Latent Space 播客举办了一期交叉节目，讨论了过去一年人工智能的快速发展，特别是从预训练转向推理时间缩放定律的转变。发言者指出，尽管 Llama 等开源模型可用，但它们并未得到广泛采用，这令人惊讶。对话还涉及了人工智能代理的兴起、垂直和水平人工智能初创公司之间的辩论，以及领域特定模型训练和代理经验日益增长的重要性。 AI
排序理由 该集群包含讨论人工智能趋势和行业人士见解的播客节目，而不是特定的研究论文、模型发布或重要的行业事件。
AI 生成摘要 · Google Gemini · 来自 4 个来源。 我们如何撰写摘要 →
<p>Today, we check in a year after the <a href="https://www.latent.space/p/unsupervised-learning" target="_blank">first </a><a href="https://www.latent.space/p/unsupervised-learning" target="_blank"><strong>Unsupervised Learning x Latent Space Crossover special</strong></a><stron…
<p><em>If you’re in SF: Join us for the </em><a href="https://lu.ma/poke" target="_blank">Claude Plays Pokemon hackathon</a><em> this Sunday!</em></p><p><em>If you’re not: Fill out </em><a href="https://www.surveymonkey.com/r/57QJSF2" target="_blank">the 2025 State of AI Eng</a><…
<p><em>Part 2 of our podcast feed swap weekend! Check out </em><a href="https://www.latent.space/p/cogrev-tinystories#details" target="_blank"><em>Cognitive Revolution</em></a><em> as well.</em></p><p>"Data" Dan Whitenack has been co-host of the Practical AI podcast for the past …
<p>Daniel had the chance to sit down with @swyx and Alessio from the <a href="https://www.latent.space/podcast">Latent Space pod</a> in SF to talk about current AI trends and to highlight some key learnings from past episodes. The discussion covers open access LLMs, smol models, …