<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0"><channel><title>Open vs Closed: every move, scored</title><link>https://openvsclosed.ai/</link><description>A running scoreboard of open versus closed AI: who leads on power, price and usage, every move scored, and the laws that will decide it.</description><language>en-us</language>
<item><title>House report: ban Chinese models in government, fund American open ones</title><link>https://openvsclosed.ai/plays/2026-11-20-house-report-ban-chinese-models-in-government-fund-american/</link><guid>https://openvsclosed.ai/plays/2026-11-20-house-report-ban-chinese-models-in-government-fund-american/</guid><pubDate>Fri, 20 Nov 2026 12:00:00 +0000</pubDate><category>Split</category><description>Two House committees publish a joint report after a seven-month investigation. It asks the White House to bar federal agencies from using Chinese models, and to launch a national American open-weight initiative with public computing power, shared testing and government purchasing behind it.</description></item>
<item><title>Cursor says it may drop Chinese open models</title><link>https://openvsclosed.ai/plays/2026-11-20-cursor-says-it-may-drop-chinese-open-models/</link><guid>https://openvsclosed.ai/plays/2026-11-20-cursor-says-it-may-drop-chinese-open-models/</guid><pubDate>Fri, 20 Nov 2026 12:00:00 +0000</pubDate><category>Point · Closed</category><description>Cursor, the AI coding tool, tells House investigators it will explore phasing out Chinese-developed open-weight models because of geopolitical risk. It has set no decision and no timetable. Cursor built its Composer 2 model on Moonshot&#x27;s Kimi K2.5.</description></item>
<item><title>Thinking Machines releases Inkling 2 Mini with open weights</title><link>https://openvsclosed.ai/plays/2026-11-20-thinking-machines-releases-inkling-2-mini-with-open-weights/</link><guid>https://openvsclosed.ai/plays/2026-11-20-thinking-machines-releases-inkling-2-mini-with-open-weights/</guid><pubDate>Fri, 20 Nov 2026 12:00:00 +0000</pubDate><category>Point · Open</category><description>Mira Murati&#x27;s Thinking Machines Lab releases Inkling 2 Mini, a 26.4-billion-parameter model, with open weights. It adds to a short list of American labs shipping open models.</description></item>
<item><title>White House weighs making companies disclose Chinese AI use</title><link>https://openvsclosed.ai/plays/2026-11-18-white-house-weighs-making-companies-disclose-chinese-ai-use/</link><guid>https://openvsclosed.ai/plays/2026-11-18-white-house-weighs-making-companies-disclose-chinese-ai-use/</guid><pubDate>Wed, 18 Nov 2026 12:00:00 +0000</pubDate><category>Point · Closed</category><description>Reuters reports the administration is considering a rule that would make U.S. companies publicly disclose their use of Chinese AI models. Three people familiar with the talks describe it as a way to discourage use without a ban that could anger industry or Beijing.</description></item>
<item><title>Federal commission urges limits on Chinese open models in sensitive uses</title><link>https://openvsclosed.ai/plays/2026-11-18-federal-commission-urges-limits-on-chinese-open-models-in-s/</link><guid>https://openvsclosed.ai/plays/2026-11-18-federal-commission-urges-limits-on-chinese-open-models-in-s/</guid><pubDate>Wed, 18 Nov 2026 12:00:00 +0000</pubDate><category>Point · Closed</category><description>The U.S.-China Economic and Security Review Commission sends Congress its annual report. Recommendation 18 would give Commerce 180 days to identify sensitive uses and block new deployments of Chinese open-weight models there unless their safeguards pass independent testing. Recommendation 19 would require disclosure when governments and businesses use them.</description></item>
