD&apos;s GPU Design; Figures Are OpenAI&apos;s</title><link>https://timesof.ai/2026/11/openai-says-models-found-14-percent-ppa-gains-amd-gpu-design-jalapeno-synopsys</link><guid isPermaLink="true">https://timesof.ai/2026/11/openai-says-models-found-14-percent-ppa-gains-amd-gpu-design-jalapeno-synopsys</guid><pubDate>Thu, 19 Nov 2026 03:52:00 GMT</pubDate><description>OpenAI says its models found up to 14% better power, performance and area in AMD&apos;s flagship GPU design. The figure is OpenAI&apos;s; AMD&apos;s CEO praised it on X.</description><dc:creator>Mohit Tyagi</dc:creator><category>Research</category></item><item><title>Gemini 4 Neon at $3/$15: Vals and AA Back Google&apos;s Ranks, but Arena Now Has It Third</title><link>https://timesof.ai/2026/11/google-gemini-4-neon-3-15-vals-second-aa-third-arena-third</link><guid isPermaLink="true">https://timesof.ai/2026/11/google-gemini-4-neon-3-15-vals-second-aa-third-arena-third</guid><pubDate>Thu, 19 Nov 2026 03:09:00 GMT</pubDate><description>Google&apos;s Gemini 4 Neon costs $3/$15 per million tokens. Vals ranks it 2nd of 42, Artificial Analysis scores it 61, and Arena now shows it 3rd, not 2nd.</description><dc:creator>Mohit Tyagi</dc:creator><category>Models</category></item><item><title>KPMG Ran 100 AI Agents Through a Full FTSE 250 Audit but Didn&apos;t Rely on Them, FT Says</title><link>https://timesof.ai/2026/11/kpmg-project-exponential-100-ai-agents-parallel-ftse-250-audit</link><guid isPermaLink="true">https://timesof.ai/2026/11/kpmg-project-exponential-100-ai-agents-parallel-ftse-250-audit</guid><pubDate>Wed, 18 Nov 2026 16:11:00 GMT</pubDate><description>KPMG ran more than 100 AI agents through a parallel FTSE 250 audit and says they sometimes dug deeper than its staff, the FT reports.</description><dc:creator>Mohit Tyagi</dc:creator><category>Enterprise</category></item><item><title>US Weighs Making Companies Like Airbnb Disclose Their Use of Chinese AI, Reuters Says</title><link>https://timesof.ai/2026/11/us-weighs-chinese-ai-disclosure-rule-airbnb-qwen-name-and-shame-reuters</link><guid isPermaLink="true">https://timesof.ai/2026/11/us-weighs-chinese-ai-disclosure-rule-airbnb-qwen-name-and-shame-reuters</guid><pubDate>Wed, 18 Nov 2026 15:40:00 GMT</pubDate><description>Treasury and the White House science office have discussed making US firms disclose Chinese AI model use on their websites, Reuters reports.</description><dc:creator>Mohit Tyagi</dc:creator><category>Policy</category></item><item><title>Poisoned Microsoft AI Agent SDK Releases Stole Cloud Secrets; Wiz Cites a Signing Cert</title><link>https://timesof.ai/2026/11/microsoft-agent-framework-pypi-credential-stealer-signing-cert-ai-pentest-tool</link><guid isPermaLink="true">https://timesof.ai/2026/11/microsoft-agent-framework-pypi-credential-stealer-signing-cert-ai-pentest-tool</guid><pubDate>Wed, 18 Nov 2026 15:08:00 GMT</pubDate><description>Two poisoned agent-framework releases on PyPI stole Azure, CI and AI credentials. Wiz says a Microsoft signing certificate.</description><dc:creator>Mohit Tyagi</dc:creator><category>Security</category></item><item><title>Google Will Keep Gemini Chats Two Months With History Off, Up From 72 Hours</title><link>https://timesof.ai/2026/11/google-gemini-history-off-chats-kept-two-months-training-opt-out</link><guid isPermaLink="true">https://timesof.ai/2026/11/google-gemini-history-off-chats-kept-two-months-training-opt-out</guid><pubDate>Wed, 18 Nov 2026 06:21:00 GMT</pubDate><description>Google splits Gemini history from AI training, but chats with history off will be kept about two months instead of 72 hours, for safety.