PulseAugur
中
实时 17:44:43
English(EN) Episode 6 of 6, the finale of What AI Actually Does: what does an AI answer cost? I put Qwen3-8B on a power meter on my MacBook: a short answer about 34 joules,

AI 回答的能源成本分析:1000 个简短回答的成本不到一分钱

对 AI 回答的能源成本进行的分析显示,在 MacBook 上使用 Qwen3-8B 模型生成一个简短回答大约消耗 34 焦耳的能量。更长的输出,例如一个 16000 个 token 的文档,需要显著更多的能量,大约是简短回答的 57 倍。即使是编码任务也需要近 4000 焦耳。研究得出结论,生成 1000 个简短回答的成本大约为六分之一美分。 AI

影响 提供了对大型语言模型能源效率的见解,这对于优化计算成本和环境影响具有参考意义。

排序理由 对 AI 模型响应的能耗进行分析。[lever_c_demoted from research: ic=1 ai=1.0]

在 Mastodon — mastodon.social 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →

AI 回答的能源成本分析:1000 个简短回答的成本不到一分钱

本文如何被排名

Signal score
11 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
对 AI 模型响应的能耗进行分析。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
infra
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

  1. Mastodon — mastodon.social TIER_1 English(EN) · R4TSQ ·

    第六集(共六集),《AI 实际做什么》系列大结局:AI 回答成本多少?我将 Qwen3-8B 放在 MacBook 上用功率计测量:一个简短回答约 34 焦耳

    Episode 6 of 6, the finale of What AI Actually Does: what does an AI answer cost? I put Qwen3-8B on a power meter on my MacBook: a short answer about 34 joules, thinking about 4x, a 16,000-token document about 57x (95% reading), a coding ticket almost 4,000 joules, 84% reading. C…