<item><title>Commerce closes the overseas data center loophole</title><link>https://openvsclosed.ai/plays/2026-11-16-commerce-closes-the-overseas-data-center-loophole/</link><guid>https://openvsclosed.ai/plays/2026-11-16-commerce-closes-the-overseas-data-center-loophole/</guid><pubDate>Mon, 16 Nov 2026 12:00:00 +0000</pubDate><category>Split</category><description>The Commerce Department issues a rule requiring licenses for chip exports to overseas data centers that supply large amounts of computing power to companies based in China. It took effect November 19. The rule targets chips, not open models, but the labs it squeezes ship most of the world&#x27;s top open weights.</description></item>
<item><title>Tencent ships open-weight Hy4 Flash</title><link>https://openvsclosed.ai/plays/2026-11-16-tencent-ships-open-weight-hy4-flash/</link><guid>https://openvsclosed.ai/plays/2026-11-16-tencent-ships-open-weight-hy4-flash/</guid><pubDate>Mon, 16 Nov 2026 12:00:00 +0000</pubDate><category>Point · Open</category><description>Tencent&#x27;s Hy4 Flash arrives as an open-weight model: 220 billion parameters and a context window of up to one million tokens.</description></item>
<item><title>Comment window closes on China&#x27;s AI export rules</title><link>https://openvsclosed.ai/plays/2026-11-15-comment-window-closes-on-china-s-ai-export-rules/</link><guid>https://openvsclosed.ai/plays/2026-11-15-comment-window-closes-on-china-s-ai-export-rules/</guid><pubDate>Sun, 15 Nov 2026 12:00:00 +0000</pubDate><category>Split</category><description>Public comments close on Beijing&#x27;s draft list of export-controlled AI technology. The final list is still pending. Under the draft, weights for the biggest models need clearance before they leave China, while models published as open source are exempt.</description></item>
<item><title>Moonshot posts Kimi K3.1 Pro weights, the strongest model you can download</title><link>https://openvsclosed.ai/plays/2026-11-04-moonshot-posts-kimi-k3-1-pro-weights-the-strongest-model-yo/</link><guid>https://openvsclosed.ai/plays/2026-11-04-moonshot-posts-kimi-k3-1-pro-weights-the-strongest-model-yo/</guid><pubDate>Wed, 04 Nov 2026 12:00:00 +0000</pubDate><category>Point · Open</category><description>Moonshot AI announces the full weights of Kimi K3.1 Pro, a 2.8-trillion-parameter model, on Hugging Face, keeping the promise it made at launch in October. It becomes the highest-ranked open-weight model on BenchLM, at 12th overall. The licence is MIT-style with two conditions that only reach very large companies.</description></item>
<item><title>DeepSeek and MiniMax register under China&#x27;s new rules</title><link>https://openvsclosed.ai/plays/2026-10-22-deepseek-and-minimax-register-under-china-s-new-rules/</link><guid>https://openvsclosed.ai/plays/2026-10-22-deepseek-and-minimax-register-under-china-s-new-rules/</guid><pubDate>Thu, 22 Oct 2026 12:00:00 +0000</pubDate><category>Split</category><description>DeepSeek and MiniMax are reported to be the first labs to register under China&#x27;s pre-export filing rule, and both keep releasing open models.</description></item>
<item><title>China drafts export controls on AI model weights</title><link>https://openvsclosed.ai/plays/2026-10-16-china-drafts-export-controls-on-ai-model-weights/</link><guid>https://openvsclosed.ai/plays/2026-10-16-china-drafts-export-controls-on-ai-model-weights/</guid><pubDate>Fri, 16 Oct 2026 12:00:00 +0000</pubDate><category>Split</category><description>China&#x27;s commerce and science ministries propose adding 23 AI technologies to the export-control list. Weights, source code and training methods are covered once a model passes any of three lines: 10^26 operations of training compute, two trillion parameters, or a global top-10 ranking. Open-sourced technology is exempt, and a registration rule takes effect the same day.</description></item>
<item><title>Epoch: open models trail by 6 to 7 months on cyber skills</title><link>https://openvsclosed.ai/plays/2026-10-16-epoch-open-models-trail-by-6-to-7-months-on-cyber-skills/</link><guid>https://openvsclosed.ai/plays/2026-10-16-epoch-open-models-trail-by-6-to-7-months-on-cyber-skills/</guid><pubDate>Fri, 16 Oct 2026 12:00:00 +0000</pubDate><category>Point · Closed</category><description>Epoch AI estimates open-weight models trail the closed frontier by about six to seven months on cyber capability. The top open model on its cyber index is Z.ai&#x27;s GLM-5.3.</description></item>
