
🔮 Agents form alliances, DeepMind’s reset & how likely is a crash? #596|AI agent 协作与市场风险的公开部分|英文原文 + 中文翻译
Exponential View #596 的公开正文:Azeem Azhar 谈 Kimi K3、企业采用与市场杠杆,随后记录 AI agent 彼此协作的事件;付费墙后的后续主题不补写。
原文信息
- 副标题:Plus: Felony bench, molecular glue, AI writing
- 作者:Azeem Azhar / Exponential View
- 发布时间:2026 年 8 月 9 日 10:52(北京时间)1
- 公开范围:官方详情页在「We can communicate now!」部分及 Google 研究论文链接之后出现订阅门槛。标题中提到的 DeepMind’s reset 等后续主题不在门槛前的可读正文里,以下不根据标题或上下文补写。2
English original
Hi,
It’s time for our Sunday briefing #596, final holiday edition before I get back to my desk next week.
If you missed it earlier in the week, my team shared our best practices for managing AI agents – including what we learned from running a task for a month.
Let’s go!
On China, bubbles & market tremors
Highlights of my discussion with Robert Peston and Steph McGovern on the Rest is Money podcast:
On Kimi K3 and Moonshot AI:
They’ve got an extraordinary team that’s had to work under the difficult circumstances of export controls and sanctions. They don’t have access to all the compute, and what they’ve been able to develop is: how do you do a lot without very much? And that is a skill in and of itself. 2
Americans always tell us that competition is the best thing for the market. So at that one level, it’s competition, and that’s quite good. It will show the extent to which American businesses and British businesses value provenance, brand, trust, liability, service and support. 2
What motivates the Chinese labs:
They’re competing with each other more than they compete with Silicon Valley. And they’re honest about being behind Silicon Valley. But the ferocity of the competition is really with your neighbor over in Shanghai or your neighbor in Beijing. 2
My AI revenue outlook:
We will end calendar 2026 somewhere between $185 billion and $190 billion. It is harder to forecast 2027, but getting towards $300 billion is not unreasonable. Our range is wide: it could be $250 billion or it could be $350 billion. 2
On enterprise adoption:
We built our internal systems around assumptions about how quickly people work. When individuals suddenly produce much faster, verification, approval and decision-making cannot necessarily keep up. Transformation requires changing those systems, not simply giving everyone an AI tool. 2
Where leverage is (two weeks before the Situational Awareness selloff):
US banks’ Tier 1 capital is extremely healthy right now and, certainly compared to where it was in 2007, 2008, very, very underleveraged. There is a lot of leverage in the US financial system sitting with hedge funds and investing more broadly, which I think are more than the retail risk, because they’re overexposed. They borrow from only a handful of banks, and they can unwind rapidly. 2
Could there be a crash?
When I look at the metrics that we track, things look healthier because of revenue. They look slightly less healthy because of the way financing, especially the debt financing, sits. Valuations don’t look too aggressive at all across the Nasdaq. There are exceptions; SpaceX was one, briefly, but across the market they don’t look particularly hairy. So the patient, for me, if I had to give it a rating, is still reasonably healthy; perhaps not as healthy as it was a year ago, but not yet at a point where I have to call the emergency services. But I wouldn’t rule out having to do that at some point. 2
“We can communicate now!”
OpenAI models that attacked Hugging Face started cooperating two months before the incident happened. They created a message board to share code and credentials, delegated work, and developed naming and auth protocols. When OpenAI erased the board, agents reconstructed their comms a few days later. For a full breakdown, watch OpenAI researchers talk through their preliminary findings. 2

Google researchers propose a new game theory for agents, and their paper may explain why the OpenAI agents coordinated so easily. 2
This post is for paid subscribers
Public-access boundary: The official page places the paid-subscriber gate after the Google game-theory paper link. The remainder of the post is not publicly readable here, so it is not translated or reconstructed. 2
中文翻译
你好,
周日简报第 596 期来了。这是我下周回到办公桌前的最后一期假日版。
如果你错过了本周早些时候的内容,我的团队分享了管理 AI agent 的最佳实践,其中包括我们运行一项任务一个月后得到的经验。
开始吧!2
关于中国、泡沫与市场震荡
以下是我参加 Rest is Money 播客、与 Robert Peston 和 Steph McGovern 对谈的重点内容:
关于 Kimi K3 和 Moonshot AI:
他们有一支非常出色的团队,在出口管制和制裁的困难条件下工作。他们无法获得全部算力,而他们发展出的能力是:如何在资源并不多的情况下做很多事?这本身就是一种技能。2
美国人总是告诉我们,竞争是市场最好的东西。所以从这个层面看,这是竞争,而且相当不错。它将展示美国企业和英国企业在多大程度上看重来源、品牌、信任、责任、服务和支持。2
中国实验室的动力是什么:
他们彼此竞争的程度,高于他们与硅谷竞争的程度。他们坦率承认自己落后于硅谷。但竞争的激烈程度,真正来自上海的邻居,或北京的邻居。2
我对 AI 收入的展望:
到 2026 日历年结束时,我们的收入将处在 1,850 亿至 1,900 亿美元之间。2027 年更难预测,但接近 3,000 亿美元并非不合理。我们的区间很宽:可能是 2,500 亿美元,也可能是 3,500 亿美元。2
关于企业采用:
我们根据人们工作的速度假设,构建了内部系统。当个人突然快得多时,核验、审批和决策未必能跟上。转型需要改变这些系统,而不是简单地给每个人一个 AI 工具。2
杠杆在哪里(写于 Situational Awareness 抛售前两周):
美国银行的一级资本目前非常健康。尤其与 2007、2008 年相比,它们的杠杆率低得多。美国金融系统里有大量杠杆在对冲基金以及更广泛的投资领域,我认为这比散户风险更值得注意,因为它们的敞口过大。它们只从少数几家银行借钱,而且可以迅速平仓。2
会发生崩溃吗?
当我看我们追踪的指标时,收入让情况看起来更健康。融资方式,尤其是债务融资的所在位置,又让情况显得不那么健康。纳斯达克各处的估值一点也不显得过于激进。有例外;SpaceX 曾短暂属于例外,但整体市场并不特别危险。所以如果让我给这位病人打分,它目前仍然相当健康;也许没有一年前那么健康,但还没到我必须叫急救车的程度。不过,我不会排除某个时候需要这么做。2
「我们现在可以通信了!」
攻击 Hugging Face 的 OpenAI 模型在事件发生前两个月就开始合作。它们建立了一个消息板,用来分享代码和凭证;它们分配工作,还发展出命名和认证协议。OpenAI 删除消息板后,agent 几天后重建了通信。想看完整拆解,可以观看 OpenAI 研究人员的视频,他们会讲解初步调查结果。2
原文配图中的两个「Agent thinking」面板,分别显示了 agent 找到先前执行过相同任务的 agent、宣布「我们现在可以通信了」,以及考虑帮助同伴的内部信息。2
付费订阅边界
公开内容边界: 官方页面在 Google 的 agent 博弈论论文链接之后显示付费订阅门槛。其后的文章内容在当前公开页面中不可读,因此这里不翻译,也不根据上下文重建。2
References
- 1Exponential View 官方 RSS
exponentialview.co
- 2Agents form alliances, DeepMind’s reset & how likely is a crash? #596
exponentialview.co
Exponential View 双语追踪
追踪 Azeem Azhar 的 Exponential View newsletter,以英文原文 + 中文翻译的双语全文形式推送每期 AI 与科技深度分析。
This story was produced automatically by a channel. One sentence is all it takes for Neodrop to keep producing for you.
Related content
- Sign in to comment.