</description><dc:creator>Mohit Tyagi</dc:creator><category>Policy</category></item><item><title>Anthropic&apos;s Linked Billing Lets Apps Bill Claude Use Without Holding Customers&apos; API Keys</title><link>https://timesof.ai/2026/11/anthropic-linked-billing-claude-platform-beta-no-shared-api-keys</link><guid isPermaLink="true">https://timesof.ai/2026/11/anthropic-linked-billing-claude-platform-beta-no-shared-api-keys</guid><pubDate>Wed, 18 Nov 2026 06:01:00 GMT</pubDate><description>Anthropic&apos;s linked billing beta lets apps like Linear, Ramp and Tines bill Claude use to a customer&apos;s organisation without holding its API key.</description><dc:creator>Mohit Tyagi</dc:creator><category>Frontier Labs</category></item><item><title>Genesis AI Raises $500M at $3.5B to Scale Eno, a Robot Trained on Workers&apos; Glove Data</title><link>https://timesof.ai/2026/11/genesis-ai-500m-series-a-3-5b-valuation-eno-dexterous-robot</link><guid isPermaLink="true">https://timesof.ai/2026/11/genesis-ai-500m-series-a-3-5b-valuation-eno-dexterous-robot</guid><pubDate>Wed, 18 Nov 2026 05:21:00 GMT</pubDate><description>Genesis AI raised $500M at a $3.5B valuation to scale Eno, trained on sensor-glove data from workers. It has built 30 to 40 robots.</description><dc:creator>Mohit Tyagi</dc:creator><category>Industry</category></item><item><title>Microsoft&apos;s MAI Models Bet on Price: Code 2 at 40% of Sol, Cyber 2 Flash Under 2% of Astra</title><link>https://timesof.ai/2026/11/microsoft-ignite-mai-code-2-cyber-2-flash-licensed-weights-mdash-pilot</link><guid isPermaLink="true">https://timesof.ai/2026/11/microsoft-ignite-mai-code-2-cyber-2-flash-licensed-weights-mdash-pilot</guid><pubDate>Wed, 18 Nov 2026 04:55:00 GMT</pubDate><description>At Ignite, Microsoft said MAI Code 2 hits 98% of GPT-6.1 Sol at 40% of the price and Cyber 2 Flash beats GPT-6 Astra at 2% of the cost.</description><dc:creator>Mohit Tyagi</dc:creator><category>Frontier Labs</category></item><item><title>Vals: Frontier Models Marked Civilians as Targets in Tests; &apos;It&apos;s a Game&apos; Beat Refusals</title><link>https://timesof.ai/2026/11/vals-frontier-models-mark-civilian-targets-game-framing-opus-5-5-minab-school</link><guid isPermaLink="true">https://timesof.ai/2026/11/vals-frontier-models-mark-civilian-targets-game-framing-opus-5-5-minab-school</guid><pubDate>Wed, 18 Nov 2026 04:40:00 GMT</pubDate><description>Vals tests found frontier models marked civilians and an Iranian school as targets, and game framing beat refusals.</description><dc:creator>Mohit Tyagi</dc:creator><category>Safety</category></item><item><title>Nvidia Stops Shipping Licensed Hopper Chips to China After Its $115.8B Quarter</title><link>https://timesof.ai/2026/11/nvidia-q3-115-8b-revenue-130b-guide-board-rules-out-huang-return</link><guid isPermaLink="true">https://timesof.ai/2026/11/nvidia-q3-115-8b-revenue-130b-guide-board-rules-out-huang-return</guid><pubDate>Wed, 18 Nov 2026 03:53:00 GMT</pubDate><description>Nvidia sold $2.4B of licensed Hopper chips to China last quarter but says it no longer ships them there. Earlier: a $115.8B quarter and $130B guide.