I havent really looked into them deeply but I find it hard to think that they are not all in on this.
That said, it does stress the fine line. How many bots do you produce knowing there is a lot of advancement in the next year? Producing a million v1 might not be sensible. But there has to be enough to keep training the system on many different tasks.
I agree. They need to find the balance. There is a very big chance that those v1 will be useless very quickly, but they still need to build a good amount. It would be great if they can make them so they are kind of reusable to save costs.