←Search results

Search “Hugging Face”

By relevance · Total 1193 results Clear search

Thursday

Wednesday

  3. Platformer

    Does intelligence need a hard cap?

    At The Curve conference held in Berkeley, several speakers proposed that limits should be considered on the level of intelligence large language models can reach; depending on how strict the limits are, this could amount to a de facto ban on superintelligence. Anthropic CEO Dario Amodei has called for "some kind of speed limit" on recursive self-improvement, but any such limit would require enforcement capabilities that do not currently exist.

    upload in progress, 0

    Matched fieldsBody

    #ai-news#Industry news#Safety / alignment#OpenAI

Tuesday

  2. Interconnects

    Why the Online Risk Debate Has Lost Focus: Open Weights, Ideology, and Trade-offs

    Nathan Lambert argues that the current debate over the online risks of open models has reached a stalemate: although Anthropic's report on GLM-5.3 as an offensive cyber tool is technically sound, it sidesteps key questions such as "what would happen if open models were banned" and "why do Chinese companies believe it is safe to release them." He points out that most recorded cyberattacks to date have been carried out by closed models, and that if open-weight models were banned on cyber-risk grounds, it might also be necessary to ban public APIs for frontier closed models.

    Nathan Lambert argues that the current debate over the online risks of open models has reached a stalemate: although Anthropic's report on GLM-5.3 as an offensive cyber tool is technically sound, it sidesteps key questions such as "what would happen if open models were banned" and "why do Chinese companies believe it is safe to release them." He points out that most recorded cyberattacks to date have been carried out by closed models, and that if open-weight models were banned on cyber-risk grou

    Matched fieldsBody

    #newsletter#Industry news#Safety / alignment#Open-source ecosystem

Monday

  3. MarkTechPost

    The Evolution of Qwen: From 7B to 2.4T, a Full Retrospective of Alibaba's AI Model Series

    Alibaba Cloud's Qwen series began with Tongyi Qianwen in April 2023, open-sourced Qwen-7B in August 2023, launched the first Qwen3-Max-Preview with over 1 trillion parameters in September 2025, and ultimately evolved to an open-weight model with 2.4 trillion parameters. In 2024, three generations—Qwen1.5, Qwen2, and Qwen2.5—were released in succession; Qwen2.5's pre-training data grew from 7 trillion to 18 trillion tokens, and model downloads exceeded 40 million. In 2025, Qwen3 adopted a hybrid thinking mode, covering 119 languages, and Qwen3-Next-80B-A3B claims training costs of only 10% of Qwen3-32B and 10x higher inference throughput above 32K tokens.

    Alibaba Cloud's Qwen series began with Tongyi Qianwen in April 2023, open-sourced Qwen-7B in August 2023, launched the first Qwen3-Max-Preview with over 1 trillion parameters in September 2025, and ultimately evolved to an open-weight model with 2.4 trillion parameters. In 2024, three generations—Qwen1.5, Qwen2, and Qwen2.5—were released in succession; Qwen2.5's pre-training data grew from 7 trillion to 18 trillion tokens, and model downloads exceeded 40 million. In 2025, Qwen3 adopted a hybrid

    Matched fieldsBody

    #ai-vertical#Industry news#Open-source ecosystem#English text is being prepared.