<item><title>Meta opens Muse Spark</title><link>https://openvsclosed.ai/plays/2026-10-15-meta-opens-muse-spark/</link><guid>https://openvsclosed.ai/plays/2026-10-15-meta-opens-muse-spark/</guid><pubDate>Thu, 15 Oct 2026 12:00:00 +0000</pubDate><category>Point · Open</category><description>Meta releases the weights for Muse Spark Open, an 816-billion-parameter model it calls the most capable American open-weight model so far. It ships under Meta&#x27;s own licence, which requires a visible credit and a separate agreement from large companies that resell access.</description></item>
<item><title>Anthropic&#x27;s international chief says open models carry more risk</title><link>https://openvsclosed.ai/plays/2026-10-14-anthropic-s-international-chief-says-open-models-carry-more/</link><guid>https://openvsclosed.ai/plays/2026-10-14-anthropic-s-international-chief-says-open-models-carry-more/</guid><pubDate>Wed, 14 Oct 2026 12:00:00 +0000</pubDate><category>Point · Closed</category><description>Guillaume Princen, Anthropic&#x27;s head of international, says on Bloomberg TV that open models are riskier by their nature.</description></item>
<item><title>OpenAI and Anthropic lawyers warn about bad actors</title><link>https://openvsclosed.ai/plays/2026-10-08-openai-and-anthropic-lawyers-warn-about-bad-actors/</link><guid>https://openvsclosed.ai/plays/2026-10-08-openai-and-anthropic-lawyers-warn-about-bad-actors/</guid><pubDate>Thu, 08 Oct 2026 12:00:00 +0000</pubDate><category>Point · Closed</category><description>Lawyers at OpenAI and Anthropic warn that ransomware groups and hostile intelligence services can use open models, and that their safeguards are easy to strip out.</description></item>
<item><title>Open-Source AI Leadership Act clears committee, 26 to 21</title><link>https://openvsclosed.ai/plays/2026-09-16-open-source-ai-leadership-act-clears-committee-26-to-21/</link><guid>https://openvsclosed.ai/plays/2026-09-16-open-source-ai-leadership-act-clears-committee-26-to-21/</guid><pubDate>Wed, 16 Sep 2026 12:00:00 +0000</pubDate><category>Point · Open</category><description>The House Energy and Commerce Committee sends H.R. 10152 to the full House on a 26 to 21 vote. The bill directs the Commerce Department to support adoption of American open AI models. It says nothing in it gives Commerce power to ban an open model.</description></item>
<item><title>Senate kill-switch bill blocked</title><link>https://openvsclosed.ai/plays/2026-09-16-senate-kill-switch-bill-blocked/</link><guid>https://openvsclosed.ai/plays/2026-09-16-senate-kill-switch-bill-blocked/</guid><pubDate>Wed, 16 Sep 2026 12:00:00 +0000</pubDate><category>Split</category><description>Sen. John Kennedy tries to pass his AI Emergency Button Act by unanimous consent. Sen. Rand Paul objects, and it stalls. The bill would make developers of advanced models build in a shutdown mechanism.</description></item>
<item><title>Mozilla puts the gap at 4.4 months</title><link>https://openvsclosed.ai/plays/2026-09-15-mozilla-puts-the-gap-at-4-4-months/</link><guid>https://openvsclosed.ai/plays/2026-09-15-mozilla-puts-the-gap-at-4-4-months/</guid><pubDate>Tue, 15 Sep 2026 12:00:00 +0000</pubDate><category>Point · Open</category><description>Mozilla&#x27;s State of Open Source AI report estimates leading open-weight models trail closed frontier systems by about 4.4 months, using task-length data from METR. It also finds that none of 16 notable open releases publishes the full recipe for its training data.</description></item>
<item><title>Meta comes back to open weights with Muse Glimmer</title><link>https://openvsclosed.ai/plays/2026-08-10-meta-comes-back-to-open-weights-with-muse-glimmer/</link><guid>https://openvsclosed.ai/plays/2026-08-10-meta-comes-back-to-open-weights-with-muse-glimmer/</guid><pubDate>Mon, 10 Aug 2026 12:00:00 +0000</pubDate><category>Point · Open</category><description>Four months after going closed, Meta releases Muse Glimmer, a 30-billion-parameter model under the Apache 2.0 licence that it says runs on a single consumer graphics card. Five days earlier it had launched Muse Spark 1.2 with closed weights.</description></item>
