They definitely are scared. Of their competitors in China eating their lunch. And dinner. And breakfast. And drink all the champagne.
These labs have no real moats except sunk cost. Anthropic itself proves that OpenAI's first-mover advantage is already minimal - they caught up in less than 18 months of product-market fit. Chinese open models will destroy any long term value these guys have, except that there is so much money riding on their faked-and-will-never-make lasting excellence, that if they get a little bit of protection by erecting barriers to entry for others, they might actually survive on sunk cost alone.
They definitely are scared. Of their competitors in China eating their lunch. And dinner. And breakfast. And drink all the champagne.
These labs have no real moats except sunk cost. Anthropic itself proves that OpenAI's first-mover advantage is already minimal - they caught up in less than 18 months of product-market fit. Chinese open models will destroy any long term value these guys have, except that there is so much money riding on their faked-and-will-never-make lasting excellence, that if they get a little bit of protection by erecting barriers to entry for others, they might actually survive on sunk cost alone.