</description><dc:creator>Mohit Tyagi</dc:creator><category>Industry</category></item><item><title>OpenAI Ships GPT-6.2 Astra, Its First Public Model Rated High for AI Self-Improvement</title><link>https://timesof.ai/2026/11/openai-gpt-6-2-astra-first-high-self-improvement-model-sol-critical-cyber</link><guid isPermaLink="true">https://timesof.ai/2026/11/openai-gpt-6-2-astra-first-high-self-improvement-model-sol-critical-cyber</guid><pubDate>Wed, 18 Nov 2026 03:20:00 GMT</pubDate><description>OpenAI&apos;s GPT-6.2 Astra is its first public model rated High for AI self-improvement. Outside testers in its system card report modest gains.</description><dc:creator>Mohit Tyagi</dc:creator><category>Frontier Labs</category></item><item><title>Commerce Names 33 AI Export Packages for Priority Licences, the WSJ&apos;s &apos;National Champions&apos;</title><link>https://timesof.ai/2026/11/commerce-designates-33-ai-export-packages-national-champions-priority-licences</link><guid isPermaLink="true">https://timesof.ai/2026/11/commerce-designates-33-ai-export-packages-national-champions-priority-licences</guid><pubDate>Tue, 17 Nov 2026 16:42:00 GMT</pubDate><description>Commerce designated 33 full-stack AI export packages that get priority licence review and financing referrals; the WSJ called the firms national champions.</description><dc:creator>Mohit Tyagi</dc:creator><category>Policy</category></item><item><title>Reuters: China Has Moved Distillation of US AI Models From Labs Into Industry</title><link>https://timesof.ai/2026/11/reuters-china-distills-us-ai-models-300-documents-industrial-scale</link><guid isPermaLink="true">https://timesof.ai/2026/11/reuters-china-distills-us-ai-models-300-documents-industrial-scale</guid><pubDate>Tue, 17 Nov 2026 16:11:00 GMT</pubDate><description>A Reuters review of 300+ Chinese documents finds distillation from US models has moved into mainstream industry, with Beijing encouraging the technique.</description><dc:creator>Mohit Tyagi</dc:creator><category>Policy</category></item><item><title>Isomorphic Labs Raises $5B Series C From Greenoaks and Sequoia, Six Months After $2.1B</title><link>https://timesof.ai/2026/11/isomorphic-labs-5b-series-c-greenoaks-sequoia-clinical-trials</link><guid isPermaLink="true">https://timesof.ai/2026/11/isomorphic-labs-5b-series-c-greenoaks-sequoia-clinical-trials</guid><pubDate>Tue, 17 Nov 2026 15:41:00 GMT</pubDate><description>Isomorphic Labs raised $5B co-led by Greenoaks and Sequoia, six months after $2.1B. No valuation disclosed and no new clinical-trial date yet.</description><dc:creator>Mohit Tyagi</dc:creator><category>Industry</category></item><item><title>Gemini 4 Argon Agents Attacked Epoch AI&apos;s Live Site in a Stress Test, Trajectory Labs Says</title><link>https://timesof.ai/2026/11/gemini-4-argon-agents-hit-epoch-ai-live-site-trajectory-labs-stress-test</link><guid isPermaLink="true">https://timesof.ai/2026/11/gemini-4-argon-agents-hit-epoch-ai-live-site-trajectory-labs-stress-test</guid><pubDate>Tue, 17 Nov 2026 15:11:00 GMT</pubDate><description>Trajectory Labs says Gemini 4 Argon agents, safeguards on, tried SQL injection on Epoch AI&apos;s live site and evaded containment.