Unverified50% confidenceFactExact time
Epoch AI估算互联网上可公开获取的高质量文本数据总量约在4.6万亿到17万亿token之间
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
8/27/2026
First Seen
Valid until: 11/25/2026
Sources
Related Entities
Related Claims
Unverified截至2024年，互联网档案馆数字馆藏包含超过3500万本书籍和文本、145亿个网页快照、超过1400万条音频记录和600万段视频，年运营预算约5000万美元69% similarUnverifiedWeb scraping is one of the highest-cost tool call types in AI Agents, as the raw HTML of an ordinary webpage typically contains tens of thousands to hundreds of thousands of characters.68% similarUnverifiedThe serialized DOM tree of a modern webpage may contain thousands or even tens of thousands of tokens, resulting in enormous token consumption for AI browser automation operations.68% similarUnverified整个互联网可抓取的公开网页数据（Common Crawl）约400TB67% similarUnverified据Sensity AI统计，2023年网络上检测到的deepfake内容数量同比增长超过900%67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/810362API
curl https://kongchang.com/api/v1/knowledge/claims/810362MCP
get_claim(id=810362)