I read the latest Silicon Valley AI article by Meng Xing, a partner at Wuyuan Capital, and it's so interesting: Is the Silicon Valley consensus to lay off 80% of humanity? Why can't Musk's xAI be achieved? Many startup companies are starting to recruit new roles as "AI builders"? Are CEOs all buying Bitcoin, building bunkers, and installing bulletproof glass in their homes? The future situation may be that 10 people do the work of the past 100 people, get 20 shares of money, and then 90 people become unemployed When AI can do everything, human value is shifting from "knowing what to do" to judging "what is worth doing and what should not be done". Some notes: 1. The entire Meta has tens of thousands of engineers, all of whom are using Claude Code to create an internal token consumption ranking, and may be laid off at the end; Meta has already laid off tens of thousands of employees, replacing the cost of personnel with token costs. The Silicon Valley consensus is that technology companies should lay off 80% of their workforce 2. Musk used to do SpaceX and Tesla, essentially doing systems engineering: the link is long, involving software, hardware, and supply chain, each with innovation space, but ultimately it is an end-to-end engineering problem. He is good at identifying key leverage points in this long chain and then compressing the timeline to overcome them. Cascade rocket engines and reusable landings are both products of this mindset. The problem with xAI is that there is no global planning, only sprinting. A co-founder of xAI said last year that there were two things he didn't expect: first, the competition was so fierce, and second, there were so few opportunities for application innovation in the AI era, all of which were eaten up by models. 3. To be more radical, today's so-called 'AI native organizations' sound very sexy - having each department streamline their workflow, digitize the parts that can be intervened by AI, and write them as skills. But in essence, it is a distillation of oneself through human flesh: you turn your abilities into machine skills, and when the company obtains your skills, it has actually completed AI transformation. Whether to lay off employees based on this is a moral question. Meta is doing this today. 4. The entire Silicon Valley is keeping an eye on Meta. If its experiment is successful - revenue doesn't fall, efficiency really improves - other big companies will quickly follow suit, and layoffs will become the norm in the industry. Moreover, layoffs have a cruel self acceleration mechanism: at the beginning, everyone dared not lay off for fear of damaging morale; Once it becomes a norm, the more it is cut, the faster and the less it feels. 5. Many startup companies are starting to hire a new role called "AI builder" - merging product managers, front-end engineers, and back-end engineers into one. Another approach is to merge composite positions of data scientists and machine learning engineers, as well as content integration operators who combine writing, delivery, and operations. 6. This seemingly distributed innovation world is actually extremely centralized at the bottom. This center is NVIDIA. If you can provide a stable API service today, such as Claude's API, with 99th percentile stability, you can sell it at two to three times the official API price. Before 2028, no AI company can significantly widen the gap by relying on heap computing power. The constraint of computing power objectively strengthens the oligopoly pattern of the large model market - it's not that anyone doesn't work hard, it's that the manufacturing cycle of the physical world is so slow. The power structure behind it is clear: whoever has the card is powerful, and who has the card is determined by Nvidia. CoreWeave, Lambda, and Nebius, which were launched today, are all backed by NVIDIA. 7. During this trip to Silicon Valley, I repeatedly heard my friends discussing the same thing seriously: buying Bitcoin, building bunkers, and installing bulletproof glass in their homes. They were not joking. The company that provides residential security for CEOs has achieved its highest growth level since 2003. Full text: https://mp.weixin. (qq.com)/s/kwErGjX231e2efVWhERzTw