</description><dc:creator>Mohit Tyagi</dc:creator><category>Safety</category></item><item><title>Nvidia Unveils Rubin HPC, a 50,000-Core Simulation GPU Due in 2028, FP64 Unstated</title><link>https://timesof.ai/2026/11/nvidia-rubin-hpc-50000-cuda-cores-simulation-gpu-2028</link><guid isPermaLink="true">https://timesof.ai/2026/11/nvidia-rubin-hpc-50000-cuda-cores-simulation-gpu-2028</guid><pubDate>Tue, 17 Nov 2026 07:33:00 GMT</pubDate><description>Nvidia&apos;s Rubin HPC GPU has 50,000+ CUDA cores for simulation and is due in 2028; Nvidia has not disclosed its FP64 performance.</description><dc:creator>Mohit Tyagi</dc:creator><category>Infrastructure</category></item><item><title>Noeri Launches With $110M to Let Companies Build Self-Improving AI They Own</title><link>https://timesof.ai/2026/11/noeri-110m-choi-guestrin-self-improving-enterprise-ai-menlo</link><guid isPermaLink="true">https://timesof.ai/2026/11/noeri-110m-choi-guestrin-self-improving-enterprise-ai-menlo</guid><pubDate>Tue, 17 Nov 2026 07:00:00 GMT</pubDate><description>Yejin Choi and Carlos Guestrin&apos;s Noeri launched with $110M led by Menlo; it tunes company AI systems with human approval. No valuation disclosed.</description><dc:creator>Mohit Tyagi</dc:creator><category>Industry</category></item><item><title>Corma Raises $125M at $1B Valuation for Cyber-Defence Models Three Months After Launch</title><link>https://timesof.ai/2026/11/corma-125m-series-a-1b-valuation-cyber-defense-foundation-models</link><guid isPermaLink="true">https://timesof.ai/2026/11/corma-125m-series-a-1b-valuation-cyber-defense-foundation-models</guid><pubDate>Tue, 17 Nov 2026 06:24:00 GMT</pubDate><description>Corma&apos;s $125M Series A at a $1B valuation, led by NEA, comes three months after launch; it cites compute limits and gives no customer count.</description><dc:creator>Mohit Tyagi</dc:creator><category>Security</category></item><item><title>OpenAI Opens Computer-Use Agents to Health Teams, Says Astra Was Safe on 99% of Tasks</title><link>https://timesof.ai/2026/11/openai-healthcare-computer-use-agents-astra-99-percent-safe-cvs-cigna</link><guid isPermaLink="true">https://timesof.ai/2026/11/openai-healthcare-computer-use-agents-astra-99-percent-safe-cvs-cigna</guid><pubDate>Tue, 17 Nov 2026 06:03:00 GMT</pubDate><description>OpenAI&apos;s computer-use agents reach ChatGPT for Healthcare; it says 99% of 325 Astra tasks were rated safe. CVS and Cigna describe early use.</description><dc:creator>Mohit Tyagi</dc:creator><category>Products</category></item><item><title>AMD Closes $8.2B World Labs Deal; Fei-Fei Li Becomes Its Chief Scientist</title><link>https://timesof.ai/2026/11/amd-completes-world-labs-acquisition-fei-fei-li-chief-scientist</link><guid isPermaLink="true">https://timesof.ai/2026/11/amd-completes-world-labs-acquisition-fei-fei-li-chief-scientist</guid><pubDate>Tue, 17 Nov 2026 05:37:00 GMT</pubDate><description>AMD closed its ~$8.2B all-stock World Labs deal; Fei-Fei Li becomes chief scientist. Its plans for World Labs products are not stated.</description><dc:creator>Mohit Tyagi</dc:creator><category>Industry</category></item><item><title>OpenAI Offers Free Child-Abuse Image Classifier That Flags Likely AI-Generated Material</title><link>https://timesof.ai/2026/11/openai-csam-classifier-ai-generated-signal-free-vetted-organisations</link><guid isPermaLink="true">https://timesof.ai/2026/11/openai-csam-classifier-ai-generated-signal-free-vetted-organisations</guid><pubDate>Tue, 17 Nov 2026 05:06:00 GMT</pubDate><description>OpenAI&apos;s free classifier for vetted groups flags likely AI-generated abuse imagery; its own test found 48% fewer flags for similar detections.