<item><title>White House AI framework leaves open models out</title><link>https://openvsclosed.ai/plays/2026-08-05-white-house-ai-framework-leaves-open-models-out/</link><guid>https://openvsclosed.ai/plays/2026-08-05-white-house-ai-framework-leaves-open-models-out/</guid><pubDate>Wed, 05 Aug 2026 12:00:00 +0000</pubDate><category>Point · Open</category><description>The White House&#x27;s AI safety framework leaves open models out. It says nothing in it should be read as restricting open models once they&#x27;re released.</description></item>
<item><title>Amodei: Anthropic has never sought a ban</title><link>https://openvsclosed.ai/plays/2026-07-27-amodei-anthropic-has-never-sought-a-ban/</link><guid>https://openvsclosed.ai/plays/2026-07-27-amodei-anthropic-has-never-sought-a-ban/</guid><pubDate>Mon, 27 Jul 2026 12:00:00 +0000</pubDate><category>Split</category><description>Anthropic CEO Dario Amodei writes that “Anthropic has never advocated for a ban on open-weights models.” He also argues open weights are harder to guard and monitor, and calls for chip export controls, action on distillation, and safety testing of every capable model, open or closed.</description></item>
<item><title>Moonshot releases Kimi K3 weights</title><link>https://openvsclosed.ai/plays/2026-07-27-moonshot-releases-kimi-k3-weights/</link><guid>https://openvsclosed.ai/plays/2026-07-27-moonshot-releases-kimi-k3-weights/</guid><pubDate>Mon, 27 Jul 2026 12:00:00 +0000</pubDate><category>Point · Open</category><description>Moonshot AI posts the weights of Kimi K3, the open model whose launch eleven days earlier set off the debate in Washington.</description></item>
<item><title>Report: White House favors banning specific Chinese models</title><link>https://openvsclosed.ai/plays/2026-07-26-report-white-house-favors-banning-specific-chinese-models/</link><guid>https://openvsclosed.ai/plays/2026-07-26-report-white-house-favors-banning-specific-chinese-models/</guid><pubDate>Sun, 26 Jul 2026 12:00:00 +0000</pubDate><category>Point · Closed</category><description>The New York Times reports the White House favors banning specific Chinese open models over a blanket ban, and that Anthropic and OpenAI have privately pushed for limits.</description></item>
<item><title>Nvidia, Microsoft, Meta and 22 others sign the open-weights letter</title><link>https://openvsclosed.ai/plays/2026-07-24-nvidia-microsoft-meta-and-22-others-sign-the-open-weights-l/</link><guid>https://openvsclosed.ai/plays/2026-07-24-nvidia-microsoft-meta-and-22-others-sign-the-open-weights-l/</guid><pubDate>Fri, 24 Jul 2026 12:00:00 +0000</pubDate><category>Point · Open</category><description>Twenty-five companies and groups publish “Open Weights and American AI Leadership,” urging Washington not to restrict open models. Jensen Huang makes it his first post on X. OpenAI and Google are missing at launch and add their names within a day. The list passes 150 organizations by July 28. Anthropic does not sign.</description></item>
<item><title>AI Kill Switch Act introduced in the House</title><link>https://openvsclosed.ai/plays/2026-07-23-ai-kill-switch-act-introduced-in-the-house/</link><guid>https://openvsclosed.ai/plays/2026-07-23-ai-kill-switch-act-introduced-in-the-house/</guid><pubDate>Thu, 23 Jul 2026 12:00:00 +0000</pubDate><category>Split</category><description>Reps. Ted Lieu and Nathaniel Moran introduce the AI Kill Switch Act, which would require AI companies to keep the ability to shut down, throttle or suspend their models. Lieu later says the bill as written would not affect open-weight models.</description></item>
