马斯克质疑大模型层零利润假设，认为供应商或难支撑

Wait, what? Are you listening to what you are writing @GavinSBaker?   If there are going to be ZERO ...

Wait, what? Are you listening to what you are writing @GavinSBaker ? If there are going to be ZERO profits at the model layer, as in your hypothetical, why would anyone value the providers at $2 trillion? In fact, why would anyone hand them a dime? If people realize that zero profits is a real possibility and cut them, off the providers will no longer be in a position to spend a trillions dollars on compute to begin with. Of course that would hurt infrastructure plays. Your verbal gymnastics won’t change that. Also, Individual companies running open source, per your example, are unlikely make up for the lost demand from the frontier labs if they go bust. (Open AI and Anthropic account for a huge hunk of current demand and the backlog of business, as I am sure you know). It’s just not clear that the compute providers could survive if the two frontier labs ran out of cash. (Remember what happened to Exodus Communications? Most people will have never heard of them but they were the largest webhosting company going into the dotcom bubble. And then … poof!) And of course survey after survey shows corporate users are not finding ROI from the product and Gen AI remains error prone and limited, so demand may not reach the heights you are imagining. Finally, you attribute to me the idea of wanting AI to be dominated by a single player, which is pretty much the opposite of what I actually have been arguing nonstop for the last three years. That’s just plain careless. To quote yourself, “This is all really basic stuff. You seem smart. Think for maybe 5-10 seconds before you post a reply.” Gavin Baker @GavinSBaker 1 is just super obviously not true. If there was one lab, and only one lab, with 99.999999% margins, that would be a disaster for the neoclouds and everyone who supplied that single lab. How can this not be obvious to you? There are somewhere between 1000 and 1 million case studies across a variety of industries showing that monopsonies exercise power over their suppliers. Fragmentation and competition at the model layer is good for everyone who is not at the model layer. At an extreme - a world where companies like Latham & Watkins run fine-tuned open-source models on their own GPUs and Nvidia funds the open-source training runs - it is possible for there to be zero profits at the model layer and yet high returns for AI infrastructure supplies (semiconductors, datacenters and power in this example). This is all really basic stuff. You seem smart. Think for maybe 5-10 seconds before you post a reply. 🔗 View Quoted Tweet 💬 2 🔄 1 ❤️ 10 👀 4112 📊 3 ⚡