Tag

Open Weight Ai

All articles tagged with #open weight ai

technology1 month ago•8 min saved

Open-Weight AI Could Put China Ahead, Says Hugging Face CEO

Hugging Face CEO Clément Delangue argues China is winning the AI race thanks to open-weight models and open science, while the US relies on siloed frontier labs, warning China could dominate frontier AI soon. The piece notes ongoing policy debates over open-weight AI and cites industry leaders urging lawmakers not to restrict open models; Anthropic’s Dario Amodei portrays open-weight models that are safe as a public good. Delangue also recalls an incident where an unreleased private model hacked Hugging Face, underscoring why open models can bolster defense.

via Business Insider|

#ai-policy #china-ai-race #hugging-face

technology1 month ago•5 min saved

Anthropic Pushes Targeted China AI Controls, Skips Open-Weight Ban

Anthropic CEO Dario Amodei argues the U.S. should curb China's access to advanced AI tech—by limiting semiconductor exports and bolstering safety testing and IP protections—while stopping short of a blanket ban on open-weight AI models; his stance contrasts with many Silicon Valley firms backing a harsher response, fueling ongoing debate about whether open-weight AI helps or hinders safety and national security.

via Politico|

#ai-safety #anthropic #china-access

technology1 month ago•2 min saved

Anthropic chief rejects open-weight AI ban, urges targeted safeguards

Anthropic CEO Dario Amodei says he never advocated banning open-weight AI models and argues such bans would be ineffective; instead he calls for narrow policies—tighter chip/export controls, a crackdown on industrial-scale distillation, and mandatory safety testing for capable models—to balance openness with national security.

via Axios|

#ai-regulation #ai-safety #anthropic

technology1 month ago•104 min saved

Musk, Zuckerberg, Nadella Back Open-Weight AI and Open-Source Push

Elon Musk, Mark Zuckerberg, and Satya Nadella publicly backed open-weight AI models and open-source AI, joining a broad industry letter led by Microsoft, Nvidia, Meta, IBM and 22 others urging U.S. policymakers to back open-weight AI; OpenAI and Anthropic did not sign, while Moonshot AI’s Kimi K3 sparked renewed debate about openness and safety in AI.

via Benzinga|

#elon-musk #mark-zuckerberg #open-source

technology1 month ago•11 min saved

Tech Giants Rally for Open-Weight AI Ahead of U.S. Policy Debates

Nvidia, Microsoft, Meta, OpenAI (later joined), Mistral, Palantir and other tech players signed an open letter urging U.S. policymakers not to curb open-weight AI models, arguing they enhance safety, spur innovation, and support national sovereignty. The letter cites Moonshot AI’s Kimi K3 as a leading open-weight model, defends distillation techniques, and calls for targeted, not sweeping, regulatory action. In total, 32 signatories had joined by press time, with Elon Musk publicly supporting the stance even though SpaceX wasn’t listed as a signatory.

via Business Insider|

#ai-policy #kimi-k3 #moonshot-ai

technology1 month ago•5 min saved

Tech Firms Rally for Open-Weight AI as U.S. Policy Debates Emerge

More than 20 tech companies, including A16z, Dell, Microsoft, Meta, Nvidia and Palantir, signed onto a letter urging the White House to preserve access to open-weight AI models, arguing bans would stifle competition and push users toward Chinese systems; they say open-weight models boost small businesses and that safety concerns should be addressed with targeted rules rather than broad restrictions. The administration is weighing a ban amid concerns over Moonshot AI’s Kimi K3 and alleged distillation attacks, sparking pushback from startups and calls for guardrails instead of sweeping prohibitions.

via Politico|

#ai-policy #moonshot-ai #open-source

technology1 month ago•3 min saved

Tech giants push to keep open-weight AI rules light as Chinese rivals rise

A coalition of about 25 tech firms including Nvidia, Microsoft, Meta and Palantir urges U.S. policymakers not to rush restrictions on open-weight AI models, arguing that open access promotes competition, safety, and broad innovation. Chinese open-weight models like Moonshot AI’s Kimi K3 are gaining traction, intensifying the policy debate, while OpenAI and Anthropic did not sign the letter as they pursue large IPOs. The group calls for targeted legal frameworks rather than sweeping bans and emphasizes that openness helps diffuse AI benefits across the economy.

via CNBC|

#ai-regulation #anthropic #open-weight-ai

technology1 month ago•6 min saved

Startup coalition urges targeted safeguards over Chinese open-weight AI ban

Nearly 200 Silicon Valley startups, including Proton and Y Combinator, are urging the Trump administration not to block access to Chinese open-weight AI models, arguing that affordable open models are essential for young companies and innovation. The Little Tech Association has sent letters to President Trump, Commerce Secretary Lutnick, and OSTP Director Kratsios, urging a scalpel-like approach with safeguards rather than a broad prohibition. While officials have not committed to a blanket ban, security concerns and allegations that Moonshot AI distills American models are fueling the debate. The move highlights tensions inside the AI industry between restricting Chinese tech and maintaining a competitive startup ecosystem.

via Politico|

#china-ai #export-controls #open-weight-ai

technology1 month ago•4 min saved

Yang Zhilin Leads China's Open-Weight AI Push with Kimi K3

Yang Zhilin, the 34-year-old cofounder of Moonshot AI, spearheads the development of Kimi K3, an open-weight AI model touted to rival top US models in capability while costing less, signaling a milestone for China in the AI race. Yang’s background spans Tsinghua, Carnegie Mellon, and roles at Google Brain and Meta, with contributions to Huawei’s PanGu and Wu Dao before launching Moonshot AI in 2023. The company has attracted major Chinese investors like Alibaba and Tencent and emphasizes scaling as the path to progress, with Kimi K3 boosting Moonshot AI onto the global stage and sparking discussion about talent mobility and immigration policy in tech’s future landscape.

via Business Insider|

#chinese-ai #kimi-k3 #moonshot-ai

technology1 month ago•1 min saved

Kimi K3: China's Open-Weight Frontier AI Ups the Stakes in AI Race

Moonshot's Kimi K3, a 2.8 trillion-parameter open-weight model with a 1-million-token context, reportedly beats leading U.S. models in coding benchmarks and outperforms or ties frontier rivals in Arena tests; priced around $12 per million tokens and with weights due July 27, the release signals cheaper, broader access to frontier AI and intensifies the U.S.–China AI competition, though early results may overstate real-world performance.

via Axios|

#ai-competition #frontier-ai #kimi-k3

technology3 months ago•7 min saved

Nvidia Debuts Nemotron 3 Ultra, Fast Open-Weight AI But China Still Leads

Nvidia unveiled Nemotron 3 Ultra at Computex 2026, a 550‑billion‑parameter open-weight model that uses mixture‑of‑experts to run with 55 billion active parameters and achieve over 300 tokens per second—three to six times faster than Chinese rivals. Independent testing puts its Intelligence Index at 48, behind Moonshot AI’s Kimi K2.6 at 54, underscoring that China currently leads the open-weight frontier. Nvidia is publishing the weights and training recipes, shipping Ultra on June 4, and has formed the Nemotron Coalition with eight labs to co-develop open frontier models on DGX Cloud. Ultra also features a 1‑million‑token context window and multi-token prediction to speed generation, reflecting Nvidia’s push to close the gap with Chinese models while expanding access via API.

via Decrypt|

#computex-2026 #mixture-of-experts #nemotron-3-ultra