The cybersecurity skills of AI models means that AI has collapsed the labor and tooling gap that used to separate well-resourced, state-sponsored operations from individual operators.
【非共识】这一观点挑战了传统网络安全认知，认为AI正在消除国家级行动与个人行动者之间的能力差距。这意味着未来网络安全防御需要重新思考如何应对这种能力民主化现象，资源有限的小型组织或个人也能实施复杂的攻击。
While some experts have speculated that general models will win out in performance over specialized models—that scale and compute will beat curation—the success of these companies shows that the market is making a more nuanced bet.
市场正在形成一种更微妙的AI发展路径认知，表明通用模型与专业化模型可能在不同场景下各有优势。这种市场分歧暗示AI领域可能不会出现单一赢家，而是形成多元化发展格局。
It is the only AI service to show a clear upward trend over this short time period
令人惊讶的是：在所有被调查的AI服务中，Claude是唯一显示出明确上升趋势的工具。这表明尽管市场竞争激烈，Claude仍然成功地在众多AI服务中脱颖而出，实现了持续增长。
We document a clear trend where Chinese models overtook their counterparts built in the U.S. in the summer of 2025 and subsequently widened the gap over their western counterparts.
令人惊讶的是：这项研究表明，在2025年夏天，中国开源语言模型已经超越美国同行，并且这一差距还在不断扩大。这表明全球AI发展速度之快超出了许多人的预期，也反映了非西方国家在AI领域的快速崛起。
Introduction: AI is now recently everywhere but we still need humans
https://web.archive.org/web/20231108195303/https://axbom.com/aipower/
Per Axbom does a nice overview of actors and stakeholders to take into account when thinking about AI's impact and ethics. Some of these are mentioned in the [[EU AI Regulation]] but not all actors mentioned there are mentioned here I think: EU act not only defines users (of the application) but also users of the output of an application separately. This to ensure that outputs from un-checked or illegal applications outside the EU market are admissable to the EU market.