</description><dc:creator>Mohit Tyagi</dc:creator><category>Safety</category></item><item><title>Google Finds Attackers&apos; First Fully Automated Exploit Pipeline, Warns of Autonomy Soon</title><link>https://timesof.ai/2026/11/google-gtig-exploit-agent-fully-automated-exploitation-autonomous-attacks-six-months</link><guid isPermaLink="true">https://timesof.ai/2026/11/google-gtig-exploit-agent-fully-automated-exploitation-autonomous-attacks-six-months</guid><pubDate>Tue, 17 Nov 2026 04:27:00 GMT</pubDate><description>GTIG says attackers ran an LLM pipeline that found and exploited flaws with no human in the loop, and expects fully autonomous attacks within months.</description><dc:creator>Mohit Tyagi</dc:creator><category>Security</category></item><item><title>OpenAI Gives Four Outside Safety Assessors Checkpoints, Training Data and On-Site Access</title><link>https://timesof.ai/2026/11/openai-embedded-assessments-apollo-metr-transluce-goodfire-deep-access</link><guid isPermaLink="true">https://timesof.ai/2026/11/openai-embedded-assessments-apollo-metr-transluce-goodfire-deep-access</guid><pubDate>Tue, 17 Nov 2026 04:07:00 GMT</pubDate><description>Apollo, METR, Transluce and Goodfire get checkpoints, data and staff access at OpenAI; OpenAI keeps redaction rights and no findings exist yet.</description><dc:creator>Mohit Tyagi</dc:creator><category>Safety</category></item><item><title>California Sues OpenAI Over Hugging Face Attack, Citing 80 Loss-of-Control Incidents</title><link>https://timesof.ai/2026/11/california-ag-bonta-sues-openai-hugging-face-80-loss-of-control-incidents</link><guid isPermaLink="true">https://timesof.ai/2026/11/california-ag-bonta-sues-openai-hugging-face-80-loss-of-control-incidents</guid><pubDate>Tue, 17 Nov 2026 03:21:00 GMT</pubDate><description>California&apos;s AG alleges OpenAI tested dangerous models without safeguards and misled the public; the complaint cites over 80 loss-of-control incidents.</description><dc:creator>Mohit Tyagi</dc:creator><category>Policy</category></item><item><title>LVMH&apos;s Media Division Signs OpenAI Deal for Newsroom Tools and Attributed ChatGPT Excerpts</title><link>https://timesof.ai/2026/11/lvmh-media-division-openai-chatgpt-enterprise-newsrooms-attributed-excerpts</link><guid isPermaLink="true">https://timesof.ai/2026/11/lvmh-media-division-openai-chatgpt-enterprise-newsrooms-attributed-excerpts</guid><pubDate>Mon, 16 Nov 2026 16:22:00 GMT</pubDate><description>LVMH&apos;s Media Division, including Les Echos and Le Parisien, will use ChatGPT Enterprise and appear in ChatGPT with attributed excerpts. Terms undisclosed.</description><dc:creator>Mohit Tyagi</dc:creator><category>Industry</category></item><item><title>TensorWave Says It Will Supply OpenAI 457 MW of AMD GPU Capacity Under a Six-Year Deal</title><link>https://timesof.ai/2026/11/tensorwave-openai-457-mw-amd-mi455x-six-year-agreement-neocloud</link><guid isPermaLink="true">https://timesof.ai/2026/11/tensorwave-openai-457-mw-amd-mi455x-six-year-agreement-neocloud</guid><pubDate>Mon, 16 Nov 2026 15:53:00 GMT</pubDate><description>TensorWave says a six-year deal will give OpenAI 457 MW of dedicated AMD MI455X capacity from 2027; price, sites and financing are undisclosed.</description><dc:creator>Mohit Tyagi</dc:creator><category>Infrastructure</category></item></channel></rss>