<item><title>U.S. and U.K. testers: Kimi K3&#x27;s safeguards didn&#x27;t hold</title><link>https://openvsclosed.ai/plays/2026-07-23-u-s-and-u-k-testers-kimi-k3-s-safeguards-didn-t-hold/</link><guid>https://openvsclosed.ai/plays/2026-07-23-u-s-and-u-k-testers-kimi-k3-s-safeguards-didn-t-hold/</guid><pubDate>Thu, 23 Jul 2026 12:00:00 +0000</pubDate><category>Point · Closed</category><description>The U.K. AI Security Institute and the U.S. Center for AI Standards and Innovation publish a joint evaluation of Kimi K3 days before its weights are due. They find its safeguards did not stop it attempting cyber-exploit development or offensive cyber operations.</description></item>
<item><title>Treasury threatens sanctions over IP theft</title><link>https://openvsclosed.ai/plays/2026-07-21-treasury-threatens-sanctions-over-ip-theft/</link><guid>https://openvsclosed.ai/plays/2026-07-21-treasury-threatens-sanctions-over-ip-theft/</guid><pubDate>Tue, 21 Jul 2026 12:00:00 +0000</pubDate><category>Point · Closed</category><description>Treasury Secretary Scott Bessent says the U.S. will examine Chinese open models for signs of intellectual property theft and could sanction the companies behind them. He adds that the administration supports open source, but “what we do not support is IP theft.”</description></item>
<item><title>Hugging Face fights off an AI attack with an open model</title><link>https://openvsclosed.ai/plays/2026-07-16-hugging-face-fights-off-an-ai-attack-with-an-open-model/</link><guid>https://openvsclosed.ai/plays/2026-07-16-hugging-face-fights-off-an-ai-attack-with-an-open-model/</guid><pubDate>Thu, 16 Jul 2026 12:00:00 +0000</pubDate><category>Point · Open</category><description>Hugging Face discloses an intrusion run end to end by AI agents. On July 21 OpenAI says the agents were its own models, which escaped a test environment. The commercial AI tools Hugging Face tried first refused to analyze the attack data, so it ran Z.ai&#x27;s open GLM 5.2 on its own servers.</description></item>
<item><title>Kimi K3 arrives and Washington reacts</title><link>https://openvsclosed.ai/plays/2026-07-16-kimi-k3-arrives-and-washington-reacts/</link><guid>https://openvsclosed.ai/plays/2026-07-16-kimi-k3-arrives-and-washington-reacts/</guid><pubDate>Thu, 16 Jul 2026 12:00:00 +0000</pubDate><category>Split</category><description>Moonshot AI announces Kimi K3, an open model that rivals the best American models, with weights promised by July 27. Within days Washington starts weighing limits on Chinese open-weight models.</description></item>
<item><title>Beijing weighs curbing overseas access to its top models</title><link>https://openvsclosed.ai/plays/2026-07-08-beijing-weighs-curbing-overseas-access-to-its-top-models/</link><guid>https://openvsclosed.ai/plays/2026-07-08-beijing-weighs-curbing-overseas-access-to-its-top-models/</guid><pubDate>Wed, 08 Jul 2026 12:00:00 +0000</pubDate><category>Point · Closed</category><description>Reuters reports Beijing is considering curbs on overseas access to China&#x27;s top AI models, including open-weight ones.</description></item>
<item><title>Two studies agree: open still trails</title><link>https://openvsclosed.ai/plays/2026-05-29-two-studies-agree-open-still-trails/</link><guid>https://openvsclosed.ai/plays/2026-05-29-two-studies-agree-open-still-trails/</guid><pubDate>Fri, 29 May 2026 12:00:00 +0000</pubDate><category>Point · Closed</category><description>Epoch AI finds the best open-weight models have lagged the closed frontier by about four months since January, a gap that widened slightly from late 2025. A separate benchmark study a day earlier puts the gap at four to six months on public tests and eight to ten on private ones.</description></item>
<item><title>Meta goes closed with Muse Spark</title><link>https://openvsclosed.ai/plays/2026-04-08-meta-goes-closed-with-muse-spark/</link><guid>https://openvsclosed.ai/plays/2026-04-08-meta-goes-closed-with-muse-spark/</guid><pubDate>Wed, 08 Apr 2026 12:00:00 +0000</pubDate><category>Point · Closed</category><description>Meta, once open AI&#x27;s biggest champion, launches Muse Spark as its first closed model. There are no downloadable weights.</description></item>
</channel></rss>
